Fix incorrect region for region capture after compute shader change

This reverts commit f85a7ab205.

.gitignore

@@ -23,3 +23,6 @@ gsr-kms-server
 *.mov
 *.webm
 *.ts
+*.jpg
+*.jpeg
+*.png

README.md

@@ -7,6 +7,8 @@ similar to shadowplay on windows. This is the fastest screen recording tool for
 This screen recorder can be used for recording your desktop offline, for live streaming and for nvidia shadowplay-like instant replay,
 where only the last few minutes are saved.
 
+This software can also take screenshots.
+
 This is a cli-only tool, if you want an UI for this check out [GPU Screen Recorder GTK](https://git.dec05eba.com/gpu-screen-recorder-gtk/) or if you prefer a ShadowPlay-like UI then check out [GPU Screen Recorder UI](https://git.dec05eba.com/gpu-screen-recorder-ui/).
 
 Supported video codecs:
@@ -20,16 +22,19 @@ Supported audio codecs:
 * Opus (default)
 * AAC
 
-## Note
-This software works on X11 and Wayland on AMD, Intel and NVIDIA.
+Supported image formats:
+* JPEG
+* PNG
+
+This software works on X11 and Wayland on AMD, Intel and NVIDIA. Replay data is stored in RAM, not disk.
 ### TEMPORARY ISSUES
 1) Videos are in variable framerate format. Use MPV to play such videos, otherwise you might experience stuttering in the video if you are using a buggy video player. You can try saving the video into a .mkv file instead as some software may have better support for .mkv files (such as kdenlive). You can use the "-fm cfr" option to to use constant framerate mode.
 2) FLAC audio codec is disabled at the moment because of temporary issues.
 ### AMD/Intel/Wayland root permission
-When recording a window or when using the `-w portal` option under AMD/Intel no special user permission is required,
-however when recording a monitor (or when using wayland) the program needs root permission (to access KMS).\
+When recording a window or when using the `-w portal` option no special user permission is required,
+however when recording a monitor the program needs root permission (to access KMS).\
 This is safe in GPU Screen Recorder as the part that needs root access has been moved to its own small program that only does one thing.\
-For you as a user this only means that if you installed GPU Screen Recorder as a flatpak then a prompt asking for root password will show up when you start recording.
+For you as a user this only means that if you installed GPU Screen Recorder as a flatpak then a prompt asking for root password will show up once when you start recording.
 # Performance
 On a system with a i5 4690k CPU and a GTX 1080 GPU:\
 When recording Legend of Zelda Breath of the Wild at 4k, fps drops from 30 to 7 when using OBS Studio + nvenc, however when using this screen recorder the fps remains at 30.\
@@ -63,7 +68,8 @@ Here are some known unofficial packages:
 * Nix: [NixOS wiki](https://wiki.nixos.org/wiki/Gpu-screen-recorder)
 * openSUSE: [openSUSE software repository](https://software.opensuse.org/package/gpu-screen-recorder)
 * Fedora: [Copr](https://copr.fedorainfracloud.org/coprs/brycensranch/gpu-screen-recorder-git/)
-* OpenMandriva: [gpu-screen-recorder](https://github.com/OpenMandrivaAssociation/gpu-screen-recorder/tree/master)
+* OpenMandriva: [gpu-screen-recorder](https://github.com/OpenMandrivaAssociation/gpu-screen-recorder)
+* Solus: [gpu-screen-recorder](https://github.com/getsolus/packages/tree/main/packages/g/gpu-screen-recorder)
 
 # Dependencies
 GPU Screen Recorder uses meson build system so you need to install `meson` to build GPU Screen Recorder.
@@ -146,12 +152,8 @@ You have to reboot your computer after installing GPU Screen Recorder for the fi
 # Examples
 Look at the [scripts](https://git.dec05eba.com/gpu-screen-recorder/tree/scripts) directory for script examples. For example if you want to automatically save a recording/replay into a folder with the same name as the game you are recording.
 
-# Reporting bugs
-Issues are reported on this Github page: [https://github.com/dec05eba/gpu-screen-recorder-issues](https://github.com/dec05eba/gpu-screen-recorder-issues).
-# Contributing patches
-See [https://git.dec05eba.com/?p=about](https://git.dec05eba.com/?p=about) for contribution steps.
-# Donations
-See [https://git.dec05eba.com/?p=about](https://git.dec05eba.com/?p=about) for donation options.
+# Reporting bugs, contributing patches, questions or donation
+See [https://git.dec05eba.com/?p=about](https://git.dec05eba.com/?p=about).
 
 # Demo
 [![Click here to watch a demo video on youtube](https://img.youtube.com/vi/n5tm0g01n6A/0.jpg)](https://www.youtube.com/watch?v=n5tm0g01n6A)
@@ -181,4 +183,7 @@ You can record with desktop portal option (`-w portal`) instead which ignores ni
 ## Kdenlive says that the video is not usable for editing because it has variable frame rate
 To fix this you can either record the video in .mkv format or constant frame rate (-fm cfr).
 ## Colors look incorrect when recording HDR (with hevc_hdr/av1_hdr) or using an ICC profile
-The latest version of KDE Plasma breaks HDR and ICC profiles for recording applications. Wayland in general doesn't properly support recording HDR yet. Use desktop portal option (`-w portal`) for now to turn HDR recording into SDR and to be able to record with correct colors when using an ICC profile.
+KDE Plasma version 6.2 broke HDR and ICC profiles for screen recorders. This was changed in KDE plasma version 6.3 and recording HDR works now, as long as you set HDR brightness to 100% (which means setting "Maximum SDR Brightness" in KDE plasma display settings to 203) and set color accuracy to "Prefer color accuracy". If you want to convert HDR to SDR then record with desktop portal option (`-w portal`) instead.
+I don't know how well recording HDR works in wayland compositors other than KDE plasma.
+## GPU Screen Recorder starts lagging after 30-40 minutes when launching GPU Screen Recorder from steam command launcher
+This is a [steam issue](https://github.com/ValveSoftware/steam-for-linux/issues/11446). Prepend the gpu-screen-recorder command with `LD_PREFIX=""`, for example `LD_PREFIX="" gpu-screen-recorder -w screen -o video.mp4`.

TODO

@@ -2,7 +2,6 @@ Check for reparent.
 Quickly changing workspace and back while recording under i3 breaks the screen recorder. i3 probably unmaps windows in other workspaces.
 See https://trac.ffmpeg.org/wiki/EncodingForStreamingSites for optimizing streaming.
 Look at VK_EXT_external_memory_dma_buf.
-Allow setting a different output resolution than the input resolution.
 Use mov+faststart.
 Allow recording all monitors/selected monitor without nvfbc by recording the compositor proxy window and only recording the part that matches the monitor(s).
 Allow recording a region by recording the compositor proxy window / nvfbc window and copying part of it.
@@ -10,7 +9,6 @@ Support amf and qsv.
 Disable flipping on nvidia? this might fix some stuttering issues on some setups. See NvCtrlGetAttribute/NvCtrlSetAttributeAndGetStatus NV_CTRL_SYNC_TO_VBLANK https://github.com/NVIDIA/nvidia-settings/blob/d5f022976368cbceb2f20b838ddb0bf992f0cfb9/src/gtk%2B-2.x/ctkopengl.c.
 Replays seem to have some issues with audio/video. Why?
 Cleanup unused gl/egl functions, macro, etc.
-Add option to disable overlapping of replays (the old behavior kinda. Remove the whole replay buffer data after saving when doing this).
 Set audio track name to audio device name (if not merge of multiple audio devices).
 Add support for webcam, but only really for amd/intel because amd/intel can get drm fd access to webcam, nvidia cant. This allows us to create an opengl texture directly from the webcam fd for optimal performance.
 Reverse engineer nvapi so we can disable "force p2 state" on linux too (nvapi profile api with the settings id 0x50166c5e).
@@ -180,7 +178,7 @@ Test if `xrandr --output DP-1 --scale 1.5` captures correct size on nvidia.
 
 Fix cursor position and scale when scaling x11 display.
 
-Support surround audio in application audio recording. Right now only stereo sound is supported.
+Support surround audio in application audio recording. Right now only stereo and mono sound is supported.
 
 Support application audio recording without pulseaudio combined sink.
 
@@ -196,8 +194,7 @@ Add option to record audio from the recorded window only.
 
 Add option to automatically select best video codec available. Add -k best, -k best_10bit and -k best_hdr.
 
-HDR is broken on kde plasma > 6.2 because of change to how HDR metadata works. See https://github.com/dec05eba/gpu-screen-recorder-issues/issues/60.
-    Use wayland color management protocol when it's available: https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/14.
+Use wayland color management protocol when it's available: https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/14.
 
 Use different exit codes for different errors. Use one for invalid -w option, another one for invalid -a option for audio devices, etc. This is to make UI error reporting better.
     Document these exit codes in an exit code .md file, or finally create a manpage where this can be documented.
@@ -206,6 +203,8 @@ Ffmpeg fixed black bars in videos on amd when using hevc and when recording at s
     https://github.com/FFmpeg/FFmpeg/commit/bcfbf2bac8f9eeeedc407b40596f5c7aaa0d5b47
     https://github.com/FFmpeg/FFmpeg/commit/d0facac679faf45d3356dff2e2cb382580d7a521
     Disable gpu screen recorder black bar handling when using hevc on amd when the libavcodec version is the one that comes after those commits.
+    Also consider the mesa version, to see if the gpu supports this.
+    The version is libavcodec >= 61.28.100
 
 Use opengl compute shader instead of graphics shader. This might allow for better performance when games are using 100% of graphics unit which might fix issue with 100% gpu usage causing gpu screen recorder to run slow when not using vaapi to convert rgb to nv12(?).
 
@@ -221,3 +220,37 @@ Allow flv av1 if recent ffmpeg version and streaming to youtube (and twitch?) an
 Use explicit sync in pipewire video code: https://docs.pipewire.org/page_dma_buf.html.
 
 Support vaapi rotation. Support for it is added in mesa here: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/32919.
+
+Replay (and recording?) fails to save properly sometimes (especially for long videos). This is noticable with mp4 files since they get corrupt and become unplayable.
+    The entire video does seem to get saved (it's a large video file) and it seems to have the correct headers but it's not playable.
+
+Make it possible to save a shorter replay clip remotely. Maybe implement ipc first, to then also allow starting recording/stream while a replay is running.
+
+Add an option to pass http headers when streaming. Some streaming services require streaming keys to be passed in a http header instead of in the url as a parameter.
+
+When adding vulkan video support add VK_VIDEO_ENCODE_TUNING_MODE_LOW_LATENCY_KHR.
+
+Implement screenshot without invoking opengl (which is slow to start on some systems).
+
+Automatically use desktop portal on wayland when hdr is enabled (or night light) by checking if kms hdr metadata exists, if hdr video codec is not used.
+    Or maybe do this in the ui?
+
+Detect if cached portal session token is no longer valid (this can happen if the user switches to another wayland compositor).
+
+Support reconnecting (and setting things up again) if the audio server is restarted (for both device recording and app recording).
+
+Find out how nvidia-smi fixes nvenc not working on opensuse and do that ourselves instead of relying on nvidia-smi that is not always installed.
+
+Pulseaudio code: add "running" variable to loops to allow stopping the running code when quitting.
+
+Scale screenshot frame libswscale or implement lanczos shader for improved scaline for video as well.
+
+Support high quality scaling with -s by using lanczos.
+
+Support spanning multiple monitors with region capture. This would also allow the user to record multiple monitors at the same time, the same way screen-direct works on nvidia x11.
+
+When webcam support is added also support v4l2loopback? this is done by using avdevice_register_all(); and -c v4l2 -o /dev/video0; but it needs to output raw data as well instead of h264 and possibly yuv420p. Maybe add a -k yuv420p option to do that.
+
+Do proper exit, to call gsr_capture_destroy which will properly stop gsr-kms-server. Otherwise there can be zombie gsr-kms-server on error.
+
+Replace all scissors with clearing textures if the cursor hits the outside of the frame image

include/capture/capture.h

@@ -13,31 +13,39 @@ typedef struct AVMasteringDisplayMetadata AVMasteringDisplayMetadata;
 typedef struct AVContentLightMetadata AVContentLightMetadata;
 typedef struct gsr_capture gsr_capture;
 
+typedef struct {
+    int width;
+    int height;
+    int fps;
+    AVCodecContext *video_codec_context; /* can be NULL */
+    AVFrame *frame; /* can be NULL, but will never be NULL if |video_codec_context| is set */
+} gsr_capture_metadata;
+
 struct gsr_capture {
     /* These methods should not be called manually. Call gsr_capture_* instead */
-    int (*start)(gsr_capture *cap, AVCodecContext *video_codec_context, AVFrame *frame);
+    int (*start)(gsr_capture *cap, gsr_capture_metadata *capture_metadata);
     void (*on_event)(gsr_capture *cap, gsr_egl *egl); /* can be NULL */
     void (*tick)(gsr_capture *cap); /* can be NULL. If there is an event then |on_event| is called before this */
     bool (*should_stop)(gsr_capture *cap, bool *err); /* can be NULL. If NULL, return false */
-    int (*capture)(gsr_capture *cap, AVFrame *frame, gsr_color_conversion *color_conversion);
+    int (*capture)(gsr_capture *cap, gsr_capture_metadata *capture_metadata, gsr_color_conversion *color_conversion);
     bool (*uses_external_image)(gsr_capture *cap); /* can be NULL. If NULL, return false */
     bool (*set_hdr_metadata)(gsr_capture *cap, AVMasteringDisplayMetadata *mastering_display_metadata, AVContentLightMetadata *light_metadata); /* can be NULL. If NULL, return false */
     uint64_t (*get_window_id)(gsr_capture *cap); /* can be NULL. Returns 0 if unknown */
     bool (*is_damaged)(gsr_capture *cap); /* can be NULL */
     void (*clear_damage)(gsr_capture *cap); /* can be NULL */
-    void (*destroy)(gsr_capture *cap, AVCodecContext *video_codec_context);
+    void (*destroy)(gsr_capture *cap);
 
     void *priv; /* can be NULL */
     bool started;
 };
 
-int gsr_capture_start(gsr_capture *cap, AVCodecContext *video_codec_context, AVFrame *frame);
+int gsr_capture_start(gsr_capture *cap, gsr_capture_metadata *capture_metadata);
 void gsr_capture_on_event(gsr_capture *cap, gsr_egl *egl);
 void gsr_capture_tick(gsr_capture *cap);
 bool gsr_capture_should_stop(gsr_capture *cap, bool *err);
-int gsr_capture_capture(gsr_capture *cap, AVFrame *frame, gsr_color_conversion *color_conversion);
+int gsr_capture_capture(gsr_capture *cap, gsr_capture_metadata *capture_metadata, gsr_color_conversion *color_conversion);
 bool gsr_capture_uses_external_image(gsr_capture *cap);
 bool gsr_capture_set_hdr_metadata(gsr_capture *cap, AVMasteringDisplayMetadata *mastering_display_metadata, AVContentLightMetadata *light_metadata);
-void gsr_capture_destroy(gsr_capture *cap, AVCodecContext *video_codec_context);
+void gsr_capture_destroy(gsr_capture *cap);
 
 #endif /* GSR_CAPTURE_CAPTURE_H */

include/capture/kms.h

@@ -5,13 +5,13 @@
 
 typedef struct {
     gsr_egl *egl;
-    const char *display_to_capture; /* if this is "screen", then the first monitor is captured. A copy is made of this */
-    gsr_color_depth color_depth;
-    gsr_color_range color_range;
+    const char *display_to_capture; /* A copy is made of this */
     bool hdr;
     bool record_cursor;
     int fps;
     vec2i output_resolution;
+    vec2i region_size;
+    vec2i region_position;
 } gsr_capture_kms_params;
 
 gsr_capture* gsr_capture_kms_create(const gsr_capture_kms_params *params);

include/capture/nvfbc.h

@@ -8,14 +8,11 @@ typedef struct {
     gsr_egl *egl;
     const char *display_to_capture; /* if this is "screen", then the entire x11 screen is captured (all displays). A copy is made of this */
     int fps;
-    vec2i pos;
-    vec2i size;
     bool direct_capture;
-    gsr_color_depth color_depth;
-    gsr_color_range color_range;
     bool record_cursor;
-    bool use_software_video_encoder;
     vec2i output_resolution;
+    vec2i region_size;
+    vec2i region_position;
 } gsr_capture_nvfbc_params;
 
 gsr_capture* gsr_capture_nvfbc_create(const gsr_capture_nvfbc_params *params);

include/capture/portal.h

@@ -5,8 +5,6 @@
 
 typedef struct {
     gsr_egl *egl;
-    gsr_color_depth color_depth;
-    gsr_color_range color_range;
     bool record_cursor;
     bool restore_portal_session;
     /* If this is set to NULL then this defaults to $XDG_CONFIG_HOME/gpu-screen-recorder/restore_token ($XDG_CONFIG_HOME defaults to $HOME/.config) */

include/capture/xcomposite.h

@@ -8,9 +8,7 @@ typedef struct {
     gsr_egl *egl;
     unsigned long window;
     bool follow_focused; /* If this is set then |window| is ignored */
-    gsr_color_range color_range;
     bool record_cursor;
-    gsr_color_depth color_depth;
     vec2i output_resolution;
 } gsr_capture_xcomposite_params;
 

include/capture/ximage.h

@@ -0,0 +1,18 @@
+#ifndef GSR_CAPTURE_XIMAGE_H
+#define GSR_CAPTURE_XIMAGE_H
+
+#include "capture.h"
+#include "../vec2.h"
+
+typedef struct {
+    gsr_egl *egl;
+    const char *display_to_capture; /* A copy is made of this */
+    bool record_cursor;
+    vec2i output_resolution;
+    vec2i region_size;
+    vec2i region_position;
+} gsr_capture_ximage_params;
+
+gsr_capture* gsr_capture_ximage_create(const gsr_capture_ximage_params *params);
+
+#endif /* GSR_CAPTURE_XIMAGE_H */

include/color_conversion.h

@@ -2,9 +2,13 @@
 #define GSR_COLOR_CONVERSION_H
 
 #include "shader.h"
+#include "defs.h"
 #include "vec2.h"
 #include <stdbool.h>
 
+#define GSR_COLOR_CONVERSION_MAX_SHADERS 6
+#define GSR_COLOR_CONVERSION_MAX_FRAMEBUFFERS 2
+
 typedef enum {
     GSR_COLOR_RANGE_LIMITED,
     GSR_COLOR_RANGE_FULL
@@ -22,12 +26,22 @@ typedef enum {
 
 typedef enum {
     GSR_DESTINATION_COLOR_NV12, /* YUV420, BT709, 8-bit */
-    GSR_DESTINATION_COLOR_P010  /* YUV420, BT2020, 10-bit */
+    GSR_DESTINATION_COLOR_P010, /* YUV420, BT2020, 10-bit */
+    GSR_DESTINATION_COLOR_RGB8
 } gsr_destination_color;
 
+typedef enum {
+    GSR_ROT_0,
+    GSR_ROT_90,
+    GSR_ROT_180,
+    GSR_ROT_270
+} gsr_rotation;
+
 typedef struct {
-    int offset;
-    int rotation;
+    int rotation_matrix;
+    int source_position;
+    int target_position;
+    int scale;
 } gsr_color_uniforms;
 
 typedef struct {
@@ -44,19 +58,23 @@ typedef struct {
 
 typedef struct {
     gsr_color_conversion_params params;
-    gsr_color_uniforms uniforms[4];
-    gsr_shader shaders[4];
+    gsr_color_uniforms uniforms[GSR_COLOR_CONVERSION_MAX_SHADERS];
+    gsr_shader shaders[GSR_COLOR_CONVERSION_MAX_SHADERS];
 
-    unsigned int framebuffers[2];
+    unsigned int framebuffers[GSR_COLOR_CONVERSION_MAX_FRAMEBUFFERS];
 
     unsigned int vertex_array_object_id;
     unsigned int vertex_buffer_object_id;
+
+    int max_local_size_dim;
 } gsr_color_conversion;
 
 int gsr_color_conversion_init(gsr_color_conversion *self, const gsr_color_conversion_params *params);
 void gsr_color_conversion_deinit(gsr_color_conversion *self);
 
-void gsr_color_conversion_draw(gsr_color_conversion *self, unsigned int texture_id, vec2i source_pos, vec2i source_size, vec2i texture_pos, vec2i texture_size, float rotation, bool external_texture, gsr_source_color source_color);
+void gsr_color_conversion_draw(gsr_color_conversion *self, unsigned int texture_id, vec2i destination_pos, vec2i destination_size, vec2i texture_pos, vec2i texture_size, gsr_rotation rotation, bool external_texture, gsr_source_color source_color);
 void gsr_color_conversion_clear(gsr_color_conversion *self);
 
+gsr_rotation gsr_monitor_rotation_to_rotation(gsr_monitor_rotation monitor_rotation);
+
 #endif /* GSR_COLOR_CONVERSION_H */

include/dbus.h

@@ -37,7 +37,11 @@ void gsr_dbus_deinit(gsr_dbus *self);
 /* The follow functions should be called in order to setup ScreenCast properly */
 /* These functions that return an int return the response status code */
 int gsr_dbus_screencast_create_session(gsr_dbus *self, char **session_handle);
-int gsr_dbus_screencast_select_sources(gsr_dbus *self, const char *session_handle, gsr_portal_capture_type capture_type, gsr_portal_cursor_mode cursor_mode);
+/*
+    |capture_type| is a bitmask of gsr_portal_capture_type values. gsr_portal_capture_type values that are not supported by the desktop portal will be ignored.
+    |gsr_portal_cursor_mode| is a bitmask of gsr_portal_cursor_mode values. gsr_portal_cursor_mode values that are not supported will be ignored.
+*/
+int gsr_dbus_screencast_select_sources(gsr_dbus *self, const char *session_handle, uint32_t capture_type, uint32_t cursor_mode);
 int gsr_dbus_screencast_start(gsr_dbus *self, const char *session_handle, uint32_t *pipewire_node);
 bool gsr_dbus_screencast_open_pipewire_remote(gsr_dbus *self, const char *session_handle, int *pipewire_fd);
 const char* gsr_dbus_screencast_get_restore_token(gsr_dbus *self);

include/defs.h

@@ -6,7 +6,8 @@
 typedef enum {
     GSR_GPU_VENDOR_AMD,
     GSR_GPU_VENDOR_INTEL,
-    GSR_GPU_VENDOR_NVIDIA
+    GSR_GPU_VENDOR_NVIDIA,
+    GSR_GPU_VENDOR_BROADCOM
 } gsr_gpu_vendor;
 
 typedef struct {

include/egl.h

@@ -98,17 +98,20 @@ typedef void(*__GLXextFuncPtr)(void);
 #define GL_TEXTURE_EXTERNAL_OES                 0x8D65
 #define GL_RED                                  0x1903
 #define GL_GREEN                                0x1904
-#define GL_BLUE                                    0x1905
+#define GL_BLUE                                 0x1905
 #define GL_ALPHA                                0x1906
 #define GL_TEXTURE_SWIZZLE_RGBA                 0x8E46
 #define GL_RG                                   0x8227
 #define GL_RGB                                  0x1907
 #define GL_RGBA                                 0x1908
+#define GL_RGB8                                 0x8051
 #define GL_RGBA8                                0x8058
 #define GL_R8                                   0x8229
 #define GL_RG8                                  0x822B
 #define GL_R16                                  0x822A
 #define GL_RG16                                 0x822C
+#define GL_RGB16                                0x8054
+#define GL_RGBA32F                              0x8814
 #define GL_UNSIGNED_BYTE                        0x1401
 #define GL_COLOR_BUFFER_BIT                     0x00004000
 #define GL_TEXTURE_WRAP_S                       0x2802
@@ -130,6 +133,18 @@ typedef void(*__GLXextFuncPtr)(void);
 #define GL_ONE_MINUS_SRC_ALPHA                  0x0303
 #define GL_DEBUG_OUTPUT                         0x92E0
 #define GL_SCISSOR_TEST                         0x0C11
+#define GL_PACK_ALIGNMENT                       0x0D05
+#define GL_UNPACK_ALIGNMENT                     0x0CF5
+#define GL_READ_ONLY                            0x88B8
+#define GL_WRITE_ONLY                           0x88B9
+#define GL_READ_WRITE                           0x88BA
+#define GL_MAX_COMPUTE_FIXED_GROUP_INVOCATIONS  0x90EB
+#define GL_TEXTURE0                             0x84C0
+#define GL_TEXTURE1                             0x84C1
+#define GL_CLAMP_TO_BORDER                      0x812D
+#define GL_TEXTURE_BORDER_COLOR                 0x1004
+#define GL_SHADER_IMAGE_ACCESS_BARRIER_BIT      0x00000020
+#define GL_ALL_BARRIER_BITS                     0xFFFFFFFF
 
 #define GL_VENDOR                               0x1F00
 #define GL_RENDERER                             0x1F01
@@ -139,6 +154,7 @@ typedef void(*__GLXextFuncPtr)(void);
 #define GL_INFO_LOG_LENGTH                      0x8B84
 #define GL_FRAGMENT_SHADER                      0x8B30
 #define GL_VERTEX_SHADER                        0x8B31
+#define GL_COMPUTE_SHADER                       0x91B9
 #define GL_COMPILE_STATUS                       0x8B81
 #define GL_LINK_STATUS                          0x8B82
 
@@ -226,15 +242,21 @@ struct gsr_egl {
     void (*glClearColor)(float red, float green, float blue, float alpha);
     void (*glGenTextures)(int n, unsigned int *textures);
     void (*glDeleteTextures)(int n, const unsigned int *texture);
+    void (*glActiveTexture)(unsigned int texture);
     void (*glBindTexture)(unsigned int target, unsigned int texture);
+    void (*glBindImageTexture)(unsigned int unit, unsigned int texture, int level, unsigned char layered, int layer, unsigned int access, unsigned int format);
     void (*glTexParameteri)(unsigned int target, unsigned int pname, int param);
     void (*glTexParameteriv)(unsigned int target, unsigned int pname, const int *params);
+    void (*glTexParameterfv)(unsigned int target, unsigned int pname, const float *params);
     void (*glGetTexLevelParameteriv)(unsigned int target, int level, unsigned int pname, int *params);
     void (*glTexImage2D)(unsigned int target, int level, int internalFormat, int width, int height, int border, unsigned int format, unsigned int type, const void *pixels);
+    void (*glTexSubImage2D)(unsigned int target, int level, int xoffset, int yoffset, int width, int height, unsigned format, unsigned type, const void *pixels);
     void (*glGetTexImage)(unsigned int target, int level, unsigned int format, unsigned int type, void *pixels);
     void (*glGenFramebuffers)(int n, unsigned int *framebuffers);
     void (*glBindFramebuffer)(unsigned int target, unsigned int framebuffer);
     void (*glDeleteFramebuffers)(int n, const unsigned int *framebuffers);
+    void (*glDispatchCompute)(unsigned int num_groups_x, unsigned int num_groups_y, unsigned int num_groups_z);
+    void (*glMemoryBarrier)(unsigned int barriers);
     void (*glViewport)(int x, int y, int width, int height);
     void (*glFramebufferTexture2D)(unsigned int target, unsigned int attachment, unsigned int textarget, unsigned int texture, int level);
     void (*glDrawBuffers)(int n, const unsigned int *bufs);
@@ -267,14 +289,19 @@ struct gsr_egl {
     void (*glEnable)(unsigned int cap);
     void (*glDisable)(unsigned int cap);
     void (*glBlendFunc)(unsigned int sfactor, unsigned int dfactor);
+    void (*glPixelStorei)(unsigned int pname, int param);
     int (*glGetUniformLocation)(unsigned int program, const char *name);
     void (*glUniform1f)(int location, float v0);
     void (*glUniform2f)(int location, float v0, float v1);
+    void (*glUniform1i)(int location, int v0);
+    void (*glUniform2i)(int location, int v0, int v1);
+    void (*glUniformMatrix2fv)(int location, int count, unsigned char transpose, const float *value);
     void (*glDebugMessageCallback)(GLDEBUGPROC callback, const void *userParam);
     void (*glScissor)(int x, int y, int width, int height);
     void (*glReadPixels)(int x, int y, int width, int height, unsigned int format, unsigned int type, void *pixels);
     void* (*glMapBuffer)(unsigned int target, unsigned int access);
     unsigned char (*glUnmapBuffer)(unsigned int target);
+    void (*glGetIntegerv)(unsigned int pname, int *params);
 };
 
 bool gsr_egl_load(gsr_egl *self, gsr_window *window, bool is_monitor_capture, bool enable_debug);

include/image_writer.h

@@ -0,0 +1,35 @@
+#ifndef GSR_IMAGE_WRITER_H
+#define GSR_IMAGE_WRITER_H
+
+#include <stdbool.h>
+
+typedef struct gsr_egl gsr_egl;
+
+typedef enum {
+    GSR_IMAGE_FORMAT_JPEG,
+    GSR_IMAGE_FORMAT_PNG
+} gsr_image_format;
+
+typedef enum {
+    GSR_IMAGE_WRITER_SOURCE_OPENGL,
+    GSR_IMAGE_WRITER_SOURCE_MEMORY
+} gsr_image_writer_source;
+
+typedef struct {
+    gsr_image_writer_source source;
+    gsr_egl *egl;
+    int width;
+    int height;
+    unsigned int texture;
+    const void *memory; /* Reference */
+} gsr_image_writer;
+
+bool gsr_image_writer_init_opengl(gsr_image_writer *self, gsr_egl *egl, int width, int height);
+/* |memory| is taken as a reference. The data is expected to be in rgba8 format (8 bit rgba) */
+bool gsr_image_writer_init_memory(gsr_image_writer *self, const void *memory, int width, int height);
+void gsr_image_writer_deinit(gsr_image_writer *self);
+
+/* Quality is between 1 and 100 where 100 is the max quality. Quality doesn't apply to lossless formats */
+bool gsr_image_writer_write_to_file(gsr_image_writer *self, const char *filepath, gsr_image_format image_format, int quality);
+
+#endif /* GSR_IMAGE_WRITER_H */

include/pipewire_audio.h

@@ -10,6 +10,7 @@
 
 #define GSR_PIPEWIRE_AUDIO_MAX_STREAM_NODES 128
 #define GSR_PIPEWIRE_AUDIO_MAX_PORTS 256
+#define GSR_PIPEWIRE_AUDIO_MAX_LINKS 256
 #define GSR_PIPEWIRE_AUDIO_MAX_REQUESTED_LINKS 32
 #define GSR_PIPEWIRE_AUDIO_MAX_VIRTUAL_SINKS 32
 
@@ -37,14 +38,31 @@ typedef struct {
     char *name;
 } gsr_pipewire_audio_port;
 
+typedef struct {
+    uint32_t id;
+    uint32_t output_node_id;
+    uint32_t input_node_id;
+} gsr_pipewire_audio_link;
+
 typedef enum {
     GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM, /* Application */
     GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK    /* Combined (virtual) sink */
 } gsr_pipewire_audio_link_input_type;
 
+typedef enum {
+    GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_STANDARD,
+    GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT,
+    GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_INPUT
+} gsr_pipewire_audio_requested_type;
+
 typedef struct {
-    char **output_names;
-    int num_output_names;
+    char *name;
+    gsr_pipewire_audio_requested_type type;
+} gsr_pipewire_audio_requested_output;
+
+typedef struct {
+    gsr_pipewire_audio_requested_output *outputs;
+    int num_outputs;
     char *input_name;
     bool inverted;
     gsr_pipewire_audio_node_type output_type;
@@ -60,12 +78,21 @@ typedef struct {
     struct spa_hook registry_listener;
     int server_version_sync;
 
+    struct pw_proxy *metadata_proxy;
+    struct spa_hook metadata_listener;
+    struct spa_hook metadata_proxy_listener;
+    char default_output_device_name[128];
+    char default_input_device_name[128];
+
     gsr_pipewire_audio_node stream_nodes[GSR_PIPEWIRE_AUDIO_MAX_STREAM_NODES];
     int num_stream_nodes;
 
     gsr_pipewire_audio_port ports[GSR_PIPEWIRE_AUDIO_MAX_PORTS];
     int num_ports;
 
+    gsr_pipewire_audio_link links[GSR_PIPEWIRE_AUDIO_MAX_LINKS];
+    int num_links;
+
     gsr_pipewire_audio_requested_link requested_links[GSR_PIPEWIRE_AUDIO_MAX_REQUESTED_LINKS];
     int num_requested_links;
 
@@ -118,6 +145,8 @@ bool gsr_pipewire_audio_add_link_from_apps_to_sink_inverted(gsr_pipewire_audio *
     If a device or a new device starts outputting audio after this function is called and the device name matches
     then it will automatically link the audio sources.
     |source_names| and |sink_name_input| are case-insensitive matches.
+    |source_names| can include "default_output" or "default_input" to use the default output/input
+    and it will automatically switch when the default output/input is changed in system audio settings.
 */
 bool gsr_pipewire_audio_add_link_from_sources_to_sink(gsr_pipewire_audio *self, const char **source_names, int num_source_names, const char *sink_name_input);
 

include/pipewire_video.h

@@ -9,7 +9,7 @@
 #include <spa/param/video/format.h>
 
 #define GSR_PIPEWIRE_VIDEO_MAX_MODIFIERS 1024
-#define GSR_PIPEWIRE_VIDEO_NUM_VIDEO_FORMATS 6
+#define GSR_PIPEWIRE_VIDEO_MAX_VIDEO_FORMATS 12
 #define GSR_PIPEWIRE_VIDEO_DMABUF_MAX_PLANES 4
 
 typedef struct gsr_egl gsr_egl;
@@ -82,7 +82,7 @@ typedef struct {
         uint32_t width, height;
     } crop;
 
-    gsr_video_format supported_video_formats[GSR_PIPEWIRE_VIDEO_NUM_VIDEO_FORMATS];
+    gsr_video_format supported_video_formats[GSR_PIPEWIRE_VIDEO_MAX_VIDEO_FORMATS];
 
     gsr_pipewire_video_data_version server_version;
     gsr_pipewire_video_video_info video_info;

include/shader.h

@@ -9,7 +9,7 @@ typedef struct {
 } gsr_shader;
 
 /* |vertex_shader| or |fragment_shader| may be NULL */
-int gsr_shader_init(gsr_shader *self, gsr_egl *egl, const char *vertex_shader, const char *fragment_shader);
+int gsr_shader_init(gsr_shader *self, gsr_egl *egl, const char *vertex_shader, const char *fragment_shader, const char *compute_shader);
 void gsr_shader_deinit(gsr_shader *self);
 
 int gsr_shader_bind_attribute_location(gsr_shader *self, const char *attribute, int location);

include/sound.hpp

@@ -61,6 +61,9 @@ typedef enum {
 
 /*
     Get a sound device by name, returning the device into the |device| parameter.
+    |device_name| can be a device name or "default_output" or "default_input".
+    If the device name is "default_output" or "default_input" then it will automatically switch which
+    device is records from when the default output/input is changed in the system audio settings.
     Returns 0 on success, or a negative value on failure.
 */
 int sound_device_get_by_name(SoundDevice *device, const char *device_name, const char *description, unsigned int num_channels, unsigned int period_frame_size, AudioFormat audio_format);

include/utils.h

@@ -15,7 +15,7 @@ typedef struct AVFrame AVFrame;
 typedef struct {
     const char *name;
     int name_len;
-    vec2i pos;
+    vec2i pos; /* This is 0, 0 on wayland. Use |drm_monitor_get_display_server_data| to get the position */
     vec2i size;
     uint32_t connector_id; /* Only on x11 and drm */
     gsr_monitor_rotation rotation; /* Only on x11 and wayland */
@@ -43,14 +43,13 @@ typedef void (*active_monitor_callback)(const gsr_monitor *monitor, void *userda
 void for_each_active_monitor_output_x11_not_cached(Display *display, active_monitor_callback callback, void *userdata);
 void for_each_active_monitor_output(const gsr_window *window, const char *card_path, gsr_connection_type connection_type, active_monitor_callback callback, void *userdata);
 bool get_monitor_by_name(const gsr_egl *egl, gsr_connection_type connection_type, const char *name, gsr_monitor *monitor);
-gsr_monitor_rotation drm_monitor_get_display_server_rotation(const gsr_window *window, const gsr_monitor *monitor);
+bool drm_monitor_get_display_server_data(const gsr_window *window, const gsr_monitor *monitor, gsr_monitor_rotation *monitor_rotation, vec2i *monitor_position);
 
 int get_connector_type_by_name(const char *name);
 drm_connector_type_count* drm_connector_types_get_index(drm_connector_type_count *type_counts, int *num_type_counts, int connector_type);
 uint32_t monitor_identifier_from_type_and_count(int monitor_type_index, int monitor_type_count);
 
 bool gl_get_gpu_info(gsr_egl *egl, gsr_gpu_info *info);
-bool gl_driver_version_greater_than(const gsr_gpu_info *gpu_info, int major, int minor, int patch);
 
 bool try_card_has_valid_plane(const char *card_path);
 /* |output| should be at least 128 bytes in size */
@@ -63,9 +62,9 @@ int create_directory_recursive(char *path);
 /* |img_attr| needs to be at least 44 in size */
 void setup_dma_buf_attrs(intptr_t *img_attr, uint32_t format, uint32_t width, uint32_t height, const int *fds, const uint32_t *offsets, const uint32_t *pitches, const uint64_t *modifiers, int num_planes, bool use_modifier);
 bool video_codec_context_is_vaapi(AVCodecContext *video_codec_context);
-bool vaapi_copy_drm_planes_to_video_surface(AVCodecContext *video_codec_context, AVFrame *video_frame, vec2i source_pos, vec2i source_size, vec2i dest_pos, vec2i dest_size, uint32_t format, vec2i size, const int *fds, const uint32_t *offsets, const uint32_t *pitches, const uint64_t *modifiers, int num_planes);
-bool vaapi_copy_egl_image_to_video_surface(gsr_egl *egl, EGLImage image, vec2i source_pos, vec2i source_size, vec2i dest_pos, vec2i dest_size, AVCodecContext *video_codec_context, AVFrame *video_frame);
 
 vec2i scale_keep_aspect_ratio(vec2i from, vec2i to);
 
+unsigned int gl_create_texture(gsr_egl *egl, int width, int height, int internal_format, unsigned int format, int filter);
+
 #endif /* GSR_UTILS_H */

kms/server/kms_server.c

@@ -25,7 +25,6 @@
 
 typedef struct {
     int drmfd;
-    drmModePlaneResPtr planes;
 } gsr_drm;
 
 typedef struct {
@@ -289,21 +288,31 @@ static int drm_prime_handles_to_fds(gsr_drm *drm, drmModeFB2Ptr drmfb, int *fb_f
     return GSR_KMS_MAX_DMA_BUFS;
 }
 
-static int kms_get_fb(gsr_drm *drm, gsr_kms_response *response, connector_to_crtc_map *c2crtc_map) {
+static int kms_get_fb(gsr_drm *drm, gsr_kms_response *response) {
     int result = -1;
 
     response->result = KMS_RESULT_OK;
     response->err_msg[0] = '\0';
     response->num_items = 0;
 
-    for(uint32_t i = 0; i < drm->planes->count_planes && response->num_items < GSR_KMS_MAX_ITEMS; ++i) {
+    connector_to_crtc_map c2crtc_map;
+    c2crtc_map.num_maps = 0;
+    map_crtc_to_connector_ids(drm, &c2crtc_map);
+
+    drmModePlaneResPtr planes = drmModeGetPlaneResources(drm->drmfd);
+    if(!planes) {
+        fprintf(stderr, "kms server error: failed to get plane resources, error: %s\n", strerror(errno));
+        goto done;
+    }
+
+    for(uint32_t i = 0; i < planes->count_planes && response->num_items < GSR_KMS_MAX_ITEMS; ++i) {
         drmModePlanePtr plane = NULL;
         drmModeFB2Ptr drmfb = NULL;
 
-        plane = drmModeGetPlane(drm->drmfd, drm->planes->planes[i]);
+        plane = drmModeGetPlane(drm->drmfd, planes->planes[i]);
         if(!plane) {
             response->result = KMS_RESULT_FAILED_TO_GET_PLANE;
-            snprintf(response->err_msg, sizeof(response->err_msg), "failed to get drm plane with id %u, error: %s\n", drm->planes->planes[i], strerror(errno));
+            snprintf(response->err_msg, sizeof(response->err_msg), "failed to get drm plane with id %u, error: %s\n", planes->planes[i], strerror(errno));
             fprintf(stderr, "kms server error: %s\n", response->err_msg);
             goto next;
         }
@@ -346,7 +355,7 @@ static int kms_get_fb(gsr_drm *drm, gsr_kms_response *response, connector_to_crt
 
         const int item_index = response->num_items;
 
-        const connector_crtc_pair *crtc_pair = get_connector_pair_by_crtc_id(c2crtc_map, plane->crtc_id);
+        const connector_crtc_pair *crtc_pair = get_connector_pair_by_crtc_id(&c2crtc_map, plane->crtc_id);
         if(crtc_pair && crtc_pair->hdr_metadata_blob_id) {
             response->items[item_index].has_hdr_metadata = get_hdr_metadata(drm->drmfd, crtc_pair->hdr_metadata_blob_id, &response->items[item_index].hdr_metadata);
         } else {
@@ -389,6 +398,11 @@ static int kms_get_fb(gsr_drm *drm, gsr_kms_response *response, connector_to_crt
             drmModeFreePlane(plane);
     }
 
+    done:
+
+    if(planes)
+        drmModeFreePlaneResources(planes);
+
     if(response->num_items > 0)
         response->result = KMS_RESULT_OK;
 
@@ -499,7 +513,6 @@ int main(int argc, char **argv) {
     int socket_fd = 0;
     gsr_drm drm;
     drm.drmfd = 0;
-    drm.planes = NULL;
 
     if(argc != 3) {
         fprintf(stderr, "usage: gsr-kms-server <domain_socket_path> <card_path>\n");
@@ -532,17 +545,6 @@ int main(int argc, char **argv) {
         fprintf(stderr, "kms server warning: drmSetClientCap DRM_CLIENT_CAP_ATOMIC failed, error: %s. The wrong monitor may be captured as a result\n", strerror(errno));
     }
 
-    drm.planes = drmModeGetPlaneResources(drm.drmfd);
-    if(!drm.planes) {
-        fprintf(stderr, "kms server error: failed to get plane resources, error: %s\n", strerror(errno));
-        res = 2;
-        goto done;
-    }
-
-    connector_to_crtc_map c2crtc_map;
-    c2crtc_map.num_maps = 0;
-    map_crtc_to_connector_ids(&drm, &c2crtc_map);
-
     fprintf(stderr, "kms server info: connecting to the client\n");
     bool connected = false;
     const double connect_timeout_sec = 5.0;
@@ -642,7 +644,7 @@ int main(int argc, char **argv) {
                 response.version = GSR_KMS_PROTOCOL_VERSION;
                 response.num_items = 0;
                 
-                if(kms_get_fb(&drm, &response, &c2crtc_map) == 0) {
+                if(kms_get_fb(&drm, &response) == 0) {
                     if(send_msg_to_client(socket_fd, &response) == -1)
                         fprintf(stderr, "kms server error: failed to respond to client KMS_REQUEST_TYPE_GET_KMS request\n");
                 } else {
@@ -681,8 +683,6 @@ int main(int argc, char **argv) {
     }
 
     done:
-    if(drm.planes)
-        drmModeFreePlaneResources(drm.planes);
     if(drm.drmfd > 0)
         close(drm.drmfd);
     if(socket_fd > 0)

meson.build

@@ -1,4 +1,4 @@
-project('gpu-screen-recorder', ['c', 'cpp'], version : '5.0.2', default_options : ['warning_level=2'])
+project('gpu-screen-recorder', ['c', 'cpp'], version : '5.3.4', default_options : ['warning_level=2'])
 
 add_project_arguments('-Wshadow', language : ['c', 'cpp'])
 if get_option('buildtype') == 'debug'
@@ -12,6 +12,7 @@ src = [
     'src/capture/capture.c',
     'src/capture/nvfbc.c',
     'src/capture/xcomposite.c',
+    'src/capture/ximage.c',
     'src/capture/kms.c',
     'src/encoder/video/video.c',
     'src/encoder/video/nvenc.c',
@@ -22,8 +23,8 @@ src = [
     'src/codec_query/vaapi.c',
     'src/codec_query/vulkan.c',
     'src/window/window.c',
-    'src/window/window_x11.c',
-    'src/window/window_wayland.c',
+    'src/window/x11.c',
+    'src/window/wayland.c',
     'src/egl.c',
     'src/cuda.c',
     'src/xnvctrl.c',
@@ -35,6 +36,7 @@ src = [
     'src/library_loader.c',
     'src/cursor.c',
     'src/damage.c',
+    'src/image_writer.c',
     'src/sound.cpp',
     'src/main.cpp',
 ]

meson_options.txt

@@ -1,5 +1,5 @@
 option('systemd', type : 'boolean', value : true, description : 'Install systemd service file')
 option('capabilities', type : 'boolean', value : true, description : 'Set binary admin capability on gsr-kms-server binary to remove password prompt when recording monitor (without desktop portal option) on amd/intel or nvidia wayland')
 option('nvidia_suspend_fix', type : 'boolean', value : true, description : 'Install nvidia modprobe config file to tell nvidia driver to preserve video memory on suspend. This is a workaround for an nvidia driver bug that breaks cuda (and gpu screen recorder) on suspend')
-option('portal', type : 'boolean', value : true, description : 'Build with support for xdg desktop portal ScreenCast capture (wayland only) (-w portal option)')
-option('app_audio', type : 'boolean', value : true, description : 'Build with support for recording a single audio source (-aa option). Requires pipewire')
+option('portal', type : 'boolean', value : true, description : 'Build with support for xdg desktop portal ScreenCast capture (wayland only) (-w portal option). Requires pipewire')
+option('app_audio', type : 'boolean', value : true, description : 'Build with support for recording a single audio source (-a app: option). Requires pipewire')

project.conf

@@ -1,7 +1,7 @@
 [package]
 name = "gpu-screen-recorder"
 type = "executable"
-version = "5.0.2"
+version = "5.3.4"
 platforms = ["posix"]
 
 [config]

scripts/replay-application-name.sh

@@ -3,4 +3,4 @@
 window=$(xdotool selectwindow)
 window_name=$(xdotool getwindowname "$window" || xdotool getwindowclassname "$window" || echo "Game")
 window_name="$(echo "$window_name" | tr '/\\' '_')"
-gpu-screen-recorder -w "$window" -f 60 -c mkv -a default_output -r 60 -o "$HOME/Videos/Replays/$window_name"
+gpu-screen-recorder -w "$window" -f 60 -c mkv -a default_output -bm cbr -q 45000 -r 60 -o "$HOME/Videos/Replays/$window_name"

scripts/start-replay.sh

@@ -3,4 +3,4 @@
 pidof -q gpu-screen-recorder && exit 0
 video_path="$HOME/Videos"
 mkdir -p "$video_path"
-gpu-screen-recorder -w screen -f 60 -a default_output -c mkv -r 30 -o "$video_path"
+gpu-screen-recorder -w screen -f 60 -a default_output -c mkv -bm cbr -q 45000 -r 30 -o "$video_path"

src/capture/capture.c

@@ -1,9 +1,9 @@
 #include "../../include/capture/capture.h"
 #include <assert.h>
 
-int gsr_capture_start(gsr_capture *cap, AVCodecContext *video_codec_context, AVFrame *frame) {
+int gsr_capture_start(gsr_capture *cap, gsr_capture_metadata *capture_metadata) {
     assert(!cap->started);
-    int res = cap->start(cap, video_codec_context, frame);
+    int res = cap->start(cap, capture_metadata);
     if(res == 0)
         cap->started = true;
 
@@ -29,9 +29,9 @@ bool gsr_capture_should_stop(gsr_capture *cap, bool *err) {
         return false;
 }
 
-int gsr_capture_capture(gsr_capture *cap, AVFrame *frame, gsr_color_conversion *color_conversion) {
+int gsr_capture_capture(gsr_capture *cap, gsr_capture_metadata *capture_metadata, gsr_color_conversion *color_conversion) {
     assert(cap->started);
-    return cap->capture(cap, frame, color_conversion);
+    return cap->capture(cap, capture_metadata, color_conversion);
 }
 
 bool gsr_capture_uses_external_image(gsr_capture *cap) {
@@ -48,6 +48,6 @@ bool gsr_capture_set_hdr_metadata(gsr_capture *cap, AVMasteringDisplayMetadata *
         return false;
 }
 
-void gsr_capture_destroy(gsr_capture *cap, AVCodecContext *video_codec_context) {
-    cap->destroy(cap, video_codec_context);
+void gsr_capture_destroy(gsr_capture *cap) {
+    cap->destroy(cap);
 }

src/capture/kms.c

@@ -14,9 +14,7 @@
 #include <xf86drm.h>
 #include <libdrm/drm_fourcc.h>
 
-#include <libavcodec/avcodec.h>
 #include <libavutil/mastering_display_metadata.h>
-#include <libavformat/avformat.h>
 
 #define FIND_CRTC_BY_NAME_TIMEOUT_SECONDS 2.0
 
@@ -55,11 +53,6 @@ typedef struct {
     bool is_x11;
     gsr_cursor x11_cursor;
 
-    AVCodecContext *video_codec_context;
-    bool performance_error_shown;
-    bool fast_path_failed;
-    bool mesa_supports_compute_only_vaapi_copy;
-
     //int drm_fd;
     //uint64_t prev_sequence;
     //bool damaged;
@@ -115,18 +108,22 @@ static int max_int(int a, int b) {
 }
 
 static void gsr_capture_kms_create_input_texture_ids(gsr_capture_kms *self) {
+    const float border_color[4] = { 0.0f, 0.0f, 0.0f, 0.0f };
+
     self->params.egl->glGenTextures(1, &self->input_texture_id);
     self->params.egl->glBindTexture(GL_TEXTURE_2D, self->input_texture_id);
-    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, border_color);
     self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
     self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->params.egl->glBindTexture(GL_TEXTURE_2D, 0);
 
     self->params.egl->glGenTextures(1, &self->external_input_texture_id);
     self->params.egl->glBindTexture(GL_TEXTURE_EXTERNAL_OES, self->external_input_texture_id);
-    self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameterfv(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_BORDER_COLOR, border_color);
     self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
     self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->params.egl->glBindTexture(GL_TEXTURE_EXTERNAL_OES, 0);
@@ -136,8 +133,9 @@ static void gsr_capture_kms_create_input_texture_ids(gsr_capture_kms *self) {
 
     self->params.egl->glGenTextures(1, &self->cursor_texture_id);
     self->params.egl->glBindTexture(cursor_texture_id_target, self->cursor_texture_id);
-    self->params.egl->glTexParameteri(cursor_texture_id_target, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->params.egl->glTexParameteri(cursor_texture_id_target, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    self->params.egl->glTexParameteri(cursor_texture_id_target, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameteri(cursor_texture_id_target, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameterfv(cursor_texture_id_target, GL_TEXTURE_BORDER_COLOR, border_color);
     self->params.egl->glTexParameteri(cursor_texture_id_target, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
     self->params.egl->glTexParameteri(cursor_texture_id_target, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->params.egl->glBindTexture(cursor_texture_id_target, 0);
@@ -177,7 +175,7 @@ static vec2i rotate_capture_size_if_rotated(gsr_capture_kms *self, vec2i capture
     return capture_size;
 }
 
-static int gsr_capture_kms_start(gsr_capture *cap, AVCodecContext *video_codec_context, AVFrame *frame) {
+static int gsr_capture_kms_start(gsr_capture *cap, gsr_capture_metadata *capture_metadata) {
     gsr_capture_kms *self = cap->priv;
 
     gsr_capture_kms_create_input_texture_ids(self);
@@ -210,7 +208,8 @@ static int gsr_capture_kms_start(gsr_capture *cap, AVCodecContext *video_codec_c
     }
 
     monitor.name = self->params.display_to_capture;
-    self->monitor_rotation = drm_monitor_get_display_server_rotation(self->params.egl->window, &monitor);
+    vec2i monitor_position = {0, 0};
+    drm_monitor_get_display_server_data(self->params.egl->window, &monitor, &self->monitor_rotation, &monitor_position);
 
     self->capture_pos = monitor.pos;
     /* Monitor size is already rotated on x11 when the monitor is rotated, no need to apply it ourselves */
@@ -219,29 +218,18 @@ static int gsr_capture_kms_start(gsr_capture *cap, AVCodecContext *video_codec_c
     else
         self->capture_size = rotate_capture_size_if_rotated(self, monitor.size);
 
-    /* Disable vsync */
-    self->params.egl->eglSwapInterval(self->params.egl->egl_display, 0);
-
-    if(self->params.output_resolution.x == 0 && self->params.output_resolution.y == 0) {
-        self->params.output_resolution = self->capture_size;
-        video_codec_context->width = FFALIGN(self->capture_size.x, 2);
-        video_codec_context->height = FFALIGN(self->capture_size.y, 2);
-    } else {
+    if(self->params.output_resolution.x > 0 && self->params.output_resolution.y > 0) {
         self->params.output_resolution = scale_keep_aspect_ratio(self->capture_size, self->params.output_resolution);
-        video_codec_context->width = FFALIGN(self->params.output_resolution.x, 2);
-        video_codec_context->height = FFALIGN(self->params.output_resolution.y, 2);
+        capture_metadata->width = self->params.output_resolution.x;
+        capture_metadata->height = self->params.output_resolution.y;
+    } else if(self->params.region_size.x > 0 && self->params.region_size.y > 0) {
+        capture_metadata->width = self->params.region_size.x;
+        capture_metadata->height = self->params.region_size.y;
+    } else {
+        capture_metadata->width = self->capture_size.x;
+        capture_metadata->height = self->capture_size.y;
     }
 
-    self->fast_path_failed = self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD && !gl_driver_version_greater_than(&self->params.egl->gpu_info, 24, 0, 9);
-    if(self->fast_path_failed)
-        fprintf(stderr, "gsr warning: gsr_capture_kms_start: your amd driver (mesa) version is known to be buggy (<= version 24.0.9), falling back to opengl copy\n");
-
-    self->mesa_supports_compute_only_vaapi_copy = self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD && gl_driver_version_greater_than(&self->params.egl->gpu_info, 24, 3, 6);
-
-    frame->width = video_codec_context->width;
-    frame->height = video_codec_context->height;
-
-    self->video_codec_context = video_codec_context;
     self->last_time_monitor_check = clock_get_monotonic_seconds();
     return 0;
 }
@@ -276,16 +264,6 @@ static void gsr_capture_kms_on_event(gsr_capture *cap, gsr_egl *egl) {
 //     }
 // }
 
-static float monitor_rotation_to_radians(gsr_monitor_rotation rot) {
-    switch(rot) {
-        case GSR_MONITOR_ROT_0:   return 0.0f;
-        case GSR_MONITOR_ROT_90:  return M_PI_2;
-        case GSR_MONITOR_ROT_180: return M_PI;
-        case GSR_MONITOR_ROT_270: return M_PI + M_PI_2;
-    }
-    return 0.0f;
-}
-
 static gsr_kms_response_item* find_drm_by_connector_id(gsr_kms_response *kms_response, uint32_t connector_id) {
     for(int i = 0; i < kms_response->num_items; ++i) {
         if(kms_response->items[i].connector_id == connector_id && !kms_response->items[i].is_cursor)
@@ -451,7 +429,7 @@ static gsr_kms_response_item* find_cursor_drm_if_on_monitor(gsr_capture_kms *sel
     return cursor_drm_fd;
 }
 
-static void render_drm_cursor(gsr_capture_kms *self, gsr_color_conversion *color_conversion, const gsr_kms_response_item *cursor_drm_fd, vec2i target_pos, float texture_rotation, vec2i output_size) {
+static void render_drm_cursor(gsr_capture_kms *self, gsr_color_conversion *color_conversion, const gsr_kms_response_item *cursor_drm_fd, vec2i target_pos, vec2i output_size, vec2i framebuffer_size) {
     const vec2d scale = {
         self->capture_size.x == 0 ? 0 : (double)output_size.x / (double)self->capture_size.x,
         self->capture_size.y == 0 ? 0 : (double)output_size.y / (double)self->capture_size.y
@@ -466,25 +444,28 @@ static void render_drm_cursor(gsr_capture_kms *self, gsr_color_conversion *color
             break;
         case GSR_MONITOR_ROT_90:
             cursor_pos = swap_vec2i(cursor_pos);
-            cursor_pos.x = self->capture_size.x - cursor_pos.x;
+            cursor_pos.x = framebuffer_size.x - cursor_pos.x;
             // TODO: Remove this horrible hack
             cursor_pos.x -= cursor_size.x;
             break;
         case GSR_MONITOR_ROT_180:
-            cursor_pos.x = self->capture_size.x - cursor_pos.x;
-            cursor_pos.y = self->capture_size.y - cursor_pos.y;
+            cursor_pos.x = framebuffer_size.x - cursor_pos.x;
+            cursor_pos.y = framebuffer_size.y - cursor_pos.y;
             // TODO: Remove this horrible hack
             cursor_pos.x -= cursor_size.x;
             cursor_pos.y -= cursor_size.y;
             break;
         case GSR_MONITOR_ROT_270:
             cursor_pos = swap_vec2i(cursor_pos);
-            cursor_pos.y = self->capture_size.y - cursor_pos.y;
+            cursor_pos.y = framebuffer_size.y - cursor_pos.y;
             // TODO: Remove this horrible hack
             cursor_pos.y -= cursor_size.y;
             break;
     }
 
+    cursor_pos.x -= self->params.region_position.x;
+    cursor_pos.y -= self->params.region_position.y;
+
     cursor_pos.x *= scale.x;
     cursor_pos.y *= scale.y;
 
@@ -522,7 +503,7 @@ static void render_drm_cursor(gsr_capture_kms *self, gsr_color_conversion *color
     gsr_color_conversion_draw(color_conversion, self->cursor_texture_id,
         cursor_pos, (vec2i){cursor_size.x * scale.x, cursor_size.y * scale.y},
         (vec2i){0, 0}, cursor_size,
-        texture_rotation, cursor_texture_id_is_external, GSR_SOURCE_COLOR_RGB);
+        gsr_monitor_rotation_to_rotation(self->monitor_rotation), cursor_texture_id_is_external, GSR_SOURCE_COLOR_RGB);
 
     self->params.egl->glDisable(GL_SCISSOR_TEST);
 }
@@ -550,7 +531,7 @@ static void render_x11_cursor(gsr_capture_kms *self, gsr_color_conversion *color
     gsr_color_conversion_draw(color_conversion, self->x11_cursor.texture_id,
         cursor_pos, (vec2i){self->x11_cursor.size.x * scale.x, self->x11_cursor.size.y * scale.y},
         (vec2i){0, 0}, self->x11_cursor.size,
-        0.0f, false, GSR_SOURCE_COLOR_RGB);
+        GSR_ROT_0, false, GSR_SOURCE_COLOR_RGB);
 
     self->params.egl->glDisable(GL_SCISSOR_TEST);
 }
@@ -592,7 +573,8 @@ static void gsr_capture_kms_update_connector_ids(gsr_capture_kms *self) {
     self->monitor_id.connector_ids[0] = monitor.connector_id;
 
     monitor.name = self->params.display_to_capture;
-    self->monitor_rotation = drm_monitor_get_display_server_rotation(self->params.egl->window, &monitor);
+    vec2i monitor_position = {0, 0};
+    drm_monitor_get_display_server_data(self->params.egl->window, &monitor, &self->monitor_rotation, &monitor_position);
 
     self->capture_pos = monitor.pos;
     /* Monitor size is already rotated on x11 when the monitor is rotated, no need to apply it ourselves */
@@ -602,17 +584,7 @@ static void gsr_capture_kms_update_connector_ids(gsr_capture_kms *self) {
         self->capture_size = rotate_capture_size_if_rotated(self, monitor.size);
 }
 
-static void gsr_capture_kms_fail_fast_path_if_not_fast(gsr_capture_kms *self, uint32_t pixel_format) {
-    const uint8_t pixel_format_color_depth_1 = (pixel_format >> 16) & 0xFF;
-    if(!self->fast_path_failed && self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD && !self->mesa_supports_compute_only_vaapi_copy && (pixel_format_color_depth_1 == '3' || pixel_format_color_depth_1 == '4')) {
-        self->fast_path_failed = true;
-        fprintf(stderr, "gsr warning: gsr_capture_kms_capture: the monitor you are recording is in 10/12-bit color format and your mesa version is <= 24.3.6, composition will be used."
-            " If you experience performance problems in the video then record on a single window on X11 or use portal capture option instead or disable 10/12-bit color option in your desktop environment settings,"
-            " or try to record the monitor on X11 instead (if you aren't already doing that) or update your mesa version.\n");
-    }
-}
-
-static int gsr_capture_kms_capture(gsr_capture *cap, AVFrame *frame, gsr_color_conversion *color_conversion) {
+static int gsr_capture_kms_capture(gsr_capture *cap, gsr_capture_metadata *capture_metadata, gsr_color_conversion *color_conversion) {
     gsr_capture_kms *self = cap->priv;
 
     gsr_capture_kms_cleanup_kms_fds(self);
@@ -643,64 +615,37 @@ static int gsr_capture_kms_capture(gsr_capture *cap, AVFrame *frame, gsr_color_c
     if(drm_fd->has_hdr_metadata && self->params.hdr && hdr_metadata_is_supported_format(&drm_fd->hdr_metadata))
         gsr_kms_set_hdr_metadata(self, drm_fd);
 
-    if(!self->performance_error_shown && self->monitor_rotation != GSR_MONITOR_ROT_0 && video_codec_context_is_vaapi(self->video_codec_context) && self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD) {
-        self->performance_error_shown = true;
-        self->fast_path_failed = true;
-        fprintf(stderr, "gsr warning: gsr_capture_kms_capture: the monitor you are recording is rotated, composition will have to be used."
-            " If you experience performance problems in the video then record a single window on X11 or use portal capture option instead\n");
-    }
-
-    gsr_capture_kms_fail_fast_path_if_not_fast(self, drm_fd->pixel_format);
-
     self->capture_size = rotate_capture_size_if_rotated(self, (vec2i){ drm_fd->src_w, drm_fd->src_h });
+    if(self->params.region_size.x > 0 && self->params.region_size.y > 0)
+        self->capture_size = self->params.region_size;
 
     const bool is_scaled = self->params.output_resolution.x > 0 && self->params.output_resolution.y > 0;
     vec2i output_size = is_scaled ? self->params.output_resolution : self->capture_size;
     output_size = scale_keep_aspect_ratio(self->capture_size, output_size);
 
-    const float texture_rotation = monitor_rotation_to_radians(self->monitor_rotation);
-    const vec2i target_pos = { max_int(0, frame->width / 2 - output_size.x / 2), max_int(0, frame->height / 2 - output_size.y / 2) };
+    const vec2i target_pos = { max_int(0, capture_metadata->width / 2 - output_size.x / 2), max_int(0, capture_metadata->height / 2 - output_size.y / 2) };
     gsr_capture_kms_update_capture_size_change(self, color_conversion, target_pos, drm_fd);
 
     vec2i capture_pos = self->capture_pos;
     if(!capture_is_combined_plane)
         capture_pos = (vec2i){drm_fd->x, drm_fd->y};
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    capture_pos.x += self->params.region_position.x;
+    capture_pos.y += self->params.region_position.y;
 
-    /* Fast opengl free path */
-    if(!self->fast_path_failed && self->monitor_rotation == GSR_MONITOR_ROT_0 && video_codec_context_is_vaapi(self->video_codec_context) && self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD) {
-        int fds[4];
-        uint32_t offsets[4];
-        uint32_t pitches[4];
-        uint64_t modifiers[4];
-        for(int i = 0; i < drm_fd->num_dma_bufs; ++i) {
-            fds[i] = drm_fd->dma_buf[i].fd;
-            offsets[i] = drm_fd->dma_buf[i].offset;
-            pitches[i] = drm_fd->dma_buf[i].pitch;
-            modifiers[i] = drm_fd->modifier;
-        }
-        if(!vaapi_copy_drm_planes_to_video_surface(self->video_codec_context, frame, (vec2i){capture_pos.x, capture_pos.y}, self->capture_size, target_pos, output_size, drm_fd->pixel_format, (vec2i){drm_fd->width, drm_fd->height}, fds, offsets, pitches, modifiers, drm_fd->num_dma_bufs)) {
-            fprintf(stderr, "gsr error: gsr_capture_kms_capture: vaapi_copy_drm_planes_to_video_surface failed, falling back to opengl copy. Please report this as an issue at https://github.com/dec05eba/gpu-screen-recorder-issues\n");
-            self->fast_path_failed = true;
-        }
-    } else {
-        self->fast_path_failed = true;
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
+
+    EGLImage image = gsr_capture_kms_create_egl_image_with_fallback(self, drm_fd);
+    if(image) {
+        gsr_capture_kms_bind_image_to_input_texture_with_fallback(self, image);
+        self->params.egl->eglDestroyImage(self->params.egl->egl_display, image);
     }
 
-    if(self->fast_path_failed) {
-        EGLImage image = gsr_capture_kms_create_egl_image_with_fallback(self, drm_fd);
-        if(image) {
-            gsr_capture_kms_bind_image_to_input_texture_with_fallback(self, image);
-            self->params.egl->eglDestroyImage(self->params.egl->egl_display, image);
-        }
-
-        gsr_color_conversion_draw(color_conversion, self->external_texture_fallback ? self->external_input_texture_id : self->input_texture_id,
-            target_pos, output_size,
-            capture_pos, self->capture_size,
-            texture_rotation, self->external_texture_fallback, GSR_SOURCE_COLOR_RGB);
-    }
+    gsr_color_conversion_draw(color_conversion, self->external_texture_fallback ? self->external_input_texture_id : self->input_texture_id,
+        target_pos, output_size,
+        capture_pos, self->capture_size,
+        gsr_monitor_rotation_to_rotation(self->monitor_rotation), self->external_texture_fallback, GSR_SOURCE_COLOR_RGB);
 
     if(self->params.record_cursor) {
         gsr_kms_response_item *cursor_drm_fd = find_cursor_drm_if_on_monitor(self, drm_fd->connector_id, capture_is_combined_plane);
@@ -709,15 +654,18 @@ static int gsr_capture_kms_capture(gsr_capture *cap, AVFrame *frame, gsr_color_c
         // TODO: This doesn't work properly with software cursor on x11 since it will draw the x11 cursor on top of the cursor already in the framebuffer.
         // Detect if software cursor is used on x11 somehow.
         if(self->is_x11) {
-            const vec2i cursor_monitor_offset = self->capture_pos;
+            vec2i cursor_monitor_offset = self->capture_pos;
+            cursor_monitor_offset.x += self->params.region_position.x;
+            cursor_monitor_offset.y += self->params.region_position.y;
             render_x11_cursor(self, color_conversion, cursor_monitor_offset, target_pos, output_size);
         } else if(cursor_drm_fd) {
-            render_drm_cursor(self, color_conversion, cursor_drm_fd, target_pos, texture_rotation, output_size);
+            const vec2i framebuffer_size = rotate_capture_size_if_rotated(self, (vec2i){ drm_fd->src_w, drm_fd->src_h });
+            render_drm_cursor(self, color_conversion, cursor_drm_fd, target_pos, output_size, framebuffer_size);
         }
     }
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
 
     gsr_capture_kms_cleanup_kms_fds(self);
 
@@ -756,8 +704,8 @@ static bool gsr_capture_kms_set_hdr_metadata(gsr_capture *cap, AVMasteringDispla
     mastering_display_metadata->min_luminance = av_make_q(self->hdr_metadata.hdmi_metadata_type1.min_display_mastering_luminance, 10000);
     mastering_display_metadata->max_luminance = av_make_q(self->hdr_metadata.hdmi_metadata_type1.max_display_mastering_luminance, 1);
 
-    mastering_display_metadata->has_primaries = mastering_display_metadata->display_primaries[0][0].num > 0;
-    mastering_display_metadata->has_luminance = mastering_display_metadata->max_luminance.num > 0;
+    mastering_display_metadata->has_primaries = true;
+    mastering_display_metadata->has_luminance = true;
 
     return true;
 }
@@ -772,8 +720,7 @@ static bool gsr_capture_kms_set_hdr_metadata(gsr_capture *cap, AVMasteringDispla
 //     self->damaged = false;
 // }
 
-static void gsr_capture_kms_destroy(gsr_capture *cap, AVCodecContext *video_codec_context) {
-    (void)video_codec_context;
+static void gsr_capture_kms_destroy(gsr_capture *cap) {
     gsr_capture_kms *self = cap->priv;
     if(cap->priv) {
         gsr_capture_kms_stop(self);

src/capture/nvfbc.c

@@ -13,7 +13,6 @@
 #include <assert.h>
 
 #include <X11/Xlib.h>
-#include <libavcodec/avcodec.h>
 
 typedef struct {
     gsr_capture_nvfbc_params params;
@@ -28,8 +27,7 @@ typedef struct {
     NVFBC_TOGL_SETUP_PARAMS setup_params;
 
     bool supports_direct_cursor;
-    bool capture_region;
-    uint32_t x, y, width, height;
+    uint32_t width, height;
     NVFBC_TRACKING_TYPE tracking_type;
     uint32_t output_id;
     uint32_t tracking_width, tracking_height;
@@ -133,31 +131,6 @@ static bool gsr_capture_nvfbc_load_library(gsr_capture *cap) {
     return true;
 }
 
-/* TODO: check for glx swap control extension string (GLX_EXT_swap_control, etc) */
-static void set_vertical_sync_enabled(gsr_egl *egl, int enabled) {
-    int result = 0;
-
-    if(egl->glXSwapIntervalEXT) {
-        assert(gsr_window_get_display_server(egl->window) == GSR_DISPLAY_SERVER_X11);
-        Display *display = gsr_window_get_display(egl->window);
-        const Window window = (Window)gsr_window_get_window(egl->window);
-        egl->glXSwapIntervalEXT(display, window, enabled ? 1 : 0);
-    } else if(egl->glXSwapIntervalMESA) {
-        result = egl->glXSwapIntervalMESA(enabled ? 1 : 0);
-    } else if(egl->glXSwapIntervalSGI) {
-        result = egl->glXSwapIntervalSGI(enabled ? 1 : 0);
-    } else {
-        static int warned = 0;
-        if (!warned) {
-            warned = 1;
-            fprintf(stderr, "gsr warning: setting vertical sync not supported\n");
-        }
-    }
-
-    if(result != 0)
-        fprintf(stderr, "gsr warning: setting vertical sync failed\n");
-}
-
 static void gsr_capture_nvfbc_destroy_session(gsr_capture_nvfbc *self) {
     if(self->fbc_handle_created && self->capture_session_created) {
         NVFBC_DESTROY_CAPTURE_SESSION_PARAMS destroy_capture_params;
@@ -248,11 +221,8 @@ static int gsr_capture_nvfbc_setup_handle(gsr_capture_nvfbc *self) {
         }
     }
 
-    if(!self->capture_region) {
-        self->width = self->tracking_width;
-        self->height = self->tracking_height;
-    }
-
+    self->width = self->tracking_width;
+    self->height = self->tracking_height;
     return 0;
 
     error_cleanup:
@@ -268,8 +238,6 @@ static int gsr_capture_nvfbc_setup_session(gsr_capture_nvfbc *self) {
     create_capture_params.bWithCursor = (!self->params.direct_capture || self->supports_direct_cursor) ? NVFBC_TRUE : NVFBC_FALSE;
     if(!self->params.record_cursor)
         create_capture_params.bWithCursor = false;
-    if(self->capture_region)
-        create_capture_params.captureBox = (NVFBC_BOX){ self->x, self->y, self->width, self->height };
     create_capture_params.eTrackingType = self->tracking_type;
     create_capture_params.dwSamplingRateMs = (uint32_t)ceilf(1000.0f / (float)self->params.fps);
     create_capture_params.bAllowDirectCapture = self->params.direct_capture ? NVFBC_TRUE : NVFBC_FALSE;
@@ -311,19 +279,12 @@ static void gsr_capture_nvfbc_stop(gsr_capture_nvfbc *self) {
     }
 }
 
-static int gsr_capture_nvfbc_start(gsr_capture *cap, AVCodecContext *video_codec_context, AVFrame *frame) {
+static int gsr_capture_nvfbc_start(gsr_capture *cap, gsr_capture_metadata *capture_metadata) {
     gsr_capture_nvfbc *self = cap->priv;
 
     if(!gsr_capture_nvfbc_load_library(cap))
         return -1;
 
-    self->x = max_int(self->params.pos.x, 0);
-    self->y = max_int(self->params.pos.y, 0);
-    self->width = max_int(self->params.size.x, 0);
-    self->height = max_int(self->params.size.y, 0);
-
-    self->capture_region = (self->x > 0 || self->y > 0 || self->width > 0 || self->height > 0);
-
     self->supports_direct_cursor = false;
     int driver_major_version = 0;
     int driver_minor_version = 0;
@@ -356,28 +317,18 @@ static int gsr_capture_nvfbc_start(gsr_capture *cap, AVCodecContext *video_codec
         goto error_cleanup;
     }
 
-    if(self->capture_region) {
-        video_codec_context->width = FFALIGN(self->width, 2);
-        video_codec_context->height = FFALIGN(self->height, 2);
-    } else {
-        video_codec_context->width = FFALIGN(self->tracking_width, 2);
-        video_codec_context->height = FFALIGN(self->tracking_height, 2);
+    capture_metadata->width = self->tracking_width;
+    capture_metadata->height = self->tracking_height;
+
+    if(self->params.output_resolution.x > 0 && self->params.output_resolution.y > 0) {
+        self->params.output_resolution = scale_keep_aspect_ratio((vec2i){capture_metadata->width, capture_metadata->height}, self->params.output_resolution);
+        capture_metadata->width = self->params.output_resolution.x;
+        capture_metadata->height = self->params.output_resolution.y;
+    } else if(self->params.region_size.x > 0 && self->params.region_size.y > 0) {
+        capture_metadata->width = self->params.region_size.x;
+        capture_metadata->height = self->params.region_size.y;
     }
 
-    if(self->params.output_resolution.x == 0 && self->params.output_resolution.y == 0) {
-        self->params.output_resolution = (vec2i){video_codec_context->width, video_codec_context->height};
-    } else {
-        self->params.output_resolution = scale_keep_aspect_ratio((vec2i){video_codec_context->width, video_codec_context->height}, self->params.output_resolution);
-        video_codec_context->width = FFALIGN(self->params.output_resolution.x, 2);
-        video_codec_context->height = FFALIGN(self->params.output_resolution.y, 2);
-    }
-
-    frame->width = video_codec_context->width;
-    frame->height = video_codec_context->height;
-
-    /* Disable vsync */
-    set_vertical_sync_enabled(self->params.egl, 0);
-
     return 0;
 
     error_cleanup:
@@ -385,7 +336,7 @@ static int gsr_capture_nvfbc_start(gsr_capture *cap, AVCodecContext *video_codec
     return -1;
 }
 
-static int gsr_capture_nvfbc_capture(gsr_capture *cap, AVFrame *frame, gsr_color_conversion *color_conversion) {
+static int gsr_capture_nvfbc_capture(gsr_capture *cap, gsr_capture_metadata *capture_metadata, gsr_color_conversion *color_conversion) {
     gsr_capture_nvfbc *self = cap->priv;
 
     const double nvfbc_recreate_retry_time_seconds = 1.0;
@@ -411,12 +362,15 @@ static int gsr_capture_nvfbc_capture(gsr_capture *cap, AVFrame *frame, gsr_color
         }
     }
 
-    const vec2i frame_size = (vec2i){self->width, self->height};
+    vec2i frame_size = (vec2i){self->width, self->height};
+    if(self->params.region_size.x > 0 && self->params.region_size.y > 0)
+        frame_size = self->params.region_size;
+
     const bool is_scaled = self->params.output_resolution.x > 0 && self->params.output_resolution.y > 0;
     vec2i output_size = is_scaled ? self->params.output_resolution : frame_size;
     output_size = scale_keep_aspect_ratio(frame_size, output_size);
 
-    const vec2i target_pos = { max_int(0, frame->width / 2 - output_size.x / 2), max_int(0, frame->height / 2 - output_size.y / 2) };
+    const vec2i target_pos = { max_int(0, capture_metadata->width / 2 - output_size.x / 2), max_int(0, capture_metadata->height / 2 - output_size.y / 2) };
 
     NVFBC_FRAME_GRAB_INFO frame_info;
     memset(&frame_info, 0, sizeof(frame_info));
@@ -436,22 +390,21 @@ static int gsr_capture_nvfbc_capture(gsr_capture *cap, AVFrame *frame, gsr_color
         return 0;
     }
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
 
     gsr_color_conversion_draw(color_conversion, self->setup_params.dwTextures[grab_params.dwTextureIndex],
         target_pos, (vec2i){output_size.x, output_size.y},
-        (vec2i){0, 0}, frame_size,
-        0.0f, false, GSR_SOURCE_COLOR_BGR);
+        self->params.region_position, frame_size,
+        GSR_ROT_0, false, GSR_SOURCE_COLOR_BGR);
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
 
     return 0;
 }
 
-static void gsr_capture_nvfbc_destroy(gsr_capture *cap, AVCodecContext *video_codec_context) {
-    (void)video_codec_context;
+static void gsr_capture_nvfbc_destroy(gsr_capture *cap) {
     gsr_capture_nvfbc *self = cap->priv;
     gsr_capture_nvfbc_stop(self);
     free(cap->priv);

src/capture/portal.c

@@ -8,10 +8,9 @@
 #include <stdlib.h>
 #include <stdio.h>
 #include <unistd.h>
+#include <limits.h>
 #include <assert.h>
 
-#include <libavcodec/avcodec.h>
-
 typedef struct {
     gsr_capture_portal_params params;
 
@@ -24,10 +23,6 @@ typedef struct {
     vec2i capture_size;
     gsr_pipewire_video_dmabuf_data dmabuf_data[GSR_PIPEWIRE_VIDEO_DMABUF_MAX_PLANES];
     int num_dmabuf_data;
-
-    AVCodecContext *video_codec_context;
-    bool fast_path_failed;
-    bool mesa_supports_compute_only_vaapi_copy;
 } gsr_capture_portal;
 
 static void gsr_capture_portal_cleanup_plane_fds(gsr_capture_portal *self) {
@@ -69,26 +64,31 @@ static void gsr_capture_portal_stop(gsr_capture_portal *self) {
 }
 
 static void gsr_capture_portal_create_input_textures(gsr_capture_portal *self) {
+    const float border_color[4] = { 0.0f, 0.0f, 0.0f, 0.0f };
+
     self->params.egl->glGenTextures(1, &self->texture_map.texture_id);
     self->params.egl->glBindTexture(GL_TEXTURE_2D, self->texture_map.texture_id);
-    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, border_color);
     self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
     self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->params.egl->glBindTexture(GL_TEXTURE_2D, 0);
 
     self->params.egl->glGenTextures(1, &self->texture_map.external_texture_id);
     self->params.egl->glBindTexture(GL_TEXTURE_EXTERNAL_OES, self->texture_map.external_texture_id);
-    self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameterfv(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_BORDER_COLOR, border_color);
     self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
     self->params.egl->glTexParameteri(GL_TEXTURE_EXTERNAL_OES, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->params.egl->glBindTexture(GL_TEXTURE_EXTERNAL_OES, 0);
 
     self->params.egl->glGenTextures(1, &self->texture_map.cursor_texture_id);
     self->params.egl->glBindTexture(GL_TEXTURE_2D, self->texture_map.cursor_texture_id);
-    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->params.egl->glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, border_color);
     self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
     self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->params.egl->glBindTexture(GL_TEXTURE_2D, 0);
@@ -257,7 +257,7 @@ static bool gsr_capture_portal_get_frame_dimensions(gsr_capture_portal *self) {
     return false;
 }
 
-static int gsr_capture_portal_start(gsr_capture *cap, AVCodecContext *video_codec_context, AVFrame *frame) {
+static int gsr_capture_portal_start(gsr_capture *cap, gsr_capture_metadata *capture_metadata) {
     gsr_capture_portal *self = cap->priv;
 
     gsr_capture_portal_create_input_textures(self);
@@ -286,7 +286,7 @@ static int gsr_capture_portal_start(gsr_capture *cap, AVCodecContext *video_code
     fprintf(stderr, "gsr info: gsr_capture_portal_start: setting up pipewire\n");
     /* TODO: support hdr when pipewire supports it */
     /* gsr_pipewire closes the pipewire fd, even on failure */
-    if(!gsr_pipewire_video_init(&self->pipewire, pipewire_fd, pipewire_node, video_codec_context->framerate.num, self->params.record_cursor, self->params.egl)) {
+    if(!gsr_pipewire_video_init(&self->pipewire, pipewire_fd, pipewire_node, capture_metadata->fps, self->params.record_cursor, self->params.egl)) {
         fprintf(stderr, "gsr error: gsr_capture_portal_start: failed to setup pipewire with fd: %d, node: %" PRIu32 "\n", pipewire_fd, pipewire_node);
         gsr_capture_portal_stop(self);
         return -1;
@@ -298,29 +298,15 @@ static int gsr_capture_portal_start(gsr_capture *cap, AVCodecContext *video_code
         return -1;
     }
 
-    /* Disable vsync */
-    self->params.egl->eglSwapInterval(self->params.egl->egl_display, 0);
-
     if(self->params.output_resolution.x == 0 && self->params.output_resolution.y == 0) {
-        self->params.output_resolution = self->capture_size;
-        video_codec_context->width = FFALIGN(self->capture_size.x, 2);
-        video_codec_context->height = FFALIGN(self->capture_size.y, 2);
+        capture_metadata->width = self->capture_size.x;
+        capture_metadata->height = self->capture_size.y;
     } else {
         self->params.output_resolution = scale_keep_aspect_ratio(self->capture_size, self->params.output_resolution);
-        video_codec_context->width = FFALIGN(self->params.output_resolution.x, 2);
-        video_codec_context->height = FFALIGN(self->params.output_resolution.y, 2);
+        capture_metadata->width = self->params.output_resolution.x;
+        capture_metadata->height = self->params.output_resolution.y;
     }
 
-    self->fast_path_failed = self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD && !gl_driver_version_greater_than(&self->params.egl->gpu_info, 24, 0, 9);
-    if(self->fast_path_failed)
-        fprintf(stderr, "gsr warning: gsr_capture_kms_start: your amd driver (mesa) version is known to be buggy (<= version 24.0.9), falling back to opengl copy\n");
-
-    self->mesa_supports_compute_only_vaapi_copy = self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD && gl_driver_version_greater_than(&self->params.egl->gpu_info, 24, 3, 6);
-
-    frame->width = video_codec_context->width;
-    frame->height = video_codec_context->height;
-
-    self->video_codec_context = video_codec_context;
     return 0;
 }
 
@@ -328,18 +314,7 @@ static int max_int(int a, int b) {
     return a > b ? a : b;
 }
 
-static void gsr_capture_portal_fail_fast_path_if_not_fast(gsr_capture_portal *self, uint32_t pixel_format) {
-    const uint8_t pixel_format_color_depth_1 = (pixel_format >> 16) & 0xFF;
-    if(!self->fast_path_failed && self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD && !self->mesa_supports_compute_only_vaapi_copy && (pixel_format_color_depth_1 == '3' || pixel_format_color_depth_1 == '4')) {
-        self->fast_path_failed = true;
-        fprintf(stderr, "gsr warning: gsr_capture_kms_capture: the monitor you are recording is in 10/12-bit color format and your mesa version is <= 24.3.6, composition will be used."
-            " If you experience performance problems in the video then record on a single window on X11 instead or disable 10/12-bit color option in your desktop environment settings,"
-            " or try to record the monitor on X11 instead (if you aren't already doing that) or update your mesa version.\n");
-    }
-}
-
-static int gsr_capture_portal_capture(gsr_capture *cap, AVFrame *frame, gsr_color_conversion *color_conversion) {
-    (void)frame;
+static int gsr_capture_portal_capture(gsr_capture *cap, gsr_capture_metadata *capture_metadata, gsr_color_conversion *color_conversion) {
     (void)color_conversion;
     gsr_capture_portal *self = cap->priv;
 
@@ -359,45 +334,21 @@ static int gsr_capture_portal_capture(gsr_capture *cap, AVFrame *frame, gsr_colo
         return 0;
     }
 
-    gsr_capture_portal_fail_fast_path_if_not_fast(self, pipewire_fourcc);
-
     const bool is_scaled = self->params.output_resolution.x > 0 && self->params.output_resolution.y > 0;
     vec2i output_size = is_scaled ? self->params.output_resolution : self->capture_size;
     output_size = scale_keep_aspect_ratio(self->capture_size, output_size);
     
-    const vec2i target_pos = { max_int(0, frame->width / 2 - output_size.x / 2), max_int(0, frame->height / 2 - output_size.y / 2) };
+    const vec2i target_pos = { max_int(0, capture_metadata->width / 2 - output_size.x / 2), max_int(0, capture_metadata->height / 2 - output_size.y / 2) };
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
 
     // TODO: Handle region crop
 
-    /* Fast opengl free path */
-    if(!self->fast_path_failed && video_codec_context_is_vaapi(self->video_codec_context) && self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD) {
-        int fds[4];
-        uint32_t offsets[4];
-        uint32_t pitches[4];
-        uint64_t modifiers[4];
-        for(int i = 0; i < self->num_dmabuf_data; ++i) {
-            fds[i] = self->dmabuf_data[i].fd;
-            offsets[i] = self->dmabuf_data[i].offset;
-            pitches[i] = self->dmabuf_data[i].stride;
-            modifiers[i] = pipewire_modifiers;
-        }
-        if(!vaapi_copy_drm_planes_to_video_surface(self->video_codec_context, frame, (vec2i){region.x, region.y}, self->capture_size, target_pos, output_size, pipewire_fourcc, self->capture_size, fds, offsets, pitches, modifiers, self->num_dmabuf_data)) {
-            fprintf(stderr, "gsr error: gsr_capture_portal_capture: vaapi_copy_drm_planes_to_video_surface failed, falling back to opengl copy. Please report this as an issue at https://github.com/dec05eba/gpu-screen-recorder-issues\n");
-            self->fast_path_failed = true;
-        }
-    } else {
-        self->fast_path_failed = true;
-    }
-
-    if(self->fast_path_failed) {
-        gsr_color_conversion_draw(color_conversion, using_external_image ? self->texture_map.external_texture_id : self->texture_map.texture_id,
-            target_pos, output_size,
-            (vec2i){region.x, region.y}, self->capture_size,
-            0.0f, using_external_image, GSR_SOURCE_COLOR_RGB);
-    }
+    gsr_color_conversion_draw(color_conversion, using_external_image ? self->texture_map.external_texture_id : self->texture_map.texture_id,
+        target_pos, output_size,
+        (vec2i){region.x, region.y}, self->capture_size,
+        GSR_ROT_0, using_external_image, GSR_SOURCE_COLOR_RGB);
 
     if(self->params.record_cursor && self->texture_map.cursor_texture_id > 0 && cursor_region.width > 0) {
         const vec2d scale = {
@@ -415,12 +366,12 @@ static int gsr_capture_portal_capture(gsr_capture *cap, AVFrame *frame, gsr_colo
         gsr_color_conversion_draw(color_conversion, self->texture_map.cursor_texture_id,
             (vec2i){cursor_pos.x, cursor_pos.y}, (vec2i){cursor_region.width * scale.x, cursor_region.height * scale.y},
             (vec2i){0, 0}, (vec2i){cursor_region.width, cursor_region.height},
-            0.0f, false, GSR_SOURCE_COLOR_RGB);
+            GSR_ROT_0, false, GSR_SOURCE_COLOR_RGB);
         self->params.egl->glDisable(GL_SCISSOR_TEST);
     }
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
 
     gsr_capture_portal_cleanup_plane_fds(self);
 
@@ -442,8 +393,7 @@ static void gsr_capture_portal_clear_damage(gsr_capture *cap) {
     gsr_pipewire_video_clear_damage(&self->pipewire);
 }
 
-static void gsr_capture_portal_destroy(gsr_capture *cap, AVCodecContext *video_codec_context) {
-    (void)video_codec_context;
+static void gsr_capture_portal_destroy(gsr_capture *cap) {
     gsr_capture_portal *self = cap->priv;
     if(cap->priv) {
         gsr_capture_portal_stop(self);

src/capture/xcomposite.c

@@ -12,9 +12,6 @@
 
 #include <X11/Xlib.h>
 
-#include <libavutil/frame.h>
-#include <libavcodec/avcodec.h>
-
 typedef struct {
     gsr_capture_xcomposite_params params;
     Display *display;
@@ -31,14 +28,12 @@ typedef struct {
     double window_resize_timer;
     
     WindowTexture window_texture;
-    AVCodecContext *video_codec_context;
 
     Atom net_active_window_atom;
 
     gsr_cursor cursor;
 
     bool clear_background;
-    bool fast_path_failed;
 } gsr_capture_xcomposite;
 
 static void gsr_capture_xcomposite_stop(gsr_capture_xcomposite *self) {
@@ -64,7 +59,7 @@ static Window get_focused_window(Display *display, Atom net_active_window_atom)
     return None;
 }
 
-static int gsr_capture_xcomposite_start(gsr_capture *cap, AVCodecContext *video_codec_context, AVFrame *frame) {
+static int gsr_capture_xcomposite_start(gsr_capture *cap, gsr_capture_metadata *capture_metadata) {
     gsr_capture_xcomposite *self = cap->priv;
 
     if(self->params.follow_focused) {
@@ -95,10 +90,8 @@ static int gsr_capture_xcomposite_start(gsr_capture *cap, AVCodecContext *video_
     // TODO: Get select and add these on top of it and then restore at the end. Also do the same in other xcomposite
     XSelectInput(self->display, self->window, StructureNotifyMask | ExposureMask);
 
-    /* Disable vsync */
-    self->params.egl->eglSwapInterval(self->params.egl->egl_display, 0);
     if(window_texture_init(&self->window_texture, self->display, self->window, self->params.egl) != 0 && !self->params.follow_focused) {
-        fprintf(stderr, "gsr error: gsr_capture_xcomposite_start: failed to get window texture for window %ld\n", self->window);
+        fprintf(stderr, "gsr error: gsr_capture_xcomposite_start: failed to get window texture for window %ld\n", (long)self->window);
         return -1;
     }
 
@@ -116,22 +109,13 @@ static int gsr_capture_xcomposite_start(gsr_capture *cap, AVCodecContext *video_
     self->params.egl->glBindTexture(GL_TEXTURE_2D, 0);
 
     if(self->params.output_resolution.x == 0 && self->params.output_resolution.y == 0) {
-        self->params.output_resolution = self->texture_size;
-        video_codec_context->width = FFALIGN(self->texture_size.x, 2);
-        video_codec_context->height = FFALIGN(self->texture_size.y, 2);
+        capture_metadata->width = self->texture_size.x;
+        capture_metadata->height = self->texture_size.y;
     } else {
-        video_codec_context->width = FFALIGN(self->params.output_resolution.x, 2);
-        video_codec_context->height = FFALIGN(self->params.output_resolution.y, 2);
+        capture_metadata->width = self->params.output_resolution.x;
+        capture_metadata->height = self->params.output_resolution.y;
     }
 
-    self->fast_path_failed = self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD && !gl_driver_version_greater_than(&self->params.egl->gpu_info, 24, 0, 9);
-    if(self->fast_path_failed)
-        fprintf(stderr, "gsr warning: gsr_capture_kms_start: your amd driver (mesa) version is known to be buggy (<= version 24.0.9), falling back to opengl copy\n");
-
-    frame->width = video_codec_context->width;
-    frame->height = video_codec_context->height;
-
-    self->video_codec_context = video_codec_context;
     self->window_resize_timer = clock_get_monotonic_seconds();
     return 0;
 }
@@ -255,9 +239,8 @@ static bool gsr_capture_xcomposite_should_stop(gsr_capture *cap, bool *err) {
     return false;
 }
 
-static int gsr_capture_xcomposite_capture(gsr_capture *cap, AVFrame *frame, gsr_color_conversion *color_conversion) {
+static int gsr_capture_xcomposite_capture(gsr_capture *cap, gsr_capture_metadata *capture_metdata, gsr_color_conversion *color_conversion) {
     gsr_capture_xcomposite *self = cap->priv;
-    (void)frame;
 
     if(self->clear_background) {
         self->clear_background = false;
@@ -268,27 +251,15 @@ static int gsr_capture_xcomposite_capture(gsr_capture *cap, AVFrame *frame, gsr_
     vec2i output_size = is_scaled ? self->params.output_resolution : self->texture_size;
     output_size = scale_keep_aspect_ratio(self->texture_size, output_size);
 
-    const vec2i target_pos = { max_int(0, frame->width / 2 - output_size.x / 2), max_int(0, frame->height / 2 - output_size.y / 2) };
+    const vec2i target_pos = { max_int(0, capture_metdata->width / 2 - output_size.x / 2), max_int(0, capture_metdata->height / 2 - output_size.y / 2) };
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
 
-    /* Fast opengl free path */
-    if(!self->fast_path_failed && video_codec_context_is_vaapi(self->video_codec_context) && self->params.egl->gpu_info.vendor == GSR_GPU_VENDOR_AMD) {
-        if(!vaapi_copy_egl_image_to_video_surface(self->params.egl, self->window_texture.image, (vec2i){0, 0}, self->texture_size, target_pos, output_size, self->video_codec_context, frame)) {
-            fprintf(stderr, "gsr error: gsr_capture_xcomposite_capture: vaapi_copy_egl_image_to_video_surface failed, falling back to opengl copy. Please report this as an issue at https://github.com/dec05eba/gpu-screen-recorder-issues\n");
-            self->fast_path_failed = true;
-        }
-    } else {
-        self->fast_path_failed = true;
-    }
-
-    if(self->fast_path_failed) {
-        gsr_color_conversion_draw(color_conversion, window_texture_get_opengl_texture_id(&self->window_texture),
-            target_pos, output_size,
-            (vec2i){0, 0}, self->texture_size,
-            0.0f, false, GSR_SOURCE_COLOR_RGB);
-    }
+    gsr_color_conversion_draw(color_conversion, window_texture_get_opengl_texture_id(&self->window_texture),
+        target_pos, output_size,
+        (vec2i){0, 0}, self->texture_size,
+        GSR_ROT_0, false, GSR_SOURCE_COLOR_RGB);
 
     if(self->params.record_cursor && self->cursor.visible) {
         const vec2d scale = {
@@ -303,19 +274,17 @@ static int gsr_capture_xcomposite_capture(gsr_capture *cap, AVFrame *frame, gsr_
             target_pos.y + (self->cursor.position.y - self->cursor.hotspot.y) * scale.y
         };
 
-        self->params.egl->glEnable(GL_SCISSOR_TEST);
-        self->params.egl->glScissor(target_pos.x, target_pos.y, output_size.x, output_size.y);
+        if(cursor_pos.x < target_pos.x || cursor_pos.x + self->cursor.size.x > target_pos.x + output_size.x || cursor_pos.y < target_pos.y || cursor_pos.y + self->cursor.size.y > target_pos.y + output_size.y)
+            self->clear_background = true;
 
         gsr_color_conversion_draw(color_conversion, self->cursor.texture_id,
             cursor_pos, (vec2i){self->cursor.size.x * scale.x, self->cursor.size.y * scale.y},
             (vec2i){0, 0}, self->cursor.size,
-            0.0f, false, GSR_SOURCE_COLOR_RGB);
-
-        self->params.egl->glDisable(GL_SCISSOR_TEST);
+            GSR_ROT_0, false, GSR_SOURCE_COLOR_RGB);
     }
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
 
     return 0;
 }
@@ -325,8 +294,7 @@ static uint64_t gsr_capture_xcomposite_get_window_id(gsr_capture *cap) {
     return self->window;
 }
 
-static void gsr_capture_xcomposite_destroy(gsr_capture *cap, AVCodecContext *video_codec_context) {
-    (void)video_codec_context;
+static void gsr_capture_xcomposite_destroy(gsr_capture *cap) {
     if(cap->priv) {
         gsr_capture_xcomposite_stop(cap->priv);
         free(cap->priv);

src/capture/ximage.c

@@ -0,0 +1,247 @@
+#include "../../include/capture/ximage.h"
+#include "../../include/utils.h"
+#include "../../include/cursor.h"
+#include "../../include/color_conversion.h"
+#include "../../include/window/window.h"
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <assert.h>
+
+#include <X11/Xlib.h>
+
+/* TODO: update when monitors are reconfigured */
+
+typedef struct {
+    gsr_capture_ximage_params params;
+    Display *display;
+    gsr_cursor cursor;
+    gsr_monitor monitor;
+    vec2i capture_pos;
+    vec2i capture_size;
+    unsigned int texture_id;
+    Window root_window;
+} gsr_capture_ximage;
+
+static void gsr_capture_ximage_stop(gsr_capture_ximage *self) {
+    gsr_cursor_deinit(&self->cursor);
+    if(self->texture_id) {
+        self->params.egl->glDeleteTextures(1, &self->texture_id);
+        self->texture_id = 0;
+    }
+}
+
+static int max_int(int a, int b) {
+    return a > b ? a : b;
+}
+
+static int gsr_capture_ximage_start(gsr_capture *cap, gsr_capture_metadata *capture_metadata) {
+    gsr_capture_ximage *self = cap->priv;
+    self->root_window = DefaultRootWindow(self->display);
+
+    if(gsr_cursor_init(&self->cursor, self->params.egl, self->display) != 0) {
+        gsr_capture_ximage_stop(self);
+        return -1;
+    }
+
+    if(!get_monitor_by_name(self->params.egl, GSR_CONNECTION_X11, self->params.display_to_capture, &self->monitor)) {
+        fprintf(stderr, "gsr error: gsr_capture_ximage_start: failed to find monitor by name \"%s\"\n", self->params.display_to_capture);
+        gsr_capture_ximage_stop(self);
+        return -1;
+    }
+
+    self->capture_pos = self->monitor.pos;
+    self->capture_size = self->monitor.size;
+
+    if(self->params.region_size.x > 0 && self->params.region_size.y > 0)
+        self->capture_size = self->params.region_size;
+
+    if(self->params.output_resolution.x > 0 && self->params.output_resolution.y > 0) {
+        self->params.output_resolution = scale_keep_aspect_ratio(self->capture_size, self->params.output_resolution);
+        capture_metadata->width = self->params.output_resolution.x;
+        capture_metadata->height = self->params.output_resolution.y;
+    } else if(self->params.region_size.x > 0 && self->params.region_size.y > 0) {
+        capture_metadata->width = self->params.region_size.x;
+        capture_metadata->height = self->params.region_size.y;
+    } else {
+        capture_metadata->width = self->capture_size.x;
+        capture_metadata->height = self->capture_size.y;
+    }
+
+    self->texture_id = gl_create_texture(self->params.egl, self->capture_size.x, self->capture_size.y, GL_RGB8, GL_RGB, GL_LINEAR);
+    if(self->texture_id == 0) {
+        fprintf(stderr, "gsr error: gsr_capture_ximage_start: failed to create texture\n");
+        gsr_capture_ximage_stop(self);
+        return -1;
+    }
+
+    return 0;
+}
+
+static void gsr_capture_ximage_on_event(gsr_capture *cap, gsr_egl *egl) {
+    gsr_capture_ximage *self = cap->priv;
+    XEvent *xev = gsr_window_get_event_data(egl->window);
+    gsr_cursor_on_event(&self->cursor, xev);
+}
+
+static bool gsr_capture_ximage_upload_to_texture(gsr_capture_ximage *self, int x, int y, int width, int height) {
+    const int max_width = XWidthOfScreen(DefaultScreenOfDisplay(self->display));
+    const int max_height = XHeightOfScreen(DefaultScreenOfDisplay(self->display));
+
+    if(x < 0)
+        x = 0;
+    else if(x >= max_width)
+        x = max_width - 1;
+
+    if(y < 0)
+        y = 0;
+    else if(y >= max_height)
+        y = max_height - 1;
+
+    if(width < 0)
+        width = 0;
+    else if(x + width >= max_width)
+        width = max_width - x;
+
+    if(height < 0)
+        height = 0;
+    else if(y + height >= max_height)
+        height = max_height - y;
+
+    XImage *image = XGetImage(self->display, self->root_window, x, y, width, height, AllPlanes, ZPixmap);
+    if(!image) {
+        fprintf(stderr, "gsr error: gsr_capture_ximage_upload_to_texture: XGetImage failed\n");
+        return false;
+    }
+
+    bool success = false;
+    uint8_t *image_data = malloc(image->width * image->height * 3);
+    if(!image_data) {
+        fprintf(stderr, "gsr error: gsr_capture_ximage_upload_to_texture: failed to allocate image data\n");
+        goto done;
+    }
+
+    for(int y = 0; y < image->height; ++y) {
+        for(int x = 0; x < image->width; ++x) {
+            unsigned long pixel = XGetPixel(image, x, y);
+            unsigned char red = (pixel & image->red_mask) >> 16;
+            unsigned char green = (pixel & image->green_mask) >> 8;
+            unsigned char blue = pixel & image->blue_mask;
+
+            const size_t texture_data_index = (x + y * image->width) * 3;
+            image_data[texture_data_index + 0] = red;
+            image_data[texture_data_index + 1] = green;
+            image_data[texture_data_index + 2] = blue;
+        }
+    }
+
+    self->params.egl->glBindTexture(GL_TEXTURE_2D, self->texture_id);
+    self->params.egl->glTexSubImage2D(GL_TEXTURE_2D, 0, 0, 0, image->width, image->height, GL_RGB, GL_UNSIGNED_BYTE, image_data);
+    self->params.egl->glBindTexture(GL_TEXTURE_2D, 0);
+    success = true;
+
+    done:
+    free(image_data);
+    XDestroyImage(image);
+    return success;
+}
+
+static int gsr_capture_ximage_capture(gsr_capture *cap, gsr_capture_metadata *capture_metdata, gsr_color_conversion *color_conversion) {
+    gsr_capture_ximage *self = cap->priv;
+
+    const bool is_scaled = self->params.output_resolution.x > 0 && self->params.output_resolution.y > 0;
+    vec2i output_size = is_scaled ? self->params.output_resolution : self->capture_size;
+    output_size = scale_keep_aspect_ratio(self->capture_size, output_size);
+
+    const vec2i target_pos = { max_int(0, capture_metdata->width / 2 - output_size.x / 2), max_int(0, capture_metdata->height / 2 - output_size.y / 2) };
+    gsr_capture_ximage_upload_to_texture(self, self->capture_pos.x + self->params.region_position.x, self->capture_pos.y + self->params.region_position.y, self->capture_size.x, self->capture_size.y);
+
+    gsr_color_conversion_draw(color_conversion, self->texture_id,
+        target_pos, output_size,
+        (vec2i){0, 0}, self->capture_size,
+        GSR_ROT_0, false, GSR_SOURCE_COLOR_RGB);
+
+    if(self->params.record_cursor && self->cursor.visible) {
+        const vec2d scale = {
+            self->capture_size.x == 0 ? 0 : (double)output_size.x / (double)self->capture_size.x,
+            self->capture_size.y == 0 ? 0 : (double)output_size.y / (double)self->capture_size.y
+        };
+
+        gsr_cursor_tick(&self->cursor, self->root_window);
+
+        const vec2i cursor_pos = {
+            target_pos.x + (self->cursor.position.x - self->cursor.hotspot.x) * scale.x - self->capture_pos.x - self->params.region_position.x,
+            target_pos.y + (self->cursor.position.y - self->cursor.hotspot.y) * scale.y - self->capture_pos.y - self->params.region_position.y
+        };
+
+        self->params.egl->glEnable(GL_SCISSOR_TEST);
+        self->params.egl->glScissor(target_pos.x, target_pos.y, output_size.x, output_size.y);
+
+        gsr_color_conversion_draw(color_conversion, self->cursor.texture_id,
+            cursor_pos, (vec2i){self->cursor.size.x * scale.x, self->cursor.size.y * scale.y},
+            (vec2i){0, 0}, self->cursor.size,
+            GSR_ROT_0, false, GSR_SOURCE_COLOR_RGB);
+
+        self->params.egl->glDisable(GL_SCISSOR_TEST);
+    }
+
+    self->params.egl->glFlush();
+    self->params.egl->glFinish();
+
+    return 0;
+}
+
+static void gsr_capture_ximage_destroy(gsr_capture *cap) {
+    gsr_capture_ximage *self = cap->priv;
+    if(cap->priv) {
+        gsr_capture_ximage_stop(self);
+        free((void*)self->params.display_to_capture);
+        self->params.display_to_capture = NULL;
+        free(self);
+        cap->priv = NULL;
+    }
+    free(cap);
+}
+
+gsr_capture* gsr_capture_ximage_create(const gsr_capture_ximage_params *params) {
+    if(!params) {
+        fprintf(stderr, "gsr error: gsr_capture_ximage_create params is NULL\n");
+        return NULL;
+    }
+
+    gsr_capture *cap = calloc(1, sizeof(gsr_capture));
+    if(!cap)
+        return NULL;
+
+    gsr_capture_ximage *cap_ximage = calloc(1, sizeof(gsr_capture_ximage));
+    if(!cap_ximage) {
+        free(cap);
+        return NULL;
+    }
+
+    const char *display_to_capture = strdup(params->display_to_capture);
+    if(!display_to_capture) {
+        free(cap);
+        free(cap_ximage);
+        return NULL;
+    }
+
+    cap_ximage->params = *params;
+    cap_ximage->display = gsr_window_get_display(params->egl->window);
+    cap_ximage->params.display_to_capture = display_to_capture;
+    
+    *cap = (gsr_capture) {
+        .start = gsr_capture_ximage_start,
+        .on_event = gsr_capture_ximage_on_event,
+        .tick = NULL,
+        .should_stop = NULL,
+        .capture = gsr_capture_ximage_capture,
+        .uses_external_image = NULL,
+        .get_window_id = NULL,
+        .destroy = gsr_capture_ximage_destroy,
+        .priv = cap_ximage
+    };
+
+    return cap;
+}

src/color_conversion.c

@@ -1,25 +1,19 @@
 #include "../include/color_conversion.h"
 #include "../include/egl.h"
 #include <stdio.h>
-#include <string.h>
 #include <math.h>
+#include <string.h>
 #include <assert.h>
 
-/* TODO: highp instead of mediump? */
+#define SHADER_INDEX_Y            0
+#define SHADER_INDEX_UV           1
+#define SHADER_INDEX_Y_EXTERNAL   2
+#define SHADER_INDEX_UV_EXTERNAL  3
+#define SHADER_INDEX_RGB          4
+#define SHADER_INDEX_RGB_EXTERNAL 5
 
-#define MAX_SHADERS 4
-#define MAX_FRAMEBUFFERS 2
-
-static float abs_f(float v) {
-    return v >= 0.0f ? v : -v;
-}
-
-#define ROTATE_Z   "mat4 rotate_z(in float angle) {\n"                        \
-                   "    return mat4(cos(angle), -sin(angle), 0.0, 0.0,\n"     \
-                   "                sin(angle),  cos(angle), 0.0, 0.0,\n"     \
-                   "                0.0,           0.0,      1.0, 0.0,\n"     \
-                   "                0.0,           0.0,      0.0, 1.0);\n"    \
-                   "}\n"
+// TODO: Scissor doesn't work with compute shader. In the compute shader this can be implemented with two step calls, and using the result
+// with a call to mix to choose source/output color.
 
 /* https://en.wikipedia.org/wiki/YCbCr, see study/color_space_transform_matrix.png */
 
@@ -49,6 +43,10 @@ static float abs_f(float v) {
                             "                           0.060118,  0.429412, -0.038049, 0.000000,\n" \
                             "                           0.062745,  0.500000,  0.500000, 1.000000);\n"
 
+static int max_int(int a, int b) {
+    return a > b ? a : b;
+}
+
 static const char* color_format_range_get_transform_matrix(gsr_destination_color color_format, gsr_color_range color_range) {
     switch(color_format) {
         case GSR_DESTINATION_COLOR_NV12: {
@@ -69,138 +67,139 @@ static const char* color_format_range_get_transform_matrix(gsr_destination_color
             }
             break;
         }
+        case GSR_DESTINATION_COLOR_RGB8:
+            return "";
         default:
             return NULL;
     }
     return NULL;
 }
 
-static int load_shader_y(gsr_shader *shader, gsr_egl *egl, gsr_color_uniforms *uniforms, gsr_destination_color color_format, gsr_color_range color_range, bool external_texture) {
+// TODO: Make alpha blending optional
+// TODO: Optimize these shaders.
+static int load_compute_shader_y(gsr_shader *shader, gsr_egl *egl, gsr_color_uniforms *uniforms, int max_local_size_dim, gsr_destination_color color_format, gsr_color_range color_range, bool external_texture) {
     const char *color_transform_matrix = color_format_range_get_transform_matrix(color_format, color_range);
 
-    char vertex_shader[2048];
-    snprintf(vertex_shader, sizeof(vertex_shader),
-        "#version 300 es                                   \n"
-        "in vec2 pos;                                      \n"
-        "in vec2 texcoords;                                \n"
-        "out vec2 texcoords_out;                           \n"
-        "uniform vec2 offset;                              \n"
-        "uniform float rotation;                           \n"
-        ROTATE_Z
-        "void main()                                       \n"
-        "{                                                 \n"
-        "  texcoords_out = (vec4(texcoords.x - 0.5, texcoords.y - 0.5, 0.0, 0.0) * rotate_z(rotation)).xy + vec2(0.5, 0.5);  \n"
-        "  gl_Position = vec4(offset.x, offset.y, 0.0, 0.0) + vec4(pos.x, pos.y, 0.0, 1.0);    \n"
-        "}                                                 \n");
+    char compute_shader[2048];
+    snprintf(compute_shader, sizeof(compute_shader),
+        "#version 310 es\n"
+        "#extension GL_OES_EGL_image_external : enable\n"
+        "#extension GL_OES_EGL_image_external_essl3 : require\n"
+        "precision highp float;\n"
+        "layout (local_size_x = %d, local_size_y = %d, local_size_z = 1) in;\n"
+        "layout(binding = 0) uniform highp %s img_input;\n"
+        "layout(binding = 1) uniform highp sampler2D img_background;\n"
+        "uniform ivec2 source_position;\n"
+        "uniform ivec2 target_position;\n"
+        "uniform vec2 scale;\n"
+        "uniform mat2 rotation_matrix;\n"
+        "layout(rgba8, binding = 0) writeonly uniform highp image2D img_output;\n"
+        "%s"
+        "void main() {\n"
+        "    ivec2 texel_coord = ivec2(gl_GlobalInvocationID.xy);\n"
+        "    ivec2 size = ivec2(vec2(textureSize(img_input, 0)) * scale + 0.5);\n"
+        "    ivec2 output_size = textureSize(img_background, 0);\n"
+        "    vec2 rotated_texel_coord = vec2(texel_coord - source_position - size/2) * rotation_matrix + vec2(size/2) + 0.5;\n"
+        "    vec2 tex_coord = vec2(rotated_texel_coord)/vec2(size);\n"
+        "    vec4 source_color = texture(img_input, tex_coord);\n"
+        "    vec4 source_color_yuv = RGBtoYUV * vec4(source_color.rgb, 1.0);\n"
+        "    vec4 output_color_yuv = texture(img_background, (rotated_texel_coord + vec2(target_position))/vec2(output_size));\n"
+        "    float y_color = mix(output_color_yuv.r, source_color_yuv.r, source_color.a);\n"
+        "    imageStore(img_output, texel_coord + target_position, vec4(y_color, 1.0, 1.0, 1.0));\n"
+        "}\n", max_local_size_dim, max_local_size_dim, external_texture ? "samplerExternalOES" : "sampler2D", color_transform_matrix);
 
-    char fragment_shader[2048];
-    if(external_texture) {
-        snprintf(fragment_shader, sizeof(fragment_shader),
-            "#version 300 es                                                                 \n"
-            "#extension GL_OES_EGL_image_external : enable                                   \n"
-            "#extension GL_OES_EGL_image_external_essl3 : require                            \n"
-            "precision mediump float;                                                        \n"
-            "in vec2 texcoords_out;                                                          \n"
-            "uniform samplerExternalOES tex1;                                                \n"
-            "out vec4 FragColor;                                                             \n"
-            "%s"
-            "void main()                                                                     \n"
-            "{                                                                               \n"
-            "  vec4 pixel = texture(tex1, texcoords_out);                                    \n"
-            "  FragColor.x = (RGBtoYUV * vec4(pixel.rgb, 1.0)).x;                            \n"
-            "  FragColor.w = pixel.a;                                                        \n"
-            "}                                                                               \n", color_transform_matrix);
-    } else {
-        snprintf(fragment_shader, sizeof(fragment_shader),
-            "#version 300 es                                                                 \n"
-            "precision mediump float;                                                        \n"
-            "in vec2 texcoords_out;                                                          \n"
-            "uniform sampler2D tex1;                                                         \n"
-            "out vec4 FragColor;                                                             \n"
-            "%s"
-            "void main()                                                                     \n"
-            "{                                                                               \n"
-            "  vec4 pixel = texture(tex1, texcoords_out);                                    \n"
-            "  FragColor.x = (RGBtoYUV * vec4(pixel.rgb, 1.0)).x;                            \n"
-            "  FragColor.w = pixel.a;                                                        \n"
-            "}                                                                               \n", color_transform_matrix);
-    }
-
-    if(gsr_shader_init(shader, egl, vertex_shader, fragment_shader) != 0)
+    if(gsr_shader_init(shader, egl, NULL, NULL, compute_shader) != 0)
         return -1;
 
-    gsr_shader_bind_attribute_location(shader, "pos", 0);
-    gsr_shader_bind_attribute_location(shader, "texcoords", 1);
-    uniforms->offset = egl->glGetUniformLocation(shader->program_id, "offset");
-    uniforms->rotation = egl->glGetUniformLocation(shader->program_id, "rotation");
+    uniforms->source_position = egl->glGetUniformLocation(shader->program_id, "source_position");
+    uniforms->target_position = egl->glGetUniformLocation(shader->program_id, "target_position");
+    uniforms->rotation_matrix = egl->glGetUniformLocation(shader->program_id, "rotation_matrix");
+    uniforms->scale = egl->glGetUniformLocation(shader->program_id, "scale");
     return 0;
 }
 
-static unsigned int load_shader_uv(gsr_shader *shader, gsr_egl *egl, gsr_color_uniforms *uniforms, gsr_destination_color color_format, gsr_color_range color_range, bool external_texture) {
+static int load_compute_shader_uv(gsr_shader *shader, gsr_egl *egl, gsr_color_uniforms *uniforms, int max_local_size_dim, gsr_destination_color color_format, gsr_color_range color_range, bool external_texture) {
     const char *color_transform_matrix = color_format_range_get_transform_matrix(color_format, color_range);
 
-    char vertex_shader[2048];
-    snprintf(vertex_shader, sizeof(vertex_shader),
-        "#version 300 es                                 \n"
-        "in vec2 pos;                                    \n"
-        "in vec2 texcoords;                              \n"
-        "out vec2 texcoords_out;                         \n"
-        "uniform vec2 offset;                              \n"
-        "uniform float rotation;                         \n"
-        ROTATE_Z
-        "void main()                                     \n"
-        "{                                               \n"
-        "  texcoords_out = (vec4(texcoords.x - 0.5, texcoords.y - 0.5, 0.0, 0.0) * rotate_z(rotation)).xy + vec2(0.5, 0.5);                      \n"
-        "  gl_Position = (vec4(offset.x, offset.y, 0.0, 0.0) + vec4(pos.x, pos.y, 0.0, 1.0)) * vec4(0.5, 0.5, 1.0, 1.0) - vec4(0.5, 0.5, 0.0, 0.0);   \n"
-        "}                                               \n");
+    char compute_shader[2048];
+    snprintf(compute_shader, sizeof(compute_shader),
+        "#version 310 es\n"
+        "#extension GL_OES_EGL_image_external : enable\n"
+        "#extension GL_OES_EGL_image_external_essl3 : require\n"
+        "precision highp float;\n"
+        "layout (local_size_x = %d, local_size_y = %d, local_size_z = 1) in;\n"
+        "layout(binding = 0) uniform highp %s img_input;\n"
+        "layout(binding = 1) uniform highp sampler2D img_background;\n"
+        "uniform ivec2 source_position;\n"
+        "uniform ivec2 target_position;\n"
+        "uniform vec2 scale;\n"
+        "uniform mat2 rotation_matrix;\n"
+        "layout(rgba8, binding = 0) writeonly uniform highp image2D img_output;\n"
+        "%s"
+        "void main() {\n"
+        "    ivec2 texel_coord = ivec2(gl_GlobalInvocationID.xy);\n"
+        "    ivec2 size = ivec2(vec2(textureSize(img_input, 0)) * scale + 0.5);\n"
+        "    ivec2 output_size = textureSize(img_background, 0);\n"
+        "    vec2 rotated_texel_coord = vec2(texel_coord - source_position/2 - size/4) * rotation_matrix + vec2(size/4) + 0.5;\n"
+        "    vec2 tex_coord = vec2(rotated_texel_coord)/vec2(size);\n"
+        "    vec4 source_color = texture(img_input, tex_coord * 2.0);\n"
+        "    vec4 source_color_yuv = RGBtoYUV * vec4(source_color.rgb, 1.0);\n"
+        "    vec4 output_color_yuv = texture(img_background, (rotated_texel_coord + vec2(target_position/2))/vec2(output_size));\n"
+        "    vec2 uv_color = mix(output_color_yuv.rg, source_color_yuv.gb, source_color.a);\n"
+        "    imageStore(img_output, texel_coord + target_position/2, vec4(uv_color, 1.0, 1.0));\n"
+        "}\n", max_local_size_dim, max_local_size_dim, external_texture ? "samplerExternalOES" : "sampler2D", color_transform_matrix);
 
-    char fragment_shader[2048];
-    if(external_texture) {
-        snprintf(fragment_shader, sizeof(fragment_shader),
-            "#version 300 es                                                                       \n"
-            "#extension GL_OES_EGL_image_external : enable                                         \n"
-            "#extension GL_OES_EGL_image_external_essl3 : require                                  \n"
-            "precision mediump float;                                                              \n"
-            "in vec2 texcoords_out;                                                                \n"
-            "uniform samplerExternalOES tex1;                                                      \n"
-            "out vec4 FragColor;                                                                   \n"
-            "%s"
-            "void main()                                                                           \n"
-            "{                                                                                     \n"
-            "  vec4 pixel = texture(tex1, texcoords_out);                                          \n"
-            "  FragColor.xy = (RGBtoYUV * vec4(pixel.rgb, 1.0)).yz;                                \n"
-            "  FragColor.w = pixel.a;                                                              \n"
-            "}                                                                                     \n", color_transform_matrix);
-    } else {
-        snprintf(fragment_shader, sizeof(fragment_shader),
-            "#version 300 es                                                                       \n"
-            "precision mediump float;                                                              \n"
-            "in vec2 texcoords_out;                                                                \n"
-            "uniform sampler2D tex1;                                                               \n"
-            "out vec4 FragColor;                                                                   \n"
-            "%s"
-            "void main()                                                                           \n"
-            "{                                                                                     \n"
-            "  vec4 pixel = texture(tex1, texcoords_out);                                          \n"
-            "  FragColor.xy = (RGBtoYUV * vec4(pixel.rgb, 1.0)).yz;                                \n"
-            "  FragColor.w = pixel.a;                                                              \n"
-            "}                                                                                     \n", color_transform_matrix);
-    }
-
-    if(gsr_shader_init(shader, egl, vertex_shader, fragment_shader) != 0)
+    if(gsr_shader_init(shader, egl, NULL, NULL, compute_shader) != 0)
         return -1;
 
-    gsr_shader_bind_attribute_location(shader, "pos", 0);
-    gsr_shader_bind_attribute_location(shader, "texcoords", 1);
-    uniforms->offset = egl->glGetUniformLocation(shader->program_id, "offset");
-    uniforms->rotation = egl->glGetUniformLocation(shader->program_id, "rotation");
+    uniforms->source_position = egl->glGetUniformLocation(shader->program_id, "source_position");
+    uniforms->target_position = egl->glGetUniformLocation(shader->program_id, "target_position");
+    uniforms->rotation_matrix = egl->glGetUniformLocation(shader->program_id, "rotation_matrix");
+    uniforms->scale = egl->glGetUniformLocation(shader->program_id, "scale");
+    return 0;
+}
+
+static int load_compute_shader_rgb(gsr_shader *shader, gsr_egl *egl, gsr_color_uniforms *uniforms, int max_local_size_dim, bool external_texture) {
+    char compute_shader[2048];
+    snprintf(compute_shader, sizeof(compute_shader),
+        "#version 310 es\n"
+        "#extension GL_OES_EGL_image_external : enable\n"
+        "#extension GL_OES_EGL_image_external_essl3 : require\n"
+        "precision highp float;\n"
+        "layout (local_size_x = %d, local_size_y = %d, local_size_z = 1) in;\n"
+        "layout(binding = 0) uniform highp %s img_input;\n"
+        "layout(binding = 1) uniform highp sampler2D img_background;\n"
+        "uniform ivec2 source_position;\n"
+        "uniform ivec2 target_position;\n"
+        "uniform vec2 scale;\n"
+        "uniform mat2 rotation_matrix;\n"
+        "layout(rgba8, binding = 0) writeonly uniform highp image2D img_output;\n"
+        "void main() {\n"
+        "    ivec2 texel_coord = ivec2(gl_GlobalInvocationID.xy);\n"
+        "    ivec2 size = ivec2(vec2(textureSize(img_input, 0)) * scale + 0.5);\n"
+        "    ivec2 output_size = textureSize(img_background, 0);\n"
+        "    vec2 rotated_texel_coord = vec2(texel_coord - source_position - size/2) * rotation_matrix + vec2(size/2) + 0.5;\n"
+        "    vec2 tex_coord = vec2(rotated_texel_coord)/vec2(size);\n"
+        "    vec4 source_color = texture(img_input, tex_coord);\n"
+        "    vec4 output_color = texture(img_background, (rotated_texel_coord + vec2(target_position))/vec2(output_size));\n"
+        "    vec3 color = mix(output_color.rgb, source_color.rgb, source_color.a);\n"
+        "    imageStore(img_output, texel_coord + target_position, vec4(color, 1.0));\n"
+        "}\n", max_local_size_dim, max_local_size_dim, external_texture ? "samplerExternalOES" : "sampler2D");
+
+    if(gsr_shader_init(shader, egl, NULL, NULL, compute_shader) != 0)
+        return -1;
+
+    uniforms->source_position = egl->glGetUniformLocation(shader->program_id, "source_position");
+    uniforms->target_position = egl->glGetUniformLocation(shader->program_id, "target_position");
+    uniforms->rotation_matrix = egl->glGetUniformLocation(shader->program_id, "rotation_matrix");
+    uniforms->scale = egl->glGetUniformLocation(shader->program_id, "scale");
     return 0;
 }
 
 static int load_framebuffers(gsr_color_conversion *self) {
     /* TODO: Only generate the necessary amount of framebuffers (self->params.num_destination_textures) */
     const unsigned int draw_buffer = GL_COLOR_ATTACHMENT0;
-    self->params.egl->glGenFramebuffers(MAX_FRAMEBUFFERS, self->framebuffers);
+    self->params.egl->glGenFramebuffers(GSR_COLOR_CONVERSION_MAX_FRAMEBUFFERS, self->framebuffers);
 
     self->params.egl->glBindFramebuffer(GL_FRAMEBUFFER, self->framebuffers[0]);
     self->params.egl->glFramebufferTexture2D(GL_FRAMEBUFFER, GL_COLOR_ATTACHMENT0, GL_TEXTURE_2D, self->params.destination_textures[0], 0);
@@ -252,6 +251,10 @@ int gsr_color_conversion_init(gsr_color_conversion *self, const gsr_color_conver
     memset(self, 0, sizeof(*self));
     self->params.egl = params->egl;
     self->params = *params;
+    
+    int max_compute_work_group_invocations = 256;
+    self->params.egl->glGetIntegerv(GL_MAX_COMPUTE_FIXED_GROUP_INVOCATIONS, &max_compute_work_group_invocations);
+    self->max_local_size_dim = sqrt(max_compute_work_group_invocations);
 
     switch(params->destination_color) {
         case GSR_DESTINATION_COLOR_NV12:
@@ -261,24 +264,43 @@ int gsr_color_conversion_init(gsr_color_conversion *self, const gsr_color_conver
                 return -1;
             }
 
-            if(load_shader_y(&self->shaders[0], self->params.egl, &self->uniforms[0], params->destination_color, params->color_range, false) != 0) {
-                fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load Y shader\n");
+            if(load_compute_shader_y(&self->shaders[SHADER_INDEX_Y], self->params.egl, &self->uniforms[SHADER_INDEX_Y], self->max_local_size_dim, params->destination_color, params->color_range, false) != 0) {
+                fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load Y compute shader\n");
                 goto err;
             }
 
-            if(load_shader_uv(&self->shaders[1], self->params.egl, &self->uniforms[1], params->destination_color, params->color_range, false) != 0) {
-                fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load UV shader\n");
+            if(load_compute_shader_uv(&self->shaders[SHADER_INDEX_UV], self->params.egl, &self->uniforms[SHADER_INDEX_UV], self->max_local_size_dim, params->destination_color, params->color_range, false) != 0) {
+                fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load UV compute shader\n");
                 goto err;
             }
 
             if(self->params.load_external_image_shader) {
-                if(load_shader_y(&self->shaders[2], self->params.egl, &self->uniforms[2], params->destination_color, params->color_range, true) != 0) {
-                    fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load Y shader\n");
+                if(load_compute_shader_y(&self->shaders[SHADER_INDEX_Y_EXTERNAL], self->params.egl, &self->uniforms[SHADER_INDEX_Y_EXTERNAL], self->max_local_size_dim, params->destination_color, params->color_range, true) != 0) {
+                    fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load Y compute shader\n");
                     goto err;
                 }
 
-                if(load_shader_uv(&self->shaders[3], self->params.egl, &self->uniforms[3], params->destination_color, params->color_range, true) != 0) {
-                    fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load UV shader\n");
+                if(load_compute_shader_uv(&self->shaders[SHADER_INDEX_UV_EXTERNAL], self->params.egl, &self->uniforms[SHADER_INDEX_UV_EXTERNAL], self->max_local_size_dim, params->destination_color, params->color_range, true) != 0) {
+                    fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load UV compute shader\n");
+                    goto err;
+                }
+            }
+            break;
+        }
+        case GSR_DESTINATION_COLOR_RGB8: {
+            if(self->params.num_destination_textures != 1) {
+                fprintf(stderr, "gsr error: gsr_color_conversion_init: expected 1 destination textures for destination color RGB8, got %d destination texture(s)\n", self->params.num_destination_textures);
+                return -1;
+            }
+
+            if(load_compute_shader_rgb(&self->shaders[SHADER_INDEX_RGB], self->params.egl, &self->uniforms[SHADER_INDEX_RGB], self->max_local_size_dim, false) != 0) {
+                fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load Y compute shader\n");
+                goto err;
+            }
+
+            if(self->params.load_external_image_shader) {
+                if(load_compute_shader_rgb(&self->shaders[SHADER_INDEX_RGB_EXTERNAL], self->params.egl, &self->uniforms[SHADER_INDEX_RGB_EXTERNAL], self->max_local_size_dim, true) != 0) {
+                    fprintf(stderr, "gsr error: gsr_color_conversion_init: failed to load Y compute shader\n");
                     goto err;
                 }
             }
@@ -313,18 +335,59 @@ void gsr_color_conversion_deinit(gsr_color_conversion *self) {
         self->vertex_array_object_id = 0;
     }
 
-    self->params.egl->glDeleteFramebuffers(MAX_FRAMEBUFFERS, self->framebuffers);
-    for(int i = 0; i < MAX_FRAMEBUFFERS; ++i) {
+    self->params.egl->glDeleteFramebuffers(GSR_COLOR_CONVERSION_MAX_FRAMEBUFFERS, self->framebuffers);
+    for(int i = 0; i < GSR_COLOR_CONVERSION_MAX_FRAMEBUFFERS; ++i) {
         self->framebuffers[i] = 0;
     }
 
-    for(int i = 0; i < MAX_SHADERS; ++i) {
+    for(int i = 0; i < GSR_COLOR_CONVERSION_MAX_SHADERS; ++i) {
         gsr_shader_deinit(&self->shaders[i]);
     }
 
     self->params.egl = NULL;
 }
 
+static void gsr_color_conversion_apply_rotation(gsr_rotation rotation, float rotation_matrix[2][2], vec2i *source_position, vec2i texture_size, vec2f scale) {
+    /*
+    rotation_matrix[0][0] =  cos(angle);
+    rotation_matrix[0][1] = -sin(angle);
+    rotation_matrix[1][0] =  sin(angle);
+    rotation_matrix[1][1] =  cos(angle);
+    The manual matrix code below is the same as this code above, but without floating-point errors.
+    This is done to remove any blurring caused by these floating-point errors.
+    */
+    switch(rotation) {
+        case GSR_ROT_0:
+            rotation_matrix[0][0] = 1.0f;
+            rotation_matrix[0][1] = 0.0f;
+            rotation_matrix[1][0] = 0.0f;
+            rotation_matrix[1][1] = 1.0f;
+            break;
+        case GSR_ROT_90:
+            rotation_matrix[0][0] =  0.0f;
+            rotation_matrix[0][1] = -1.0f;
+            rotation_matrix[1][0] =  1.0f;
+            rotation_matrix[1][1] =  0.0f;
+            source_position->x += (((double)texture_size.x*0.5 - (double)texture_size.y*0.5) * scale.x + 0.5);
+            source_position->y += (((double)texture_size.y*0.5 - (double)texture_size.x*0.5) * scale.y + 0.5);
+            break;
+        case GSR_ROT_180:
+            rotation_matrix[0][0] = -1.0f;
+            rotation_matrix[0][1] =  0.0f;
+            rotation_matrix[1][0] =  0.0f;
+            rotation_matrix[1][1] = -1.0f;
+            break;
+        case GSR_ROT_270:
+            rotation_matrix[0][0] =  0.0f;
+            rotation_matrix[0][1] =  1.0f;
+            rotation_matrix[1][0] = -1.0f;
+            rotation_matrix[1][1] =  0.0f;
+            source_position->x += (((double)texture_size.x*0.5 - (double)texture_size.y*0.5) * scale.x + 0.5);
+            source_position->y += (((double)texture_size.y*0.5 - (double)texture_size.x*0.5) * scale.y + 0.5);
+            break;
+    }
+}
+
 static void gsr_color_conversion_swizzle_texture_source(gsr_color_conversion *self, gsr_source_color source_color) {
     if(source_color == GSR_SOURCE_COLOR_BGR) {
         const int swizzle_mask[] = { GL_BLUE, GL_GREEN, GL_RED, 1 };
@@ -339,106 +402,99 @@ static void gsr_color_conversion_swizzle_reset(gsr_color_conversion *self, gsr_s
     }
 }
 
-/* |source_pos| is in pixel coordinates and |source_size|  */
-void gsr_color_conversion_draw(gsr_color_conversion *self, unsigned int texture_id, vec2i source_pos, vec2i source_size, vec2i texture_pos, vec2i texture_size, float rotation, bool external_texture, gsr_source_color source_color) {
-    // TODO: Remove this crap
-    rotation = M_PI*2.0f - rotation;
+typedef enum {
+    GSR_COLOR_COMP_Y,
+    GSR_COLOR_COMP_UV,
+    GSR_COLOR_COMP_RGB
+} gsr_color_component;
 
-    /* TODO: Do not call this every frame? */
-    vec2i dest_texture_size = {0, 0};
-    self->params.egl->glBindTexture(GL_TEXTURE_2D, self->params.destination_textures[0]);
-    self->params.egl->glGetTexLevelParameteriv(GL_TEXTURE_2D, 0, GL_TEXTURE_WIDTH, &dest_texture_size.x);
-    self->params.egl->glGetTexLevelParameteriv(GL_TEXTURE_2D, 0, GL_TEXTURE_HEIGHT, &dest_texture_size.y);
-    self->params.egl->glBindTexture(GL_TEXTURE_2D, 0);
+static int color_component_get_texture_index(gsr_color_component color_component) {
+    switch(color_component) {
+        case GSR_COLOR_COMP_Y:   return 0;
+        case GSR_COLOR_COMP_UV:  return 1;
+        case GSR_COLOR_COMP_RGB: return 0;
+    }
+    assert(false);
+    return 0;
+}
+
+static unsigned int color_component_get_color_format(gsr_color_component color_component, bool use_16bit_colors) {
+    switch(color_component) {
+        case GSR_COLOR_COMP_Y:   return use_16bit_colors ? GL_R16 : GL_R8;
+        case GSR_COLOR_COMP_UV:  return use_16bit_colors ? GL_RG16 : GL_RG8;
+        case GSR_COLOR_COMP_RGB: return GL_RGBA8; // TODO: 16-bit color support
+    }
+    assert(false);
+    return GL_RGBA8;
+}
+
+static int color_component_get_shader_index(gsr_color_component color_component, bool external_texture) {
+    switch(color_component) {
+        case GSR_COLOR_COMP_Y:   return external_texture ? SHADER_INDEX_Y_EXTERNAL : SHADER_INDEX_Y;
+        case GSR_COLOR_COMP_UV:  return external_texture ? SHADER_INDEX_UV_EXTERNAL : SHADER_INDEX_UV;
+        case GSR_COLOR_COMP_RGB: return external_texture ? SHADER_INDEX_RGB_EXTERNAL : SHADER_INDEX_RGB;
+    }
+    assert(false);
+    return SHADER_INDEX_RGB;
+}
+
+static void gsr_color_conversion_dispatch_compute_shader(gsr_color_conversion *self, bool external_texture, float rotation_matrix[2][2], vec2i source_position, vec2i destination_pos, vec2i destination_size, vec2f scale, bool use_16bit_colors, gsr_color_component color_component) {
+    const int shader_index = color_component_get_shader_index(color_component, external_texture);
+    const int texture_index = color_component_get_texture_index(color_component);
+    const unsigned int color_format = color_component_get_color_format(color_component, use_16bit_colors);
+
+    self->params.egl->glActiveTexture(GL_TEXTURE1);
+    self->params.egl->glBindTexture(GL_TEXTURE_2D, self->params.destination_textures[texture_index]);
+    self->params.egl->glActiveTexture(GL_TEXTURE0);
+
+    gsr_color_uniforms *uniform = &self->uniforms[shader_index];
+    gsr_shader_use(&self->shaders[shader_index]);
+    self->params.egl->glUniformMatrix2fv(uniform->rotation_matrix, 1, GL_TRUE, (const float*)rotation_matrix);
+    self->params.egl->glUniform2i(uniform->source_position, source_position.x, source_position.y);
+    self->params.egl->glUniform2i(uniform->target_position, destination_pos.x, destination_pos.y);
+    self->params.egl->glUniform2f(uniform->scale, scale.x, scale.y);
+    self->params.egl->glBindImageTexture(0, self->params.destination_textures[texture_index], 0, GL_FALSE, 0, GL_WRITE_ONLY, color_format);
+    const double num_groups_x = ceil((double)destination_size.x/(double)self->max_local_size_dim);
+    const double num_groups_y = ceil((double)destination_size.y/(double)self->max_local_size_dim);
+    self->params.egl->glDispatchCompute(max_int(1, num_groups_x), max_int(1, num_groups_y), 1);
+}
+
+void gsr_color_conversion_draw(gsr_color_conversion *self, unsigned int texture_id, vec2i destination_pos, vec2i destination_size, vec2i texture_pos, vec2i texture_size, gsr_rotation rotation, bool external_texture, gsr_source_color source_color) {
+    vec2f scale = {0.0f, 0.0f};
+    if(texture_size.x > 0 && texture_size.y > 0)
+        scale = (vec2f){ (double)destination_size.x/(double)texture_size.x, (double)destination_size.y/(double)texture_size.y };
+
+    vec2i source_position = {0, 0};
+    float rotation_matrix[2][2] = {{0, 0}, {0, 0}};
+    gsr_color_conversion_apply_rotation(rotation, rotation_matrix, &source_position, texture_size, scale);
+
+    source_position.x -= (texture_pos.x * scale.x + 0.5);
+    source_position.y -= (texture_pos.y * scale.y + 0.5);
 
     const int texture_target = external_texture ? GL_TEXTURE_EXTERNAL_OES : GL_TEXTURE_2D;
-
     self->params.egl->glBindTexture(texture_target, texture_id);
-
-    vec2i source_texture_size = {0, 0};
-    if(external_texture) {
-        assert(self->params.load_external_image_shader);
-        source_texture_size = source_size;
-    } else {
-        /* TODO: Do not call this every frame? */
-        self->params.egl->glGetTexLevelParameteriv(texture_target, 0, GL_TEXTURE_WIDTH, &source_texture_size.x);
-        self->params.egl->glGetTexLevelParameteriv(texture_target, 0, GL_TEXTURE_HEIGHT, &source_texture_size.y);
-    }
-
-    // TODO: Remove this crap
-    if(abs_f(M_PI * 0.5f - rotation) <= 0.001f || abs_f(M_PI * 1.5f - rotation) <= 0.001f) {
-        float tmp = source_texture_size.x;
-        source_texture_size.x = source_texture_size.y;
-        source_texture_size.y = tmp;
-    }
-
-    const vec2f pos_norm = {
-        ((float)source_pos.x / (dest_texture_size.x == 0 ? 1.0f : (float)dest_texture_size.x)) * 2.0f,
-        ((float)source_pos.y / (dest_texture_size.y == 0 ? 1.0f : (float)dest_texture_size.y)) * 2.0f,
-    };
-
-    const vec2f size_norm = {
-        ((float)source_size.x / (dest_texture_size.x == 0 ? 1.0f : (float)dest_texture_size.x)) * 2.0f,
-        ((float)source_size.y / (dest_texture_size.y == 0 ? 1.0f : (float)dest_texture_size.y)) * 2.0f,
-    };
-
-    const vec2f texture_pos_norm = {
-        (float)texture_pos.x / (source_texture_size.x == 0 ? 1.0f : (float)source_texture_size.x),
-        (float)texture_pos.y / (source_texture_size.y == 0 ? 1.0f : (float)source_texture_size.y),
-    };
-
-    const vec2f texture_size_norm = {
-        (float)texture_size.x / (source_texture_size.x == 0 ? 1.0f : (float)source_texture_size.x),
-        (float)texture_size.y / (source_texture_size.y == 0 ? 1.0f : (float)source_texture_size.y),
-    };
-
-    const float vertices[] = {
-        -1.0f + 0.0f,               -1.0f + 0.0f + size_norm.y, texture_pos_norm.x,                       texture_pos_norm.y + texture_size_norm.y,
-        -1.0f + 0.0f,               -1.0f + 0.0f,               texture_pos_norm.x,                       texture_pos_norm.y,
-        -1.0f + 0.0f + size_norm.x, -1.0f + 0.0f,               texture_pos_norm.x + texture_size_norm.x, texture_pos_norm.y,
-
-        -1.0f + 0.0f,               -1.0f + 0.0f + size_norm.y, texture_pos_norm.x,                       texture_pos_norm.y + texture_size_norm.y,
-        -1.0f + 0.0f + size_norm.x, -1.0f + 0.0f,               texture_pos_norm.x + texture_size_norm.x, texture_pos_norm.y,
-        -1.0f + 0.0f + size_norm.x, -1.0f + 0.0f + size_norm.y, texture_pos_norm.x + texture_size_norm.x, texture_pos_norm.y + texture_size_norm.y
-    };
-
     gsr_color_conversion_swizzle_texture_source(self, source_color);
 
-    self->params.egl->glBindVertexArray(self->vertex_array_object_id);
-    self->params.egl->glViewport(0, 0, dest_texture_size.x, dest_texture_size.y);
-
-    /* TODO: this, also cleanup */
-    //self->params.egl->glBindBuffer(GL_ARRAY_BUFFER, self->vertex_buffer_object_id);
-    self->params.egl->glBufferSubData(GL_ARRAY_BUFFER, 0, 24 * sizeof(float), vertices);
-
-    {
-        self->params.egl->glBindFramebuffer(GL_FRAMEBUFFER, self->framebuffers[0]);
-        //cap_xcomp->params.egl->glClear(GL_COLOR_BUFFER_BIT); // TODO: Do this in a separate clear_ function. We want to do that when using multiple drm to create the final image (multiple monitors for example)
-
-        const int shader_index = external_texture ? 2 : 0;
-        gsr_shader_use(&self->shaders[shader_index]);
-        self->params.egl->glUniform1f(self->uniforms[shader_index].rotation, rotation);
-        self->params.egl->glUniform2f(self->uniforms[shader_index].offset, pos_norm.x, pos_norm.y);
-        self->params.egl->glDrawArrays(GL_TRIANGLES, 0, 6);
+    switch(self->params.destination_color) {
+        case GSR_DESTINATION_COLOR_NV12:
+        case GSR_DESTINATION_COLOR_P010: {
+            const bool use_16bit_colors = self->params.destination_color == GSR_DESTINATION_COLOR_P010;
+            gsr_color_conversion_dispatch_compute_shader(self, external_texture, rotation_matrix, source_position, destination_pos, destination_size, scale, use_16bit_colors, GSR_COLOR_COMP_Y);
+            gsr_color_conversion_dispatch_compute_shader(self, external_texture, rotation_matrix, source_position, destination_pos, (vec2i){destination_size.x/2, destination_size.y/2}, scale, use_16bit_colors, GSR_COLOR_COMP_UV);
+            break;
+        }
+        case GSR_DESTINATION_COLOR_RGB8: {
+            gsr_color_conversion_dispatch_compute_shader(self, external_texture, rotation_matrix, source_position, destination_pos, destination_size, scale, false, GSR_COLOR_COMP_RGB);
+            break;
+        }
     }
 
-    if(self->params.num_destination_textures > 1) {
-        self->params.egl->glBindFramebuffer(GL_FRAMEBUFFER, self->framebuffers[1]);
-        //cap_xcomp->params.egl->glClear(GL_COLOR_BUFFER_BIT);
-
-        const int shader_index = external_texture ? 3 : 1;
-        gsr_shader_use(&self->shaders[shader_index]);
-        self->params.egl->glUniform1f(self->uniforms[shader_index].rotation, rotation);
-        self->params.egl->glUniform2f(self->uniforms[shader_index].offset, pos_norm.x, pos_norm.y);
-        self->params.egl->glDrawArrays(GL_TRIANGLES, 0, 6);
-    }
-
-    self->params.egl->glBindVertexArray(0);
-    gsr_shader_use_none(&self->shaders[0]);
-    self->params.egl->glBindTexture(texture_target, 0);
-    self->params.egl->glBindFramebuffer(GL_FRAMEBUFFER, 0);
+    // TODO: Use the minimal barrier required
+    self->params.egl->glMemoryBarrier(GL_ALL_BARRIER_BITS); // GL_SHADER_IMAGE_ACCESS_BARRIER_BIT
+    self->params.egl->glUseProgram(0);
 
     gsr_color_conversion_swizzle_reset(self, source_color);
+    self->params.egl->glBindTexture(texture_target, 0);
 }
 
 void gsr_color_conversion_clear(gsr_color_conversion *self) {
@@ -454,6 +510,13 @@ void gsr_color_conversion_clear(gsr_color_conversion *self) {
             color2[3] = 1.0f;
             break;
         }
+        case GSR_DESTINATION_COLOR_RGB8: {
+            color2[0] = 0.0f;
+            color2[1] = 0.0f;
+            color2[2] = 0.0f;
+            color2[3] = 1.0f;
+            break;
+        }
     }
 
     self->params.egl->glBindFramebuffer(GL_FRAMEBUFFER, self->framebuffers[0]);
@@ -468,3 +531,7 @@ void gsr_color_conversion_clear(gsr_color_conversion *self) {
 
     self->params.egl->glBindFramebuffer(GL_FRAMEBUFFER, 0);
 }
+
+gsr_rotation gsr_monitor_rotation_to_rotation(gsr_monitor_rotation monitor_rotation) {
+    return (gsr_rotation)monitor_rotation;
+}

src/cursor.c

@@ -56,8 +56,10 @@ static bool gsr_cursor_set_from_x11_cursor_image(gsr_cursor *self, XFixesCursorI
     self->egl->glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA8, self->size.x, self->size.y, 0, GL_RGBA, GL_UNSIGNED_BYTE, cursor_data);
     free(cursor_data);
 
-    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    const float border_color[4] = { 0.0f, 0.0f, 0.0f, 0.0f };
+    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->egl->glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, border_color);
     self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
 

src/dbus.c

@@ -614,9 +614,41 @@ int gsr_dbus_screencast_create_session(gsr_dbus *self, char **session_handle) {
     return 0;
 }
 
-int gsr_dbus_screencast_select_sources(gsr_dbus *self, const char *session_handle, gsr_portal_capture_type capture_type, gsr_portal_cursor_mode cursor_mode) {
+static uint32_t unset_unsupported_capture_types(uint32_t requested_capture_types, uint32_t available_capture_types) {
+    if(!(available_capture_types & GSR_PORTAL_CAPTURE_TYPE_MONITOR))
+        requested_capture_types &= ~GSR_PORTAL_CAPTURE_TYPE_MONITOR;
+    if(!(available_capture_types & GSR_PORTAL_CAPTURE_TYPE_WINDOW))
+        requested_capture_types &= ~GSR_PORTAL_CAPTURE_TYPE_WINDOW;
+    if(!(available_capture_types & GSR_PORTAL_CAPTURE_TYPE_VIRTUAL))
+        requested_capture_types &= ~GSR_PORTAL_CAPTURE_TYPE_VIRTUAL;
+    return requested_capture_types;
+}
+
+static uint32_t unset_unsupported_cursor_modes(uint32_t requested_cursor_modes, uint32_t available_cursor_modes) {
+    if(!(available_cursor_modes & GSR_PORTAL_CURSOR_MODE_HIDDEN))
+        requested_cursor_modes &= ~GSR_PORTAL_CURSOR_MODE_HIDDEN;
+    if(!(available_cursor_modes & GSR_PORTAL_CURSOR_MODE_EMBEDDED))
+        requested_cursor_modes &= ~GSR_PORTAL_CURSOR_MODE_EMBEDDED;
+    if(!(available_cursor_modes & GSR_PORTAL_CURSOR_MODE_METADATA))
+        requested_cursor_modes &= ~GSR_PORTAL_CURSOR_MODE_METADATA;
+    return requested_cursor_modes;
+}
+
+int gsr_dbus_screencast_select_sources(gsr_dbus *self, const char *session_handle, uint32_t capture_type, uint32_t cursor_mode) {
     assert(session_handle);
 
+    uint32_t available_source_types = 0;
+    gsr_dbus_desktop_portal_get_property(self, "org.freedesktop.portal.ScreenCast", "AvailableSourceTypes", &available_source_types);
+    if(available_source_types == 0)
+        fprintf(stderr, "gsr error: gsr_dbus_screencast_select_sources: no source types are available\n");
+    capture_type = unset_unsupported_capture_types(capture_type, available_source_types);
+
+    uint32_t available_cursor_modes = 0;
+    gsr_dbus_desktop_portal_get_property(self, "org.freedesktop.portal.ScreenCast", "AvailableCursorModes", &available_cursor_modes);
+    if(available_cursor_modes == 0)
+        fprintf(stderr, "gsr error: gsr_dbus_screencast_select_sources: no cursors modes are available\n");
+    cursor_mode = unset_unsupported_cursor_modes(cursor_mode, available_cursor_modes);
+
     char handle_token[64];
     gsr_dbus_portal_get_unique_handle_token(self, handle_token, sizeof(handle_token));
 

src/egl.c

@@ -283,15 +283,21 @@ static bool gsr_egl_load_gl(gsr_egl *self, void *library) {
         { (void**)&self->glClearColor, "glClearColor" },
         { (void**)&self->glGenTextures, "glGenTextures" },
         { (void**)&self->glDeleteTextures, "glDeleteTextures" },
+        { (void**)&self->glActiveTexture, "glActiveTexture" },
         { (void**)&self->glBindTexture, "glBindTexture" },
+        { (void**)&self->glBindImageTexture, "glBindImageTexture" },
         { (void**)&self->glTexParameteri, "glTexParameteri" },
         { (void**)&self->glTexParameteriv, "glTexParameteriv" },
+        { (void**)&self->glTexParameterfv, "glTexParameterfv" },
         { (void**)&self->glGetTexLevelParameteriv, "glGetTexLevelParameteriv" },
         { (void**)&self->glTexImage2D, "glTexImage2D" },
+        { (void**)&self->glTexSubImage2D, "glTexSubImage2D" },
         { (void**)&self->glGetTexImage, "glGetTexImage" },
         { (void**)&self->glGenFramebuffers, "glGenFramebuffers" },
         { (void**)&self->glBindFramebuffer, "glBindFramebuffer" },
         { (void**)&self->glDeleteFramebuffers, "glDeleteFramebuffers" },
+        { (void**)&self->glDispatchCompute, "glDispatchCompute" },
+        { (void**)&self->glMemoryBarrier, "glMemoryBarrier" },
         { (void**)&self->glViewport, "glViewport" },
         { (void**)&self->glFramebufferTexture2D, "glFramebufferTexture2D" },
         { (void**)&self->glDrawBuffers, "glDrawBuffers" },
@@ -324,14 +330,19 @@ static bool gsr_egl_load_gl(gsr_egl *self, void *library) {
         { (void**)&self->glEnable, "glEnable" },
         { (void**)&self->glDisable, "glDisable" },
         { (void**)&self->glBlendFunc, "glBlendFunc" },
+        { (void**)&self->glPixelStorei, "glPixelStorei" },
         { (void**)&self->glGetUniformLocation, "glGetUniformLocation" },
         { (void**)&self->glUniform1f, "glUniform1f" },
         { (void**)&self->glUniform2f, "glUniform2f" },
+        { (void**)&self->glUniform1i, "glUniform1i" },
+        { (void**)&self->glUniform2i, "glUniform2i" },
+        { (void**)&self->glUniformMatrix2fv, "glUniformMatrix2fv" },
         { (void**)&self->glDebugMessageCallback, "glDebugMessageCallback" },
         { (void**)&self->glScissor, "glScissor" },
         { (void**)&self->glReadPixels, "glReadPixels" },
         { (void**)&self->glMapBuffer, "glMapBuffer" },
         { (void**)&self->glUnmapBuffer, "glUnmapBuffer" },
+        { (void**)&self->glGetIntegerv, "glGetIntegerv" },
 
         { NULL, NULL }
     };
@@ -355,6 +366,44 @@ static void debug_callback(unsigned int source, unsigned int type, unsigned int
         fprintf(stderr, "gsr info: gl callback: %s type = 0x%x, severity = 0x%x, message = %s\n", type == GL_DEBUG_TYPE_ERROR ? "** GL ERROR **" : "", type, severity, message);
 }
 
+/* TODO: check for glx swap control extension string (GLX_EXT_swap_control, etc) */
+static void set_vertical_sync_enabled(gsr_egl *egl, int enabled) {
+    int result = 0;
+
+    if(egl->glXSwapIntervalEXT) {
+        assert(gsr_window_get_display_server(egl->window) == GSR_DISPLAY_SERVER_X11);
+        Display *display = gsr_window_get_display(egl->window);
+        const Window window = (Window)gsr_window_get_window(egl->window);
+        egl->glXSwapIntervalEXT(display, window, enabled ? 1 : 0);
+    } else if(egl->glXSwapIntervalMESA) {
+        result = egl->glXSwapIntervalMESA(enabled ? 1 : 0);
+    } else if(egl->glXSwapIntervalSGI) {
+        result = egl->glXSwapIntervalSGI(enabled ? 1 : 0);
+    } else {
+        static int warned = 0;
+        if (!warned) {
+            warned = 1;
+            fprintf(stderr, "gsr warning: setting vertical sync not supported\n");
+        }
+    }
+
+    if(result != 0)
+        fprintf(stderr, "gsr warning: setting vertical sync failed\n");
+}
+
+static void gsr_egl_disable_vsync(gsr_egl *self) {
+    switch(self->context_type) {
+        case GSR_GL_CONTEXT_TYPE_EGL: {
+            self->eglSwapInterval(self->egl_display, 0);
+            break;
+        }
+        case GSR_GL_CONTEXT_TYPE_GLX: {
+            set_vertical_sync_enabled(self, 0);
+            break;
+        }
+    }
+}
+
 bool gsr_egl_load(gsr_egl *self, gsr_window *window, bool is_monitor_capture, bool enable_debug) {
     memset(self, 0, sizeof(gsr_egl));
     self->context_type = GSR_GL_CONTEXT_TYPE_EGL;
@@ -410,12 +459,25 @@ bool gsr_egl_load(gsr_egl *self, gsr_window *window, bool is_monitor_capture, bo
 
     self->glEnable(GL_BLEND);
     self->glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA);
+    self->glPixelStorei(GL_PACK_ALIGNMENT, 1);
+    self->glPixelStorei(GL_UNPACK_ALIGNMENT, 1);
 
     if(enable_debug) {
         self->glEnable(GL_DEBUG_OUTPUT);
         self->glDebugMessageCallback(debug_callback, NULL);
     }
 
+    gsr_egl_disable_vsync(self);
+
+    if(self->gpu_info.vendor == GSR_GPU_VENDOR_NVIDIA) {
+        /* This fixes nvenc codecs unable to load on openSUSE tumbleweed because of a cuda error. Don't ask me why */
+        const bool inside_flatpak = getenv("FLATPAK_ID") != NULL;
+        if(inside_flatpak)
+            system("flatpak-spawn --host -- nvidia-smi -f /dev/null");
+        else
+            system("nvidia-smi -f /dev/null");
+    }
+
     return true;
 
     fail:

src/encoder/video/nvenc.c

@@ -65,21 +65,6 @@ static bool gsr_video_encoder_nvenc_setup_context(gsr_video_encoder_nvenc *self,
     return true;
 }
 
-static unsigned int gl_create_texture(gsr_egl *egl, int width, int height, int internal_format, unsigned int format) {
-    unsigned int texture_id = 0;
-    egl->glGenTextures(1, &texture_id);
-    egl->glBindTexture(GL_TEXTURE_2D, texture_id);
-    egl->glTexImage2D(GL_TEXTURE_2D, 0, internal_format, width, height, 0, format, GL_UNSIGNED_BYTE, NULL);
-
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
-
-    egl->glBindTexture(GL_TEXTURE_2D, 0);
-    return texture_id;
-}
-
 static bool cuda_register_opengl_texture(gsr_cuda *cuda, CUgraphicsResource *cuda_graphics_resource, CUarray *mapped_array, unsigned int texture_id) {
     CUresult res;
     res = cuda->cuGraphicsGLRegisterImage(cuda_graphics_resource, texture_id, GL_TEXTURE_2D, CU_GRAPHICS_REGISTER_FLAGS_NONE);
@@ -110,7 +95,7 @@ static bool gsr_video_encoder_nvenc_setup_textures(gsr_video_encoder_nvenc *self
     const int div[2] = {1, 2}; // divide UV texture size by 2 because chroma is half size
 
     for(int i = 0; i < 2; ++i) {
-        self->target_textures[i] = gl_create_texture(self->params.egl, video_codec_context->width / div[i], video_codec_context->height / div[i], self->params.color_depth == GSR_COLOR_DEPTH_8_BITS ? internal_formats_nv12[i] : internal_formats_p010[i], formats[i]);
+        self->target_textures[i] = gl_create_texture(self->params.egl, video_codec_context->width / div[i], video_codec_context->height / div[i], self->params.color_depth == GSR_COLOR_DEPTH_8_BITS ? internal_formats_nv12[i] : internal_formats_p010[i], formats[i], GL_NEAREST);
         if(self->target_textures[i] == 0) {
             fprintf(stderr, "gsr error: gsr_video_encoder_nvenc_setup_textures: failed to create opengl texture\n");
             return false;
@@ -138,6 +123,18 @@ static bool gsr_video_encoder_nvenc_start(gsr_video_encoder *encoder, AVCodecCon
         return false;
     }
 
+    video_codec_context->width = FFALIGN(video_codec_context->width, 2);
+    video_codec_context->height = FFALIGN(video_codec_context->height, 2);
+
+    if(video_codec_context->width < 128)
+        video_codec_context->width = 128;
+
+    if(video_codec_context->height < 128)
+        video_codec_context->height = 128;
+
+    frame->width = video_codec_context->width;
+    frame->height = video_codec_context->height;
+
     if(!gsr_video_encoder_nvenc_setup_context(self, video_codec_context)) {
         gsr_video_encoder_nvenc_stop(self, video_codec_context);
         return false;

src/encoder/video/software.c

@@ -1,5 +1,6 @@
 #include "../../../include/encoder/video/software.h"
 #include "../../../include/egl.h"
+#include "../../../include/utils.h"
 
 #include <libavcodec/avcodec.h>
 #include <libavutil/frame.h>
@@ -14,21 +15,6 @@ typedef struct {
     unsigned int target_textures[2];
 } gsr_video_encoder_software;
 
-static unsigned int gl_create_texture(gsr_egl *egl, int width, int height, int internal_format, unsigned int format) {
-    unsigned int texture_id = 0;
-    egl->glGenTextures(1, &texture_id);
-    egl->glBindTexture(GL_TEXTURE_2D, texture_id);
-    egl->glTexImage2D(GL_TEXTURE_2D, 0, internal_format, width, height, 0, format, GL_UNSIGNED_BYTE, NULL);
-
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
-
-    egl->glBindTexture(GL_TEXTURE_2D, 0);
-    return texture_id;
-}
-
 static bool gsr_video_encoder_software_setup_textures(gsr_video_encoder_software *self, AVCodecContext *video_codec_context, AVFrame *frame) {
     int res = av_frame_get_buffer(frame, LINESIZE_ALIGNMENT);
     if(res < 0) {
@@ -48,7 +34,7 @@ static bool gsr_video_encoder_software_setup_textures(gsr_video_encoder_software
     const int div[2] = {1, 2}; // divide UV texture size by 2 because chroma is half size
 
     for(int i = 0; i < 2; ++i) {
-        self->target_textures[i] = gl_create_texture(self->params.egl, video_codec_context->width / div[i], video_codec_context->height / div[i], self->params.color_depth == GSR_COLOR_DEPTH_8_BITS ? internal_formats_nv12[i] : internal_formats_p010[i], formats[i]);
+        self->target_textures[i] = gl_create_texture(self->params.egl, video_codec_context->width / div[i], video_codec_context->height / div[i], self->params.color_depth == GSR_COLOR_DEPTH_8_BITS ? internal_formats_nv12[i] : internal_formats_p010[i], formats[i], GL_NEAREST);
         if(self->target_textures[i] == 0) {
             fprintf(stderr, "gsr error: gsr_capture_kms_setup_cuda_textures: failed to create opengl texture\n");
             return false;
@@ -97,8 +83,8 @@ static void gsr_video_encoder_software_copy_textures_to_frame(gsr_video_encoder
     self->params.egl->glBindTexture(GL_TEXTURE_2D, 0);
     // cap_kms->kms.base.egl->eglSwapBuffers(cap_kms->kms.base.egl->egl_display, cap_kms->kms.base.egl->egl_surface);
 
-    self->params.egl->glFlush();
-    self->params.egl->glFinish();
+    //self->params.egl->glFlush();
+    //self->params.egl->glFinish();
 }
 
 static void gsr_video_encoder_software_get_textures(gsr_video_encoder *encoder, unsigned int *textures, int *num_textures, gsr_destination_color *destination_color) {

src/encoder/video/vaapi.c

@@ -120,11 +120,13 @@ static bool gsr_video_encoder_vaapi_setup_textures(gsr_video_encoder_vaapi *self
                 return false;
             }
 
+            const float border_color[4] = { 0.0f, 0.0f, 0.0f, 0.0f };
             self->params.egl->glBindTexture(GL_TEXTURE_2D, self->target_textures[i]);
-            self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-            self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
-            self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
-            self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
+            self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+            self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+            self->params.egl->glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, border_color);
+            self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
+            self->params.egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
 
             while(self->params.egl->glGetError()) {}
             while(self->params.egl->eglGetError() != EGL_SUCCESS){}
@@ -167,32 +169,21 @@ static bool gsr_video_encoder_vaapi_start(gsr_video_encoder *encoder, AVCodecCon
         } else {
             video_codec_context->height = FFALIGN(video_codec_context->height, 16);
         }
+    } else {
+        video_codec_context->width = FFALIGN(video_codec_context->width, 2);
+        video_codec_context->height = FFALIGN(video_codec_context->height, 2);
     }
 
-    const int crop_top = (video_codec_context->height - frame->height) / 2;
-    const int crop_left = (video_codec_context->width - frame->width) / 2;
-    if(crop_top != 0 || crop_left != 0) {
+    if(FFALIGN(video_codec_context->width, 2) != FFALIGN(frame->width, 2) || FFALIGN(video_codec_context->height, 2) != FFALIGN(frame->height, 2)) {
         fprintf(stderr, "gsr warning: gsr_video_encoder_vaapi_start: black bars have been added to the video because of a bug in AMD drivers/hardware. Record with h264 codec instead (-k h264) to get around this issue\n");
-#if 0
-        #if LIBAVCODEC_VERSION_INT >= AV_VERSION_INT(61, 10, 100)
-        const int crop_bottom = crop_top;
-        const int crop_right = crop_left;
-        fprintf(stderr, "gsr info: cropping metadata has been added to the file to try and workaround this issue. Video players that support this will remove the black bars when the video is playing\n");
-        const int frame_cropping_data_size = 4 * sizeof(uint32_t);
-        uint8_t *frame_cropping = av_malloc(frame_cropping_data_size);
-        if(frame_cropping) {
-            AV_WL32(frame_cropping + 0, crop_top);
-            AV_WL32(frame_cropping + 4, crop_bottom);
-            AV_WL32(frame_cropping + 8, crop_left);
-            AV_WL32(frame_cropping + 12, crop_right);
-            const bool sidedata_added = av_packet_side_data_add(&video_stream->codecpar->coded_side_data, &video_stream->codecpar->nb_coded_side_data, AV_PKT_DATA_FRAME_CROPPING, frame_cropping, frame_cropping_data_size, 0) != NULL;
-            if(!sidedata_added)
-                av_free(frame_cropping);
-        }
-        #endif
-#endif
     }
 
+    if(video_codec_context->width < 128)
+        video_codec_context->width = 128;
+
+    if(video_codec_context->height < 128)
+        video_codec_context->height = 128;
+
     frame->width = video_codec_context->width;
     frame->height = video_codec_context->height;
 

src/encoder/video/vulkan.c

@@ -71,22 +71,6 @@ static bool gsr_video_encoder_vulkan_setup_context(gsr_video_encoder_vulkan *sel
     return true;
 }
 
-static unsigned int gl_create_texture(gsr_egl *egl, int width, int height, int internal_format, unsigned int format) {
-    unsigned int texture_id = 0;
-    egl->glGenTextures(1, &texture_id);
-    egl->glBindTexture(GL_TEXTURE_2D, texture_id);
-    //egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_TILING_EXT, GL_OPTIMAL_TILING_EXT);
-    egl->glTexImage2D(GL_TEXTURE_2D, 0, internal_format, width, height, 0, format, GL_UNSIGNED_BYTE, NULL);
-
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
-    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
-
-    egl->glBindTexture(GL_TEXTURE_2D, 0);
-    return texture_id;
-}
-
 static AVVulkanDeviceContext* video_codec_context_get_vulkan_data(AVCodecContext *video_codec_context) {
     AVBufferRef *hw_frames_ctx = video_codec_context->hw_frames_ctx;
     if(!hw_frames_ctx)
@@ -116,7 +100,7 @@ static bool gsr_video_encoder_vulkan_setup_textures(gsr_video_encoder_vulkan *se
     const int div[2] = {1, 2}; // divide UV texture size by 2 because chroma is half size
 
     for(int i = 0; i < 2; ++i) {
-        self->target_textures[i] = gl_create_texture(self->params.egl, video_codec_context->width / div[i], video_codec_context->height / div[i], self->params.color_depth == GSR_COLOR_DEPTH_8_BITS ? internal_formats_nv12[i] : internal_formats_p010[i], formats[i]);
+        self->target_textures[i] = gl_create_texture(self->params.egl, video_codec_context->width / div[i], video_codec_context->height / div[i], self->params.color_depth == GSR_COLOR_DEPTH_8_BITS ? internal_formats_nv12[i] : internal_formats_p010[i], formats[i], GL_NEAREST);
         if(self->target_textures[i] == 0) {
             fprintf(stderr, "gsr error: gsr_video_encoder_cuda_setup_textures: failed to create opengl texture\n");
             return false;
@@ -165,6 +149,18 @@ static void gsr_video_encoder_vulkan_stop(gsr_video_encoder_vulkan *self, AVCode
 static bool gsr_video_encoder_vulkan_start(gsr_video_encoder *encoder, AVCodecContext *video_codec_context, AVFrame *frame) {
     gsr_video_encoder_vulkan *self = encoder->priv;
 
+    video_codec_context->width = FFALIGN(video_codec_context->width, 2);
+    video_codec_context->height = FFALIGN(video_codec_context->height, 2);
+
+    if(video_codec_context->width < 128)
+        video_codec_context->width = 128;
+
+    if(video_codec_context->height < 128)
+        video_codec_context->height = 128;
+
+    frame->width = video_codec_context->width;
+    frame->height = video_codec_context->height;
+
     if(!gsr_video_encoder_vulkan_setup_context(self, video_codec_context)) {
         gsr_video_encoder_vulkan_stop(self, video_codec_context);
         return false;

src/image_writer.c

@@ -0,0 +1,96 @@
+#include "../include/image_writer.h"
+#include "../include/egl.h"
+#include "../include/utils.h"
+
+#define STB_IMAGE_WRITE_IMPLEMENTATION
+#include "../external/stb_image_write.h"
+
+#include <stdlib.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <assert.h>
+
+/* TODO: Support hdr/10-bit */
+bool gsr_image_writer_init_opengl(gsr_image_writer *self, gsr_egl *egl, int width, int height) {
+    memset(self, 0, sizeof(*self));
+    self->source = GSR_IMAGE_WRITER_SOURCE_OPENGL;
+    self->egl = egl;
+    self->width = width;
+    self->height = height;
+    self->texture = gl_create_texture(self->egl, self->width, self->height, GL_RGBA8, GL_RGBA, GL_NEAREST); /* TODO: use GL_RGB16 instead of GL_RGB8 for hdr/10-bit */
+    if(self->texture == 0) {
+        fprintf(stderr, "gsr error: gsr_image_writer_init: failed to create texture\n");
+        return false;
+    }
+    return true;
+}
+
+bool gsr_image_writer_init_memory(gsr_image_writer *self, const void *memory, int width, int height) {
+    memset(self, 0, sizeof(*self));
+    self->source = GSR_IMAGE_WRITER_SOURCE_OPENGL;
+    self->width = width;
+    self->height = height;
+    self->memory = memory;
+    return true;
+}
+
+void gsr_image_writer_deinit(gsr_image_writer *self) {
+    if(self->texture) {
+        self->egl->glDeleteTextures(1, &self->texture);
+        self->texture = 0;
+    }
+}
+
+static bool gsr_image_writer_write_memory_to_file(gsr_image_writer *self, const char *filepath, gsr_image_format image_format, int quality, const void *data) {
+    if(quality < 1)
+        quality = 1;
+    else if(quality > 100)
+        quality = 100;
+
+    bool success = false;
+    switch(image_format) {
+        case GSR_IMAGE_FORMAT_JPEG:
+            success = stbi_write_jpg(filepath, self->width, self->height, 4, data, quality);
+            break;
+        case GSR_IMAGE_FORMAT_PNG:
+            success = stbi_write_png(filepath, self->width, self->height, 4, data, 0);
+            break;
+    }
+
+    if(!success)
+        fprintf(stderr, "gsr error: gsr_image_writer_write_to_file: failed to write image data to output file %s\n", filepath);
+
+    return success;
+}
+
+static bool gsr_image_writer_write_opengl_texture_to_file(gsr_image_writer *self, const char *filepath, gsr_image_format image_format, int quality) {
+    assert(self->source == GSR_IMAGE_WRITER_SOURCE_OPENGL);
+    uint8_t *frame_data = malloc(self->width * self->height * 4);
+    if(!frame_data) {
+        fprintf(stderr, "gsr error: gsr_image_writer_write_to_file: failed to allocate memory for image frame\n");
+        return false;
+    }
+
+    // TODO: hdr support
+    self->egl->glBindTexture(GL_TEXTURE_2D, self->texture);
+    // We could use glGetTexSubImage, but it's only available starting from opengl 4.5
+    self->egl->glGetTexImage(GL_TEXTURE_2D, 0, GL_RGBA, GL_UNSIGNED_BYTE, frame_data);
+    self->egl->glBindTexture(GL_TEXTURE_2D, 0);
+
+    self->egl->glFlush();
+    self->egl->glFinish();
+    
+    const bool success = gsr_image_writer_write_memory_to_file(self, filepath, image_format, quality, frame_data);
+    free(frame_data);
+    return success;
+}
+
+bool gsr_image_writer_write_to_file(gsr_image_writer *self, const char *filepath, gsr_image_format image_format, int quality) {
+    switch(self->source) {
+        case GSR_IMAGE_WRITER_SOURCE_OPENGL:
+            return gsr_image_writer_write_opengl_texture_to_file(self, filepath, image_format, quality);
+        case GSR_IMAGE_WRITER_SOURCE_MEMORY:
+            return gsr_image_writer_write_memory_to_file(self, filepath, image_format, quality, self->memory);
+    }
+    return false;
+}

src/pipewire_audio.c

@@ -1,6 +1,8 @@
 #include "../include/pipewire_audio.h"
 
 #include <pipewire/pipewire.h>
+#include <pipewire/extensions/metadata.h>
+#include <pipewire/impl-module.h>
 
 static void on_core_info_cb(void *user_data, const struct pw_core_info *info) {
     gsr_pipewire_audio *self = user_data;
@@ -44,13 +46,106 @@ static gsr_pipewire_audio_port* gsr_pipewire_audio_get_node_port_by_name(gsr_pip
 }
 
 static bool requested_link_matches_name_case_insensitive(const gsr_pipewire_audio_requested_link *requested_link, const char *name) {
-    for(int i = 0; i < requested_link->num_output_names; ++i) {
-        if(strcasecmp(requested_link->output_names[i], name) == 0)
+    for(int i = 0; i < requested_link->num_outputs; ++i) {
+        if(requested_link->outputs[i].type == GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_STANDARD && strcasecmp(requested_link->outputs[i].name, name) == 0)
             return true;
     }
     return false;
 }
 
+static bool requested_link_has_type(const gsr_pipewire_audio_requested_link *requested_link, gsr_pipewire_audio_requested_type type) {
+    for(int i = 0; i < requested_link->num_outputs; ++i) {
+        if(requested_link->outputs[i].type == type)
+            return true;
+    }
+    return false;
+}
+
+static void gsr_pipewire_get_node_input_port_by_type(gsr_pipewire_audio *self, const gsr_pipewire_audio_node *input_node, gsr_pipewire_audio_link_input_type input_type,
+    const gsr_pipewire_audio_port **input_fl_port, const gsr_pipewire_audio_port **input_fr_port)
+{
+    *input_fl_port = NULL;
+    *input_fr_port = NULL;
+
+    switch(input_type) {
+        case GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM: {
+            *input_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, input_node->id, "input_FL");
+            *input_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, input_node->id, "input_FR");
+            break;
+        }
+        case GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK: {
+            *input_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, input_node->id, "playback_FL");
+            *input_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, input_node->id, "playback_FR");
+            break;
+        }
+    }
+}
+
+static void gsr_pipewire_get_node_output_port_by_type(gsr_pipewire_audio *self, const gsr_pipewire_audio_node *output_node, gsr_pipewire_audio_node_type output_type,
+    const gsr_pipewire_audio_port **output_fl_port, const gsr_pipewire_audio_port **output_fr_port)
+{
+    *output_fl_port = NULL;
+    *output_fr_port = NULL;
+
+    switch(output_type) {
+        case GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT:
+            *output_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "output_FL");
+            *output_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "output_FR");
+            break;
+        case GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_INPUT:
+            *output_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_FL");
+            *output_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_FR");
+            break;
+        case GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE: {
+            *output_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_FL");
+            *output_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_FR");
+            if(!*output_fl_port || !*output_fr_port) {
+                *output_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "capture_FL");
+                *output_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "capture_FR");
+            }
+            if(!*output_fl_port || !*output_fr_port) {
+                const gsr_pipewire_audio_port *output_mono_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_MONO");
+                if(!output_mono_port)
+                    output_mono_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "capture_MONO");
+
+                if(output_mono_port) {
+                    *output_fl_port = output_mono_port;
+                    *output_fr_port = output_mono_port;
+                }
+            }
+            break;
+        }
+    }
+}
+
+static void gsr_pipewire_audio_establish_link(gsr_pipewire_audio *self, const gsr_pipewire_audio_port *input_fl_port, const gsr_pipewire_audio_port *input_fr_port,
+    const gsr_pipewire_audio_port *output_fl_port, const gsr_pipewire_audio_port *output_fr_port)
+{
+    // TODO: Detect if link already exists before so we dont create these proxies when not needed
+
+    //fprintf(stderr, "linking!\n");
+    // TODO: error check and cleanup
+    {
+        struct pw_properties *props = pw_properties_new(NULL, NULL);
+        pw_properties_setf(props, PW_KEY_LINK_OUTPUT_PORT, "%u", output_fl_port->id);
+        pw_properties_setf(props, PW_KEY_LINK_INPUT_PORT, "%u", input_fl_port->id);
+        // TODO: Clean this up when removing node
+        struct pw_proxy *proxy = pw_core_create_object(self->core, "link-factory", PW_TYPE_INTERFACE_Link, PW_VERSION_LINK, &props->dict, 0);
+        //self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, self->server_version_sync);
+        pw_properties_free(props);
+    }
+
+    {
+        struct pw_properties *props = pw_properties_new(NULL, NULL);
+        pw_properties_setf(props, PW_KEY_LINK_OUTPUT_PORT, "%u", output_fr_port->id);
+        pw_properties_setf(props, PW_KEY_LINK_INPUT_PORT, "%u", input_fr_port->id);
+        // TODO: Clean this up when removing node
+        struct pw_proxy *proxy = pw_core_create_object(self->core, "link-factory", PW_TYPE_INTERFACE_Link, PW_VERSION_LINK, &props->dict, 0);
+        //self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, self->server_version_sync);
+        pw_properties_free(props);
+    }
+}
+
 static void gsr_pipewire_audio_create_link(gsr_pipewire_audio *self, const gsr_pipewire_audio_requested_link *requested_link) {
     const gsr_pipewire_audio_node_type requested_link_node_type = requested_link->input_type == GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM ? GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_INPUT : GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE;
     const gsr_pipewire_audio_node *stream_input_node = gsr_pipewire_audio_get_node_by_name_case_insensitive(self, requested_link->input_name, requested_link_node_type);
@@ -59,20 +154,7 @@ static void gsr_pipewire_audio_create_link(gsr_pipewire_audio *self, const gsr_p
 
     const gsr_pipewire_audio_port *input_fl_port = NULL;
     const gsr_pipewire_audio_port *input_fr_port = NULL;
-
-    switch(requested_link->input_type) {
-        case GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM: {
-            input_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, stream_input_node->id, "input_FL");
-            input_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, stream_input_node->id, "input_FR");
-            break;
-        }
-        case GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK: {
-            input_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, stream_input_node->id, "playback_FL");
-            input_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, stream_input_node->id, "playback_FR");
-            break;
-        }
-    }
-
+    gsr_pipewire_get_node_input_port_by_type(self, stream_input_node, requested_link->input_type, &input_fl_port, &input_fr_port);
     if(!input_fl_port || !input_fr_port)
         return;
 
@@ -92,68 +174,208 @@ static void gsr_pipewire_audio_create_link(gsr_pipewire_audio *self, const gsr_p
 
         const gsr_pipewire_audio_port *output_fl_port = NULL;
         const gsr_pipewire_audio_port *output_fr_port = NULL;
-
-        switch(requested_link->output_type) {
-            case GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT:
-                output_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "output_FL");
-                output_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "output_FR");
-                break;
-            case GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_INPUT:
-                output_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_FL");
-                output_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_FR");
-                break;
-            case GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE: {
-                output_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_FL");
-                output_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "monitor_FR");
-                if(!output_fl_port || !output_fr_port) {
-                    output_fl_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "capture_FL");
-                    output_fr_port = gsr_pipewire_audio_get_node_port_by_name(self, output_node->id, "capture_FR");
-                }
-                break;
-            }
-        }
-
+        gsr_pipewire_get_node_output_port_by_type(self, output_node, requested_link->output_type, &output_fl_port, &output_fr_port);
         if(!output_fl_port || !output_fr_port)
             continue;
 
-        // TODO: Detect if link already exists before so we dont create these proxies when not needed
-
-        //fprintf(stderr, "linking!\n");
-        // TODO: error check and cleanup
-        {
-            struct pw_properties *props = pw_properties_new(NULL, NULL);
-            pw_properties_setf(props, PW_KEY_LINK_OUTPUT_PORT, "%u", output_fl_port->id);
-            pw_properties_setf(props, PW_KEY_LINK_INPUT_PORT, "%u", input_fl_port->id);
-            // TODO: Clean this up when removing node
-            struct pw_proxy *proxy = pw_core_create_object(self->core, "link-factory", PW_TYPE_INTERFACE_Link, PW_VERSION_LINK, &props->dict, 0);
-            //self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, self->server_version_sync);
-            pw_properties_free(props);
-        }
-
-        {
-            struct pw_properties *props = pw_properties_new(NULL, NULL);
-            pw_properties_setf(props, PW_KEY_LINK_OUTPUT_PORT, "%u", output_fr_port->id);
-            pw_properties_setf(props, PW_KEY_LINK_INPUT_PORT, "%u", input_fr_port->id);
-            // TODO: Clean this up when removing node
-            struct pw_proxy *proxy = pw_core_create_object(self->core, "link-factory", PW_TYPE_INTERFACE_Link, PW_VERSION_LINK, &props->dict, 0);
-            //self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, self->server_version_sync);
-            pw_properties_free(props);
-        }
+        gsr_pipewire_audio_establish_link(self, input_fl_port, input_fr_port, output_fl_port, output_fr_port);
     }
 }
 
 static void gsr_pipewire_audio_create_links(gsr_pipewire_audio *self) {
-    for(int j = 0; j < self->num_requested_links; ++j) {
-        gsr_pipewire_audio_create_link(self, &self->requested_links[j]);
+    for(int i = 0; i < self->num_requested_links; ++i) {
+        gsr_pipewire_audio_create_link(self, &self->requested_links[i]);
     }
 }
 
+static void gsr_pipewire_audio_create_link_for_default_devices(gsr_pipewire_audio *self, const gsr_pipewire_audio_requested_link *requested_link, gsr_pipewire_audio_requested_type default_device_type) {
+    if(default_device_type == GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_STANDARD)
+        return;
+
+    const char *device_name = default_device_type == GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT ? self->default_output_device_name : self->default_input_device_name;
+    if(device_name[0] == '\0')
+        return;
+
+    if(!requested_link_has_type(requested_link, default_device_type))
+        return;
+
+    const gsr_pipewire_audio_node_type requested_link_node_type = requested_link->input_type == GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM ? GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_INPUT : GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE;
+    const gsr_pipewire_audio_node *stream_input_node = gsr_pipewire_audio_get_node_by_name_case_insensitive(self, requested_link->input_name, requested_link_node_type);
+    if(!stream_input_node)
+        return;
+
+    const gsr_pipewire_audio_port *input_fl_port = NULL;
+    const gsr_pipewire_audio_port *input_fr_port = NULL;
+    gsr_pipewire_get_node_input_port_by_type(self, stream_input_node, requested_link->input_type, &input_fl_port, &input_fr_port);
+    if(!input_fl_port || !input_fr_port)
+        return;
+
+    const gsr_pipewire_audio_node *stream_output_node = gsr_pipewire_audio_get_node_by_name_case_insensitive(self, device_name, GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE);
+    if(!stream_output_node)
+        return;
+
+    const gsr_pipewire_audio_port *output_fl_port = NULL;
+    const gsr_pipewire_audio_port *output_fr_port = NULL;
+    gsr_pipewire_get_node_output_port_by_type(self, stream_output_node, requested_link->output_type, &output_fl_port, &output_fr_port);
+    if(!output_fl_port || !output_fr_port)
+        return;
+
+    gsr_pipewire_audio_establish_link(self, input_fl_port, input_fr_port, output_fl_port, output_fr_port);
+    //fprintf(stderr, "establishing a link from %u to %u\n", stream_output_node->id, stream_input_node->id);
+}
+
+static void gsr_pipewire_audio_create_links_for_default_devices(gsr_pipewire_audio *self, gsr_pipewire_audio_requested_type default_device_type) {
+    for(int i = 0; i < self->num_requested_links; ++i) {
+        gsr_pipewire_audio_create_link_for_default_devices(self, &self->requested_links[i], default_device_type);
+    }
+}
+
+static void gsr_pipewire_audio_destroy_links_by_output_to_input(gsr_pipewire_audio *self, uint32_t output_node_id, uint32_t input_node_id) {
+    for(int i = 0; i < self->num_links; ++i) {
+        if(self->links[i].output_node_id == output_node_id && self->links[i].input_node_id == input_node_id)
+            pw_registry_destroy(self->registry, self->links[i].id);
+    }
+}
+
+static void gsr_pipewire_destroy_default_device_link(gsr_pipewire_audio *self, const gsr_pipewire_audio_requested_link *requested_link, gsr_pipewire_audio_requested_type default_device_type) {
+    if(default_device_type == GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_STANDARD)
+        return;
+
+    const char *device_name = default_device_type == GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT ? self->default_output_device_name : self->default_input_device_name;
+    if(device_name[0] == '\0')
+        return;
+
+    if(!requested_link_has_type(requested_link, default_device_type))
+        return;
+
+    /* default_output and default_input can be the same device. In that case both are the same link and we dont want to remove the link */
+    const gsr_pipewire_audio_requested_type opposite_device_type = default_device_type == GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT ? GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_INPUT : GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT;
+    const char *opposite_device_name = opposite_device_type == GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT ? self->default_output_device_name : self->default_input_device_name;
+    if(requested_link_has_type(requested_link, opposite_device_type) && strcmp(device_name, opposite_device_name) == 0)
+        return;
+
+    const gsr_pipewire_audio_node_type requested_link_node_type = requested_link->input_type == GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM ? GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_INPUT : GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE;
+    const gsr_pipewire_audio_node *stream_input_node = gsr_pipewire_audio_get_node_by_name_case_insensitive(self, requested_link->input_name, requested_link_node_type);
+    if(!stream_input_node)
+        return;
+
+    const gsr_pipewire_audio_node *stream_output_node = gsr_pipewire_audio_get_node_by_name_case_insensitive(self, device_name, GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE);
+    if(!stream_output_node)
+        return;
+
+    if(requested_link_matches_name_case_insensitive(requested_link, stream_output_node->name))
+        return;
+
+    gsr_pipewire_audio_destroy_links_by_output_to_input(self, stream_output_node->id, stream_input_node->id);
+    //fprintf(stderr, "destroying a link from %u to %u\n", stream_output_node->id, stream_input_node->id);
+}
+
+static void gsr_pipewire_destroy_default_device_links(gsr_pipewire_audio *self, gsr_pipewire_audio_requested_type default_device_type) {
+    for(int i = 0; i < self->num_requested_links; ++i) {
+        gsr_pipewire_destroy_default_device_link(self, &self->requested_links[i], default_device_type);
+    }
+}
+
+static bool json_get_value(const char *json_str, const char *key, char *value, size_t value_size) {
+    char key_full[32];
+    const int key_full_size = snprintf(key_full, sizeof(key_full), "\"%s\":", key);
+    const char *start = strstr(json_str, key_full);
+    if(!start)
+        return false;
+    
+    start += key_full_size;
+    const char *value_start = strchr(start, '"');
+    if(!value_start)
+        return false;
+
+    value_start += 1;
+    const char *value_end = strchr(value_start, '"');
+    if(!value_end)
+        return false;
+
+    snprintf(value, value_size, "%.*s", (int)(value_end - value_start), value_start);
+    return true;
+}
+
+static int on_metadata_property_cb(void *data, uint32_t id, const char *key, const char *type, const char *value) {
+	(void)type;
+    gsr_pipewire_audio *self = data;
+
+	if(id == PW_ID_CORE && key && value) {
+        char value_decoded[128];
+        if(strcmp(key, "default.audio.sink") == 0) {
+            if(json_get_value(value, "name", value_decoded, sizeof(value_decoded)) && strcmp(value_decoded, self->default_output_device_name) != 0) {
+                gsr_pipewire_destroy_default_device_links(self, GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT);
+                snprintf(self->default_output_device_name, sizeof(self->default_output_device_name), "%s", value_decoded);
+                gsr_pipewire_audio_create_links_for_default_devices(self, GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT);
+            }
+        } else if(strcmp(key, "default.audio.source") == 0) {
+            if(json_get_value(value, "name", value_decoded, sizeof(value_decoded)) && strcmp(value_decoded, self->default_input_device_name) != 0) {
+                gsr_pipewire_destroy_default_device_links(self, GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_INPUT);
+                snprintf(self->default_input_device_name, sizeof(self->default_input_device_name), "%s", value_decoded);
+                gsr_pipewire_audio_create_links_for_default_devices(self, GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_INPUT);
+            }
+        }
+	}
+
+	return 0;
+}
+
+static const struct pw_metadata_events metadata_events = {
+	PW_VERSION_METADATA_EVENTS,
+	.property = on_metadata_property_cb,
+};
+
+static void on_metadata_proxy_removed_cb(void *data) {
+    gsr_pipewire_audio *self = data;
+    if(self->metadata_proxy) {
+        pw_proxy_destroy(self->metadata_proxy);
+        self->metadata_proxy = NULL;
+    }
+}
+
+static void on_metadata_proxy_destroy_cb(void *data) {
+	gsr_pipewire_audio *self = data;
+
+	spa_hook_remove(&self->metadata_listener);
+	spa_hook_remove(&self->metadata_proxy_listener);
+	spa_zero(self->metadata_listener);
+	spa_zero(self->metadata_proxy_listener);
+
+	self->metadata_proxy = NULL;
+}
+
+static const struct pw_proxy_events metadata_proxy_events = {
+	PW_VERSION_PROXY_EVENTS,
+	.removed = on_metadata_proxy_removed_cb,
+	.destroy = on_metadata_proxy_destroy_cb,
+};
+
+static bool gsr_pipewire_audio_listen_on_metadata(gsr_pipewire_audio *self, uint32_t id) {
+    if(self->metadata_proxy) {
+        pw_proxy_destroy(self->metadata_proxy);
+        self->metadata_proxy = NULL;
+    }
+
+    self->metadata_proxy = pw_registry_bind(self->registry, id, PW_TYPE_INTERFACE_Metadata, PW_VERSION_METADATA, 0);
+    if(!self->metadata_proxy) {
+        fprintf(stderr, "gsr error: gsr_pipewire_audio_listen_on_metadata: failed to bind to registry\n");
+        return false;
+    }
+
+    pw_proxy_add_object_listener(self->metadata_proxy, &self->metadata_listener, &metadata_events, self);
+    pw_proxy_add_listener(self->metadata_proxy, &self->metadata_proxy_listener, &metadata_proxy_events, self);
+
+    self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, self->server_version_sync);
+    return true;
+}
+
 static void registry_event_global(void *data, uint32_t id, uint32_t permissions,
                   const char *type, uint32_t version,
                   const struct spa_dict *props)
 {
     //fprintf(stderr, "add: id: %d, type: %s\n", (int)id, type);
-    if (props == NULL)
+    if(!props || !type)
         return;
 
     //pw_properties_new_dict(props);
@@ -162,7 +384,7 @@ static void registry_event_global(void *data, uint32_t id, uint32_t permissions,
     if(strcmp(type, PW_TYPE_INTERFACE_Node) == 0) {
         const char *node_name = spa_dict_lookup(props, PW_KEY_NODE_NAME);
         const char *media_class = spa_dict_lookup(props, PW_KEY_MEDIA_CLASS);
-        //fprintf(stderr, "  node name: %s, media class: %s\n", node_name, media_class);
+        //fprintf(stderr, "  node id: %u, node name: %s, media class: %s\n", id, node_name, media_class);
         const bool is_stream_output = media_class && strcmp(media_class, "Stream/Output/Audio") == 0;
         const bool is_stream_input = media_class && strcmp(media_class, "Stream/Input/Audio") == 0;
         const bool is_sink = media_class && strcmp(media_class, "Audio/Sink") == 0;
@@ -206,6 +428,7 @@ static void registry_event_global(void *data, uint32_t id, uint32_t permissions,
             //fprintf(stderr, "  port name: %s, node id: %d, direction: %s\n", port_name, node_id_num, port_direction);
             char *port_name_copy = strdup(port_name);
             if(port_name_copy) {
+                //fprintf(stderr, "  port id: %u, node id: %u, name: %s\n", id, node_id_num, port_name_copy);
                 self->ports[self->num_ports].id = id;
                 self->ports[self->num_ports].node_id = node_id_num;
                 self->ports[self->num_ports].direction = direction;
@@ -217,6 +440,25 @@ static void registry_event_global(void *data, uint32_t id, uint32_t permissions,
         } else if(self->num_ports >= GSR_PIPEWIRE_AUDIO_MAX_PORTS) {
             fprintf(stderr, "gsr error: reached the maximum amount of audio ports\n");
         }
+    } else if(strcmp(type, PW_TYPE_INTERFACE_Link) == 0) {
+        const char *output_node = spa_dict_lookup(props, PW_KEY_LINK_OUTPUT_NODE);
+        const char *input_node = spa_dict_lookup(props, PW_KEY_LINK_INPUT_NODE);
+
+        const uint32_t output_node_id_num = output_node ? atoi(output_node) : 0;
+        const uint32_t input_node_id_num = input_node ? atoi(input_node) : 0;
+        if(self->num_links < GSR_PIPEWIRE_AUDIO_MAX_LINKS && output_node_id_num > 0 && input_node_id_num > 0) {
+            //fprintf(stderr, "  new link (%u): %u -> %u\n", id, output_node_id_num, input_node_id_num);
+            self->links[self->num_links].id = id;
+            self->links[self->num_links].output_node_id = output_node_id_num;
+            self->links[self->num_links].input_node_id = input_node_id_num;
+            ++self->num_links;
+        } else if(self->num_ports >= GSR_PIPEWIRE_AUDIO_MAX_LINKS) {
+            fprintf(stderr, "gsr error: reached the maximum amount of audio links\n");
+        }
+    } else if(strcmp(type, PW_TYPE_INTERFACE_Metadata) == 0) {
+        const char *name = spa_dict_lookup(props, PW_KEY_METADATA_NAME);
+        if(name && strcmp(name, "default") == 0)
+            gsr_pipewire_audio_listen_on_metadata(self, id);
     }
 }
 
@@ -226,9 +468,7 @@ static bool gsr_pipewire_audio_remove_node_by_id(gsr_pipewire_audio *self, uint3
             continue;
 
         free(self->stream_nodes[i].name);
-        for(int j = i + 1; j < self->num_stream_nodes; ++j) {
-            self->stream_nodes[j - 1] = self->stream_nodes[j];
-        }
+        self->stream_nodes[i] = self->stream_nodes[self->num_stream_nodes - 1];
         --self->num_stream_nodes;
         return true;
     }
@@ -241,15 +481,25 @@ static bool gsr_pipewire_audio_remove_port_by_id(gsr_pipewire_audio *self, uint3
             continue;
 
         free(self->ports[i].name);
-        for(int j = i + 1; j < self->num_ports; ++j) {
-            self->ports[j - 1] = self->ports[j];
-        }
+        self->ports[i] = self->ports[self->num_ports - 1];
         --self->num_ports;
         return true;
     }
     return false;
 }
 
+static bool gsr_pipewire_audio_remove_link_by_id(gsr_pipewire_audio *self, uint32_t link_id) {
+    for(int i = 0; i < self->num_links; ++i) {
+        if(self->links[i].id != link_id)
+            continue;
+
+        self->links[i] = self->links[self->num_links - 1];
+        --self->num_links;
+        return true;
+    }
+    return false;
+}
+
 static void registry_event_global_remove(void *data, uint32_t id) {
     //fprintf(stderr, "remove: %d\n", (int)id);
     gsr_pipewire_audio *self = (gsr_pipewire_audio*)data;
@@ -262,6 +512,11 @@ static void registry_event_global_remove(void *data, uint32_t id) {
         //fprintf(stderr, "removed port\n");
         return;
     }
+
+    if(gsr_pipewire_audio_remove_link_by_id(self, id)) {
+        //fprintf(stderr, "removed link\n");
+        return;
+    }
 }
 
 static const struct pw_registry_events registry_events = {
@@ -289,6 +544,8 @@ bool gsr_pipewire_audio_init(gsr_pipewire_audio *self) {
         return false;
     }
 
+    pw_context_load_module(self->context, "libpipewire-module-link-factory", NULL, NULL);
+
     if(pw_thread_loop_start(self->thread_loop) < 0) {
         fprintf(stderr, "gsr error: gsr_pipewire_audio_init: failed to start thread\n");
         gsr_pipewire_audio_deinit(self);
@@ -310,8 +567,9 @@ bool gsr_pipewire_audio_init(gsr_pipewire_audio *self) {
     self->registry = pw_core_get_registry(self->core, PW_VERSION_REGISTRY, 0);
     pw_registry_add_listener(self->registry, &self->registry_listener, &registry_events, self);
 
-    self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, 0);
+    self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, self->server_version_sync);
     pw_thread_loop_wait(self->thread_loop);
+
     pw_thread_loop_unlock(self->thread_loop);
     return true;
 }
@@ -330,6 +588,18 @@ void gsr_pipewire_audio_deinit(gsr_pipewire_audio *self) {
     }
     self->num_virtual_sink_proxies = 0;
 
+    if(self->metadata_proxy) {
+        spa_hook_remove(&self->metadata_listener);
+        spa_hook_remove(&self->metadata_proxy_listener);
+        pw_proxy_destroy(self->metadata_proxy);
+        spa_zero(self->metadata_listener);
+        spa_zero(self->metadata_proxy_listener);
+        self->metadata_proxy = NULL;
+    }
+
+    spa_hook_remove(&self->registry_listener);
+    spa_hook_remove(&self->core_listener);
+
     if(self->core) {
         pw_core_disconnect(self->core);
         self->core = NULL;
@@ -355,11 +625,13 @@ void gsr_pipewire_audio_deinit(gsr_pipewire_audio *self) {
     }
     self->num_ports = 0;
 
+    self->num_links = 0;
+
     for(int i = 0; i < self->num_requested_links; ++i) {
-        for(int j = 0; j < self->requested_links[i].num_output_names; ++j) {
-            free(self->requested_links[i].output_names[j]);
+        for(int j = 0; j < self->requested_links[i].num_outputs; ++j) {
+            free(self->requested_links[i].outputs[j].name);
         }
-        free(self->requested_links[i].output_names);
+        free(self->requested_links[i].outputs);
         free(self->requested_links[i].input_name);
     }
     self->num_requested_links = 0;
@@ -428,14 +700,14 @@ static bool string_remove_suffix(char *str, const char *suffix) {
     }
 }
 
-static bool gsr_pipewire_audio_add_link_from_apps_to_output(gsr_pipewire_audio *self, const char **output_names, int num_output_names, const char *input_name, gsr_pipewire_audio_node_type output_type, gsr_pipewire_audio_link_input_type input_type, bool inverted) {
+static bool gsr_pipewire_audio_add_links_to_output(gsr_pipewire_audio *self, const char **output_names, int num_output_names, const char *input_name, gsr_pipewire_audio_node_type output_type, gsr_pipewire_audio_link_input_type input_type, bool inverted) {
     if(self->num_requested_links >= GSR_PIPEWIRE_AUDIO_MAX_REQUESTED_LINKS) {
         fprintf(stderr, "gsr error: reached the maximum amount of audio links\n");
         return false;
     }
     
-    char **output_names_copy = calloc(num_output_names, sizeof(char*));
-    if(!output_names_copy)
+    gsr_pipewire_audio_requested_output *outputs = calloc(num_output_names, sizeof(gsr_pipewire_audio_requested_output));
+    if(!outputs)
         return false;
 
     char *input_name_copy = strdup(input_name);
@@ -446,23 +718,34 @@ static bool gsr_pipewire_audio_add_link_from_apps_to_output(gsr_pipewire_audio *
         string_remove_suffix(input_name_copy, ".monitor");
 
     for(int i = 0; i < num_output_names; ++i) {
-        output_names_copy[i] = strdup(output_names[i]);
-        if(!output_names_copy[i])
+        outputs[i].name = strdup(output_names[i]);
+        if(!outputs[i].name)
             goto error;
 
-        if(output_type == GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE)
-            string_remove_suffix(output_names_copy[i], ".monitor");
+        outputs[i].type = GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_STANDARD;
+        if(output_type == GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE) {
+            string_remove_suffix(outputs[i].name, ".monitor");
+
+            if(strcmp(outputs[i].name, "default_output") == 0)
+                outputs[i].type = GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT;
+            else if(strcmp(outputs[i].name, "default_input") == 0)
+                outputs[i].type = GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_INPUT;
+            else
+                outputs[i].type = GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_STANDARD;
+        }
     }
 
     pw_thread_loop_lock(self->thread_loop);
-    self->requested_links[self->num_requested_links].output_names = output_names_copy;
-    self->requested_links[self->num_requested_links].num_output_names = num_output_names;
+    self->requested_links[self->num_requested_links].outputs = outputs;
+    self->requested_links[self->num_requested_links].num_outputs = num_output_names;
     self->requested_links[self->num_requested_links].input_name = input_name_copy;
     self->requested_links[self->num_requested_links].output_type = output_type;
     self->requested_links[self->num_requested_links].input_type = input_type;
     self->requested_links[self->num_requested_links].inverted = inverted;
     ++self->num_requested_links;
     gsr_pipewire_audio_create_link(self, &self->requested_links[self->num_requested_links - 1]);
+    gsr_pipewire_audio_create_link_for_default_devices(self, &self->requested_links[self->num_requested_links - 1], GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_OUTPUT);
+    gsr_pipewire_audio_create_link_for_default_devices(self, &self->requested_links[self->num_requested_links - 1], GSR_PIPEWIRE_AUDIO_REQUESTED_TYPE_DEFAULT_INPUT);
     pw_thread_loop_unlock(self->thread_loop);
 
     return true;
@@ -470,30 +753,30 @@ static bool gsr_pipewire_audio_add_link_from_apps_to_output(gsr_pipewire_audio *
     error:
     free(input_name_copy);
     for(int i = 0; i < num_output_names; ++i) {
-        free(output_names_copy[i]);
+        free(outputs[i].name);
     }
-    free(output_names_copy);
+    free(outputs);
     return false;
 }
 
 bool gsr_pipewire_audio_add_link_from_apps_to_stream(gsr_pipewire_audio *self, const char **app_names, int num_app_names, const char *stream_name_input) {
-    return gsr_pipewire_audio_add_link_from_apps_to_output(self, app_names, num_app_names, stream_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM, false);
+    return gsr_pipewire_audio_add_links_to_output(self, app_names, num_app_names, stream_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM, false);
 }
 
 bool gsr_pipewire_audio_add_link_from_apps_to_stream_inverted(gsr_pipewire_audio *self, const char **app_names, int num_app_names, const char *stream_name_input) {
-    return gsr_pipewire_audio_add_link_from_apps_to_output(self, app_names, num_app_names, stream_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM, true);
+    return gsr_pipewire_audio_add_links_to_output(self, app_names, num_app_names, stream_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_STREAM, true);
 }
 
 bool gsr_pipewire_audio_add_link_from_apps_to_sink(gsr_pipewire_audio *self, const char **app_names, int num_app_names, const char *sink_name_input) {
-    return gsr_pipewire_audio_add_link_from_apps_to_output(self, app_names, num_app_names, sink_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK, false);
+    return gsr_pipewire_audio_add_links_to_output(self, app_names, num_app_names, sink_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK, false);
 }
 
 bool gsr_pipewire_audio_add_link_from_apps_to_sink_inverted(gsr_pipewire_audio *self, const char **app_names, int num_app_names, const char *sink_name_input) {
-    return gsr_pipewire_audio_add_link_from_apps_to_output(self, app_names, num_app_names, sink_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK, true);
+    return gsr_pipewire_audio_add_links_to_output(self, app_names, num_app_names, sink_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_STREAM_OUTPUT, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK, true);
 }
 
 bool gsr_pipewire_audio_add_link_from_sources_to_sink(gsr_pipewire_audio *self, const char **source_names, int num_source_names, const char *sink_name_input) {
-    return gsr_pipewire_audio_add_link_from_apps_to_output(self, source_names, num_source_names, sink_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK, false);
+    return gsr_pipewire_audio_add_links_to_output(self, source_names, num_source_names, sink_name_input, GSR_PIPEWIRE_AUDIO_NODE_TYPE_SINK_OR_SOURCE, GSR_PIPEWIRE_AUDIO_LINK_INPUT_TYPE_SINK, false);
 }
 
 void gsr_pipewire_audio_for_each_app(gsr_pipewire_audio *self, gsr_pipewire_audio_app_query_callback callback, void *userdata) {

src/pipewire_video.c

@@ -14,7 +14,7 @@
 /* This code is partially based on xr-video-player pipewire implementation which is based on obs-studio's pipewire implementation */
 
 /* TODO: Make gsr_pipewire_video_init asynchronous */
-/* TODO: Support 10-bit capture (hdr) when pipewire supports it */
+/* TODO: Support hdr when pipewire supports it */
 /* TODO: Test all of the image formats */
 
 #ifndef SPA_POD_PROP_FLAG_DONT_FIXATE
@@ -65,14 +65,20 @@ static void on_core_done_cb(void *user_data, uint32_t id, int seq) {
 
 static bool is_cursor_format_supported(const enum spa_video_format format) {
     switch(format) {
-        case SPA_VIDEO_FORMAT_RGBx: return true;
-        case SPA_VIDEO_FORMAT_BGRx: return true;
-        case SPA_VIDEO_FORMAT_xRGB: return true;
-        case SPA_VIDEO_FORMAT_xBGR: return true;
-        case SPA_VIDEO_FORMAT_RGBA: return true;
-        case SPA_VIDEO_FORMAT_BGRA: return true;
-        case SPA_VIDEO_FORMAT_ARGB: return true;
-        case SPA_VIDEO_FORMAT_ABGR: return true;
+        case SPA_VIDEO_FORMAT_RGBx:       return true;
+        case SPA_VIDEO_FORMAT_BGRx:       return true;
+        case SPA_VIDEO_FORMAT_RGBA:       return true;
+        case SPA_VIDEO_FORMAT_BGRA:       return true;
+        case SPA_VIDEO_FORMAT_RGB:        return true;
+        case SPA_VIDEO_FORMAT_BGR:        return true;
+        case SPA_VIDEO_FORMAT_ARGB:       return true;
+        case SPA_VIDEO_FORMAT_ABGR:       return true;
+#if PW_CHECK_VERSION(0, 3, 41)
+        case SPA_VIDEO_FORMAT_xRGB_210LE: return true;
+        case SPA_VIDEO_FORMAT_xBGR_210LE: return true;
+        case SPA_VIDEO_FORMAT_ARGB_210LE: return true;
+        case SPA_VIDEO_FORMAT_ABGR_210LE: return true;
+#endif
         default:                    break;
     }
     return false;
@@ -338,24 +344,46 @@ static inline struct spa_pod *build_format(struct spa_pod_builder *b,
 /* For some reason gstreamer formats are in opposite order to drm formats */
 static int64_t spa_video_format_to_drm_format(const enum spa_video_format format) {
     switch(format) {
-        case SPA_VIDEO_FORMAT_RGBx: return DRM_FORMAT_XBGR8888;
-        case SPA_VIDEO_FORMAT_BGRx: return DRM_FORMAT_XRGB8888;
-        case SPA_VIDEO_FORMAT_RGBA: return DRM_FORMAT_ABGR8888;
-        case SPA_VIDEO_FORMAT_BGRA: return DRM_FORMAT_ARGB8888;
-        case SPA_VIDEO_FORMAT_RGB:  return DRM_FORMAT_XBGR8888;
-        case SPA_VIDEO_FORMAT_BGR:  return DRM_FORMAT_XRGB8888;
+        case SPA_VIDEO_FORMAT_RGBx:       return DRM_FORMAT_XBGR8888;
+        case SPA_VIDEO_FORMAT_BGRx:       return DRM_FORMAT_XRGB8888;
+        case SPA_VIDEO_FORMAT_RGBA:       return DRM_FORMAT_ABGR8888;
+        case SPA_VIDEO_FORMAT_BGRA:       return DRM_FORMAT_ARGB8888;
+        case SPA_VIDEO_FORMAT_RGB:        return DRM_FORMAT_XBGR8888;
+        case SPA_VIDEO_FORMAT_BGR:        return DRM_FORMAT_XRGB8888;
+        case SPA_VIDEO_FORMAT_ARGB:       return DRM_FORMAT_XRGB8888;
+        case SPA_VIDEO_FORMAT_ABGR:       return DRM_FORMAT_XRGB8888;
+#if PW_CHECK_VERSION(0, 3, 41)
+        case SPA_VIDEO_FORMAT_xRGB_210LE: return DRM_FORMAT_XRGB2101010;
+        case SPA_VIDEO_FORMAT_xBGR_210LE: return DRM_FORMAT_XBGR2101010;
+        case SPA_VIDEO_FORMAT_ARGB_210LE: return DRM_FORMAT_ARGB2101010;
+        case SPA_VIDEO_FORMAT_ABGR_210LE: return DRM_FORMAT_ABGR2101010;
+#endif
         default:                    break;
     }
     return DRM_FORMAT_INVALID;
 }
 
-static const enum spa_video_format video_formats[] = {
+#if PW_CHECK_VERSION(0, 3, 41)
+#define GSR_PIPEWIRE_VIDEO_NUM_VIDEO_FORMATS GSR_PIPEWIRE_VIDEO_MAX_VIDEO_FORMATS
+#else
+#define GSR_PIPEWIRE_VIDEO_NUM_VIDEO_FORMATS 8
+#endif
+
+static const enum spa_video_format video_formats[GSR_PIPEWIRE_VIDEO_MAX_VIDEO_FORMATS] = {
     SPA_VIDEO_FORMAT_BGRA,
     SPA_VIDEO_FORMAT_BGRx,
     SPA_VIDEO_FORMAT_BGR,
     SPA_VIDEO_FORMAT_RGBx,
     SPA_VIDEO_FORMAT_RGBA,
     SPA_VIDEO_FORMAT_RGB,
+    SPA_VIDEO_FORMAT_ARGB,
+    SPA_VIDEO_FORMAT_ABGR,
+#if PW_CHECK_VERSION(0, 3, 41)
+    SPA_VIDEO_FORMAT_xRGB_210LE,
+    SPA_VIDEO_FORMAT_xBGR_210LE,
+    SPA_VIDEO_FORMAT_ARGB_210LE,
+    SPA_VIDEO_FORMAT_ABGR_210LE
+#endif
 };
 
 static bool gsr_pipewire_video_build_format_params(gsr_pipewire_video *self, struct spa_pod_builder *pod_builder, struct spa_pod **params, uint32_t *num_params) {
@@ -367,7 +395,7 @@ static bool gsr_pipewire_video_build_format_params(gsr_pipewire_video *self, str
     for(size_t i = 0; i < GSR_PIPEWIRE_VIDEO_NUM_VIDEO_FORMATS; i++) {
         if(self->supported_video_formats[i].modifiers_size == 0)
             continue;
-        params[i] = build_format(pod_builder, &self->video_info, self->supported_video_formats[i].format, self->modifiers + self->supported_video_formats[i].modifiers_index, self->supported_video_formats[i].modifiers_size);
+        params[*num_params] = build_format(pod_builder, &self->video_info, self->supported_video_formats[i].format, self->modifiers + self->supported_video_formats[i].modifiers_index, self->supported_video_formats[i].modifiers_size);
         ++(*num_params);
     }
 
@@ -382,9 +410,10 @@ static void renegotiate_format(void *data, uint64_t expirations) {
 
     struct spa_pod *params[GSR_PIPEWIRE_VIDEO_NUM_VIDEO_FORMATS];
     uint32_t num_video_formats = 0;
-    uint8_t params_buffer[2048];
+    uint8_t params_buffer[4096];
     struct spa_pod_builder pod_builder = SPA_POD_BUILDER_INIT(params_buffer, sizeof(params_buffer));
     if (!gsr_pipewire_video_build_format_params(self, &pod_builder, params, &num_video_formats)) {
+        fprintf(stderr, "gsr error: renegotiate_format: failed to build formats\n");
         pw_thread_loop_unlock(self->thread_loop);
         return;
     }
@@ -413,6 +442,11 @@ static bool spa_video_format_get_modifiers(gsr_pipewire_video *self, const enum
     }
 
     const int64_t drm_format = spa_video_format_to_drm_format(format);
+    if(drm_format == DRM_FORMAT_INVALID) {
+        fprintf(stderr, "gsr error: spa_video_format_get_modifiers: unsupported format: %d\n", (int)format);
+        return false;
+    }
+
     if(!self->egl->eglQueryDmaBufModifiersEXT(self->egl->egl_display, drm_format, max_modifiers, modifiers, NULL, num_modifiers)) {
         fprintf(stderr, "gsr error: spa_video_format_get_modifiers: eglQueryDmaBufModifiersEXT failed with drm format %d, %" PRIi64 "\n", (int)format, drm_format);
         //modifiers[0] = DRM_FORMAT_MOD_LINEAR;
@@ -443,7 +477,7 @@ static void gsr_pipewire_video_init_modifiers(gsr_pipewire_video *self) {
 static bool gsr_pipewire_video_setup_stream(gsr_pipewire_video *self) {
     struct spa_pod *params[GSR_PIPEWIRE_VIDEO_NUM_VIDEO_FORMATS];
     uint32_t num_video_formats = 0;
-    uint8_t params_buffer[2048];
+    uint8_t params_buffer[4096];
     struct spa_pod_builder pod_builder = SPA_POD_BUILDER_INIT(params_buffer, sizeof(params_buffer));
 
     self->thread_loop = pw_thread_loop_new("gsr screen capture", NULL);
@@ -476,6 +510,9 @@ static bool gsr_pipewire_video_setup_stream(gsr_pipewire_video *self) {
     // TODO: Error check
     pw_core_add_listener(self->core, &self->core_listener, &core_events, self);
 
+    self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, 0);
+    pw_thread_loop_wait(self->thread_loop);
+
     gsr_pipewire_video_init_modifiers(self);
 
     // TODO: Cleanup?
@@ -486,9 +523,6 @@ static bool gsr_pipewire_video_setup_stream(gsr_pipewire_video *self) {
         goto error;
     }
 
-    self->server_version_sync = pw_core_sync(self->core, PW_ID_CORE, 0);
-    pw_thread_loop_wait(self->thread_loop);
-
     self->stream = pw_stream_new(self->core, "com.dec05eba.gpu_screen_recorder",
         pw_properties_new(PW_KEY_MEDIA_TYPE, "Video",
                           PW_KEY_MEDIA_CATEGORY, "Capture",
@@ -702,11 +736,13 @@ static void gsr_pipewire_video_update_cursor_texture(gsr_pipewire_video *self, g
     if(!self->cursor.data)
         return;
 
+    const float border_color[4] = { 0.0f, 0.0f, 0.0f, 0.0f };
     self->egl->glBindTexture(GL_TEXTURE_2D, texture_map.cursor_texture_id);
     // TODO: glTextureSubImage2D if same size
     self->egl->glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA8, self->cursor.width, self->cursor.height, 0, GL_RGBA, GL_UNSIGNED_BYTE, self->cursor.data);
-    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->egl->glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, border_color);
     self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);
     self->egl->glBindTexture(GL_TEXTURE_2D, 0);

src/shader.c

@@ -36,28 +36,36 @@ static unsigned int loader_shader(gsr_egl *egl, unsigned int type, const char *s
     return shader_id;
 }
 
-static unsigned int load_program(gsr_egl *egl, const char *vertex_shader, const char *fragment_shader) {
+static unsigned int load_program(gsr_egl *egl, const char *vertex_shader, const char *fragment_shader, const char *compute_shader) {
     unsigned int vertex_shader_id = 0;
     unsigned int fragment_shader_id = 0;
+    unsigned int compute_shader_id = 0;
     unsigned int program_id = 0;
     int linked = 0;
+    bool success = false;
 
     if(vertex_shader) {
         vertex_shader_id = loader_shader(egl, GL_VERTEX_SHADER, vertex_shader);
         if(vertex_shader_id == 0)
-            goto err;
+            goto done;
     }
 
     if(fragment_shader) {
         fragment_shader_id = loader_shader(egl, GL_FRAGMENT_SHADER, fragment_shader);
         if(fragment_shader_id == 0)
-            goto err;
+            goto done;
+    }
+
+    if(compute_shader) {
+        compute_shader_id = loader_shader(egl, GL_COMPUTE_SHADER, compute_shader);
+        if(compute_shader_id == 0)
+            goto done;
     }
 
     program_id = egl->glCreateProgram();
     if(program_id == 0) {
         fprintf(stderr, "gsr error: load_program: failed to create shader program, error: %d\n", egl->glGetError());
-        goto err;
+        goto done;
     }
 
     if(vertex_shader_id)
@@ -66,6 +74,9 @@ static unsigned int load_program(gsr_egl *egl, const char *vertex_shader, const
     if(fragment_shader_id)
         egl->glAttachShader(program_id, fragment_shader_id);
 
+    if(compute_shader_id)
+        egl->glAttachShader(program_id, compute_shader_id);
+
     egl->glLinkProgram(program_id);
 
     egl->glGetProgramiv(program_id, GL_LINK_STATUS, &linked);
@@ -79,37 +90,36 @@ static unsigned int load_program(gsr_egl *egl, const char *vertex_shader, const
             fprintf(stderr, "gsr error: load program: linking shader program failed, error:\n%s\n", info_log);            
         }
 
-        goto err;
+        goto done;
     }
 
+    success = true;
+    done:
+
+    if(!success) {
+        if(program_id)
+            egl->glDeleteProgram(program_id);
+    }
+    if(compute_shader_id)
+        egl->glDeleteShader(compute_shader_id);
     if(fragment_shader_id)
         egl->glDeleteShader(fragment_shader_id);
     if(vertex_shader_id)
         egl->glDeleteShader(vertex_shader_id);
-
     return program_id;
-
-    err:
-    if(program_id)
-        egl->glDeleteProgram(program_id);
-    if(fragment_shader_id)
-        egl->glDeleteShader(fragment_shader_id);
-    if(vertex_shader_id)
-        egl->glDeleteShader(vertex_shader_id);
-    return 0;
 }
 
-int gsr_shader_init(gsr_shader *self, gsr_egl *egl, const char *vertex_shader, const char *fragment_shader) {
+int gsr_shader_init(gsr_shader *self, gsr_egl *egl, const char *vertex_shader, const char *fragment_shader, const char *compute_shader) {
     assert(egl);
     self->egl = egl;
     self->program_id = 0;
 
-    if(!vertex_shader && !fragment_shader) {
-        fprintf(stderr, "gsr error: gsr_shader_init: vertex shader and fragment shader can't be NULL at the same time\n");
+    if(!vertex_shader && !fragment_shader && !compute_shader) {
+        fprintf(stderr, "gsr error: gsr_shader_init: vertex, fragment shader and compute shaders can't be NULL at the same time\n");
         return -1;
     }
 
-    self->program_id = load_program(self->egl, vertex_shader, fragment_shader);
+    self->program_id = load_program(self->egl, vertex_shader, fragment_shader, compute_shader);
     if(self->program_id == 0)
         return -1;
 

src/sound.cpp

@@ -8,12 +8,16 @@ extern "C" {
 #include <string.h>
 #include <cmath>
 #include <time.h>
+#include <mutex>
 
 #include <pulse/pulseaudio.h>
 #include <pulse/mainloop.h>
 #include <pulse/xmalloc.h>
 #include <pulse/error.h>
 
+#define RECONNECT_TRY_TIMEOUT_SECONDS 0.5
+#define DEVICE_NAME_MAX_SIZE 128
+
 #define CHECK_DEAD_GOTO(p, rerror, label)                               \
     do {                                                                \
         if (!(p)->context || !PA_CONTEXT_IS_GOOD(pa_context_get_state((p)->context)) || \
@@ -29,6 +33,12 @@ extern "C" {
         }                                                               \
     } while(false);
 
+enum class DeviceType {
+    STANDARD,
+    DEFAULT_OUTPUT,
+    DEFAULT_INPUT
+};
+
 struct pa_handle {
     pa_context *context;
     pa_stream *stream;
@@ -42,6 +52,19 @@ struct pa_handle {
 
     int operation_success;
     double latency_seconds;
+
+    pa_buffer_attr attr;
+    pa_sample_spec ss;
+
+    std::mutex reconnect_mutex;
+    DeviceType device_type;
+    char stream_name[256];
+    bool reconnect;
+    double reconnect_last_tried_seconds;
+
+    char device_name[DEVICE_NAME_MAX_SIZE];
+    char default_output_device_name[DEVICE_NAME_MAX_SIZE];
+    char default_input_device_name[DEVICE_NAME_MAX_SIZE];
 };
 
 static void pa_sound_device_free(pa_handle *p) {
@@ -71,22 +94,110 @@ static void pa_sound_device_free(pa_handle *p) {
     pa_xfree(p);
 }
 
+static void subscribe_update_default_devices(pa_context*, const pa_server_info *server_info, void *userdata) {
+    pa_handle *handle = (pa_handle*)userdata;
+    std::lock_guard<std::mutex> lock(handle->reconnect_mutex);
+
+    if(server_info->default_sink_name) {
+        // TODO: Size check
+        snprintf(handle->default_output_device_name, sizeof(handle->default_output_device_name), "%s.monitor", server_info->default_sink_name);
+        if(handle->device_type == DeviceType::DEFAULT_OUTPUT && strcmp(handle->device_name, handle->default_output_device_name) != 0) {
+            handle->reconnect = true;
+            handle->reconnect_last_tried_seconds = clock_get_monotonic_seconds();
+            // TODO: Size check
+            snprintf(handle->device_name, sizeof(handle->device_name), "%s", handle->default_output_device_name);
+        }
+    }
+
+    if(server_info->default_source_name) {
+        // TODO: Size check
+        snprintf(handle->default_input_device_name, sizeof(handle->default_input_device_name), "%s", server_info->default_source_name);
+        if(handle->device_type == DeviceType::DEFAULT_INPUT && strcmp(handle->device_name, handle->default_input_device_name) != 0) {
+            handle->reconnect = true;
+            handle->reconnect_last_tried_seconds = clock_get_monotonic_seconds();
+            // TODO: Size check
+            snprintf(handle->device_name, sizeof(handle->device_name), "%s", handle->default_input_device_name);
+        }
+    }
+}
+
+static void subscribe_cb(pa_context *c, pa_subscription_event_type_t t, uint32_t idx, void *userdata) {
+    (void)idx;
+    pa_handle *handle = (pa_handle*)userdata;
+    if((t & PA_SUBSCRIPTION_EVENT_FACILITY_MASK) == PA_SUBSCRIPTION_EVENT_SERVER) {
+        pa_operation *pa = pa_context_get_server_info(c, subscribe_update_default_devices, handle);
+        if(pa)
+            pa_operation_unref(pa);
+    }
+}
+
+static void store_default_devices(pa_context*, const pa_server_info *server_info, void *userdata) {
+    pa_handle *handle = (pa_handle*)userdata;
+    if(server_info->default_sink_name)
+        snprintf(handle->default_output_device_name, sizeof(handle->default_output_device_name), "%s.monitor", server_info->default_sink_name);
+    if(server_info->default_source_name)
+        snprintf(handle->default_input_device_name, sizeof(handle->default_input_device_name), "%s", server_info->default_source_name);
+}
+
+static bool startup_get_default_devices(pa_handle *p, const char *device_name) {
+    pa_operation *pa = pa_context_get_server_info(p->context, store_default_devices, p);
+    while(pa) {
+        pa_operation_state state = pa_operation_get_state(pa);
+        if(state == PA_OPERATION_DONE) {
+            pa_operation_unref(pa);
+            break;
+        } else if(state == PA_OPERATION_CANCELLED) {
+            pa_operation_unref(pa);
+            return false;
+        }
+        pa_mainloop_iterate(p->mainloop, 1, NULL);
+    }
+
+    if(p->default_output_device_name[0] == '\0') {
+        fprintf(stderr, "Error: failed to find default audio output device\n");
+        return false;
+    }
+
+    if(strcmp(device_name, "default_output") == 0) {
+        snprintf(p->device_name, sizeof(p->device_name), "%s", p->default_output_device_name);
+        p->device_type = DeviceType::DEFAULT_OUTPUT;
+    } else if(strcmp(device_name, "default_input") == 0) {
+        snprintf(p->device_name, sizeof(p->device_name), "%s", p->default_input_device_name);
+        p->device_type = DeviceType::DEFAULT_INPUT;
+    } else {
+        snprintf(p->device_name, sizeof(p->device_name), "%s", device_name);
+        p->device_type = DeviceType::STANDARD;
+    }
+
+    return true;
+}
+
 static pa_handle* pa_sound_device_new(const char *server,
         const char *name,
-        const char *dev,
+        const char *device_name,
         const char *stream_name,
         const pa_sample_spec *ss,
         const pa_buffer_attr *attr,
         int *rerror) {
     pa_handle *p;
-    int error = PA_ERR_INTERNAL, r;
+    int error = PA_ERR_INTERNAL;
+    pa_operation *pa = NULL;
 
     p = pa_xnew0(pa_handle, 1);
+    p->attr = *attr;
+    p->ss = *ss;
+    snprintf(p->stream_name, sizeof(p->stream_name), "%s", stream_name);
+
+    p->reconnect = true;
+    p->reconnect_last_tried_seconds = clock_get_monotonic_seconds() - 1000.0;
+    p->default_output_device_name[0] = '\0';
+    p->default_input_device_name[0] = '\0';
+    p->device_type = DeviceType::STANDARD;
 
     const int buffer_size = attr->fragsize;
     void *buffer = malloc(buffer_size);
     if(!buffer) {
-        fprintf(stderr, "failed to allocate buffer for audio\n");
+        fprintf(stderr, "Error: failed to allocate buffer for audio\n");
         *rerror = -1;
         return NULL;
     }
@@ -120,32 +231,13 @@ static pa_handle* pa_sound_device_new(const char *server,
         pa_mainloop_iterate(p->mainloop, 1, NULL);
     }
 
-    if (!(p->stream = pa_stream_new(p->context, stream_name, ss, NULL))) {
-        error = pa_context_errno(p->context);
+    if(!startup_get_default_devices(p, device_name))
         goto fail;
-    }
 
-    r = pa_stream_connect_record(p->stream, dev, attr,
-        (pa_stream_flags_t)(PA_STREAM_INTERPOLATE_TIMING|PA_STREAM_ADJUST_LATENCY|PA_STREAM_AUTO_TIMING_UPDATE));
-
-    if (r < 0) {
-        error = pa_context_errno(p->context);
-        goto fail;
-    }
-
-    for (;;) {
-        pa_stream_state_t state = pa_stream_get_state(p->stream);
-
-        if (state == PA_STREAM_READY)
-            break;
-
-        if (!PA_STREAM_IS_GOOD(state)) {
-            error = pa_context_errno(p->context);
-            goto fail;
-        }
-
-        pa_mainloop_iterate(p->mainloop, 1, NULL);
-    }
+    pa_context_set_subscribe_callback(p->context, subscribe_cb, p);
+    pa = pa_context_subscribe(p->context, PA_SUBSCRIPTION_MASK_SERVER, NULL, NULL);
+    if(pa)
+        pa_operation_unref(pa);
 
     return p;
 
@@ -156,10 +248,65 @@ fail:
     return NULL;
 }
 
+static bool pa_sound_device_should_reconnect(pa_handle *p, double now, char *device_name, size_t device_name_size) {
+    std::lock_guard<std::mutex> lock(p->reconnect_mutex);
+    if(p->reconnect && now - p->reconnect_last_tried_seconds >= RECONNECT_TRY_TIMEOUT_SECONDS) {
+        p->reconnect_last_tried_seconds = now;
+        // TODO: Size check
+        snprintf(device_name, device_name_size, "%s", p->device_name);
+        return true;
+    }
+    return false;
+}
+
+static bool pa_sound_device_handle_reconnect(pa_handle *p, char *device_name, size_t device_name_size, double now) {
+    int r;
+    if(!pa_sound_device_should_reconnect(p, now, device_name, device_name_size))
+        return true;
+
+    if(p->stream) {
+        pa_stream_disconnect(p->stream);
+        pa_stream_unref(p->stream);
+        p->stream = NULL;
+    }
+
+    if(!(p->stream = pa_stream_new(p->context, p->stream_name, &p->ss, NULL))) {
+        //pa_context_errno(p->context);
+        return false;
+    }
+
+    r = pa_stream_connect_record(p->stream, device_name, &p->attr,
+        (pa_stream_flags_t)(PA_STREAM_INTERPOLATE_TIMING|PA_STREAM_ADJUST_LATENCY|PA_STREAM_AUTO_TIMING_UPDATE));
+
+    if(r < 0) {
+        //pa_context_errno(p->context);
+        return false;
+    }
+
+    for(;;) {
+        pa_stream_state_t state = pa_stream_get_state(p->stream);
+
+        if(state == PA_STREAM_READY)
+            break;
+
+        if(!PA_STREAM_IS_GOOD(state)) {
+            //pa_context_errno(p->context);
+            return false;
+        }
+
+        pa_mainloop_iterate(p->mainloop, 1, NULL);
+    }
+
+    std::lock_guard<std::mutex> lock(p->reconnect_mutex);
+    p->reconnect = false;
+    return true;
+}
+
 static int pa_sound_device_read(pa_handle *p, double timeout_seconds) {
     assert(p);
 
     const double start_time = clock_get_monotonic_seconds();
+    char device_name[DEVICE_NAME_MAX_SIZE];
 
     bool success = false;
     int r = 0;
@@ -167,6 +314,9 @@ static int pa_sound_device_read(pa_handle *p, double timeout_seconds) {
     pa_usec_t latency = 0;
     int negative = 0;
 
+    if(!pa_sound_device_handle_reconnect(p, device_name, sizeof(device_name), start_time))
+        goto fail;
+
     CHECK_DEAD_GOTO(p, rerror, fail);
 
     while (p->output_index < p->output_length) {
@@ -276,7 +426,7 @@ int sound_device_get_by_name(SoundDevice *device, const char *device_name, const
     int error = 0;
     pa_handle *handle = pa_sound_device_new(nullptr, description, device_name, description, &ss, &buffer_attr, &error);
     if(!handle) {
-        fprintf(stderr, "pa_sound_device_new() failed: %s. Audio input device %s might not be valid\n", pa_strerror(error), device_name);
+        fprintf(stderr, "Error: pa_sound_device_new() failed: %s. Audio input device %s might not be valid\n", pa_strerror(error), device_name);
         return -1;
     }
 

src/utils.c

@@ -14,10 +14,8 @@
 
 #include <xf86drmMode.h>
 #include <xf86drm.h>
-#include <libdrm/drm_fourcc.h>
 #include <X11/Xatom.h>
 #include <X11/extensions/Xrandr.h>
-#include <va/va_drmcommon.h>
 #include <libavcodec/avcodec.h>
 #include <libavutil/hwcontext_vaapi.h>
 
@@ -289,6 +287,7 @@ bool get_monitor_by_name(const gsr_egl *egl, gsr_connection_type connection_type
 typedef struct {
     const gsr_monitor *monitor;
     gsr_monitor_rotation rotation;
+    vec2i position;
     bool match_found;
 } get_monitor_by_connector_id_userdata;
 
@@ -300,6 +299,7 @@ static void get_monitor_by_name_and_size_callback(const gsr_monitor *monitor, vo
     get_monitor_by_connector_id_userdata *data = (get_monitor_by_connector_id_userdata*)userdata;
     if(monitor->name && data->monitor->name && strcmp(monitor->name, data->monitor->name) == 0 && vec2i_eql(monitor->size, data->monitor->size)) {
         data->rotation = monitor->rotation;
+        data->position = monitor->pos;
         data->match_found = true;
     }
 }
@@ -310,39 +310,51 @@ static void get_monitor_by_connector_id_callback(const gsr_monitor *monitor, voi
         (!monitor->connector_id && monitor->monitor_identifier == data->monitor->monitor_identifier))
     {
         data->rotation = monitor->rotation;
+        data->position = monitor->pos;
         data->match_found = true;
     }
 }
 
-gsr_monitor_rotation drm_monitor_get_display_server_rotation(const gsr_window *window, const gsr_monitor *monitor) {
+bool drm_monitor_get_display_server_data(const gsr_window *window, const gsr_monitor *monitor, gsr_monitor_rotation *monitor_rotation, vec2i *monitor_position) {
+    *monitor_rotation = GSR_MONITOR_ROT_0;
+    *monitor_position = (vec2i){0, 0};
+
     if(gsr_window_get_display_server(window) == GSR_DISPLAY_SERVER_WAYLAND) {
         {
             get_monitor_by_connector_id_userdata userdata;
             userdata.monitor = monitor;
             userdata.rotation = GSR_MONITOR_ROT_0;
+            userdata.position = (vec2i){0, 0};
             userdata.match_found = false;
             gsr_window_for_each_active_monitor_output_cached(window, get_monitor_by_name_and_size_callback, &userdata);
-            if(userdata.match_found)
-                return userdata.rotation;
+            if(userdata.match_found) {
+                *monitor_rotation = userdata.rotation;
+                *monitor_position = userdata.position;
+                return true;
+            }
         }
         {
             get_monitor_by_connector_id_userdata userdata;
             userdata.monitor = monitor;
             userdata.rotation = GSR_MONITOR_ROT_0;
+            userdata.position = (vec2i){0, 0};
             userdata.match_found = false;
             gsr_window_for_each_active_monitor_output_cached(window, get_monitor_by_connector_id_callback, &userdata);
-            return userdata.rotation;
+            *monitor_rotation = userdata.rotation;
+            *monitor_position = userdata.position;
+            return userdata.match_found;
         }
     } else {
         get_monitor_by_connector_id_userdata userdata;
         userdata.monitor = monitor;
         userdata.rotation = GSR_MONITOR_ROT_0;
+        userdata.position = (vec2i){0, 0};
         userdata.match_found = false;
         gsr_window_for_each_active_monitor_output_cached(window, get_monitor_by_connector_id_callback, &userdata);
-        return userdata.rotation;
+        *monitor_rotation = userdata.rotation;
+        *monitor_position = userdata.position;
+        return userdata.match_found;
     }
-
-    return GSR_MONITOR_ROT_0;
 }
 
 bool gl_get_gpu_info(gsr_egl *egl, gsr_gpu_info *info) {
@@ -382,6 +394,8 @@ bool gl_get_gpu_info(gsr_egl *egl, gsr_gpu_info *info) {
         info->vendor = GSR_GPU_VENDOR_INTEL;
     else if(strstr((const char*)gl_vendor, "NVIDIA"))
         info->vendor = GSR_GPU_VENDOR_NVIDIA;
+    else if(strstr((const char*)gl_vendor, "Broadcom"))
+        info->vendor = GSR_GPU_VENDOR_BROADCOM;
     else {
         fprintf(stderr, "gsr error: unknown gpu vendor: %s\n", gl_vendor);
         supported = false;
@@ -413,14 +427,6 @@ bool gl_get_gpu_info(gsr_egl *egl, gsr_gpu_info *info) {
     return supported;
 }
 
-static bool version_greater_than(int major, int minor, int patch, int other_major, int other_minor, int other_patch) {
-    return (major > other_major) || (major == other_major && minor > other_minor) || (major == other_major && minor == other_minor && patch > other_patch);
-}
-
-bool gl_driver_version_greater_than(const gsr_gpu_info *gpu_info, int major, int minor, int patch) {
-    return version_greater_than(gpu_info->driver_major, gpu_info->driver_minor, gpu_info->driver_patch, major, minor, patch);
-}
-
 bool try_card_has_valid_plane(const char *card_path) {
     drmVersion *ver = NULL;
     drmModePlaneResPtr planes = NULL;
@@ -635,250 +641,18 @@ static VADisplay video_codec_context_get_vaapi_display(AVCodecContext *video_cod
 }
 
 bool video_codec_context_is_vaapi(AVCodecContext *video_codec_context) {
+    if(!video_codec_context)
+        return false;
+
     AVBufferRef *hw_frames_ctx = video_codec_context->hw_frames_ctx;
     if(!hw_frames_ctx)
-        return NULL;
+        return false;
 
     AVHWFramesContext *hw_frame_context = (AVHWFramesContext*)hw_frames_ctx->data;
     AVHWDeviceContext *device_context = (AVHWDeviceContext*)hw_frame_context->device_ctx;
     return device_context->type == AV_HWDEVICE_TYPE_VAAPI;
 }
 
-static uint32_t drm_fourcc_to_va_fourcc(uint32_t drm_fourcc) {
-    switch(drm_fourcc) {
-        case DRM_FORMAT_XRGB8888: return VA_FOURCC_BGRX;
-        case DRM_FORMAT_XBGR8888: return VA_FOURCC_RGBX;
-        case DRM_FORMAT_RGBX8888: return VA_FOURCC_XBGR;
-        case DRM_FORMAT_BGRX8888: return VA_FOURCC_XRGB;
-        case DRM_FORMAT_ARGB8888: return VA_FOURCC_BGRA;
-        case DRM_FORMAT_ABGR8888: return VA_FOURCC_RGBA;
-        case DRM_FORMAT_RGBA8888: return VA_FOURCC_ABGR;
-        case DRM_FORMAT_BGRA8888: return VA_FOURCC_ARGB;
-        default:                  return drm_fourcc;
-    }
-}
-
-bool vaapi_copy_drm_planes_to_video_surface(AVCodecContext *video_codec_context, AVFrame *video_frame, vec2i source_pos, vec2i source_size, vec2i dest_pos, vec2i dest_size, uint32_t format, vec2i size, const int *fds, const uint32_t *offsets, const uint32_t *pitches, const uint64_t *modifiers, int num_planes) {
-    VAConfigID config_id = 0;
-    VAContextID context_id = 0;
-    VASurfaceID input_surface_id = 0;
-    VABufferID buffer_id = 0;
-    bool success = true;
-
-    VADisplay va_dpy = video_codec_context_get_vaapi_display(video_codec_context);
-    if(!va_dpy) {
-        success = false;
-        goto done;
-    }
-
-    VAStatus va_status = vaCreateConfig(va_dpy, VAProfileNone, VAEntrypointVideoProc, NULL, 0, &config_id);
-    if(va_status != VA_STATUS_SUCCESS) {
-        fprintf(stderr, "gsr error: vaapi_copy_drm_planes_to_video_surface: vaCreateConfig failed, error: %s\n", vaErrorStr(va_status));
-        success = false;
-        goto done;
-    }
-
-    VASurfaceID output_surface_id = (uintptr_t)video_frame->data[3];
-    va_status = vaCreateContext(va_dpy, config_id, size.x, size.y, VA_PROGRESSIVE, &output_surface_id, 1, &context_id);
-    if(va_status != VA_STATUS_SUCCESS) {
-        fprintf(stderr, "gsr error: vaapi_copy_drm_planes_to_video_surface: vaCreateContext failed, error: %s\n", vaErrorStr(va_status));
-        success = false;
-        goto done;
-    }
-
-    VADRMPRIMESurfaceDescriptor buf = {0};
-    buf.fourcc = drm_fourcc_to_va_fourcc(format);//VA_FOURCC_BGRX; // TODO: VA_FOURCC_BGRA, VA_FOURCC_X2R10G10B10
-    buf.width = size.x;
-    buf.height = size.y;
-    buf.num_objects = num_planes;
-    buf.num_layers = 1;
-    buf.layers[0].drm_format = format;
-    buf.layers[0].num_planes = buf.num_objects;
-    for(int i = 0; i < num_planes; ++i) {
-        buf.objects[i].fd = fds[i];
-        buf.objects[i].size = size.y * pitches[i]; // TODO:
-        buf.objects[i].drm_format_modifier = modifiers[i];
-
-        buf.layers[0].object_index[i] = i;
-        buf.layers[0].offset[i] = offsets[i];
-        buf.layers[0].pitch[i] = pitches[i];
-    }
-
-    VASurfaceAttrib attribs[2] = {0};
-    attribs[0].type = VASurfaceAttribMemoryType;
-    attribs[0].flags = VA_SURFACE_ATTRIB_SETTABLE;
-    attribs[0].value.type = VAGenericValueTypeInteger;
-    attribs[0].value.value.i = VA_SURFACE_ATTRIB_MEM_TYPE_DRM_PRIME_2;
-    attribs[1].type = VASurfaceAttribExternalBufferDescriptor;
-    attribs[1].flags = VA_SURFACE_ATTRIB_SETTABLE;
-    attribs[1].value.type = VAGenericValueTypePointer;
-    attribs[1].value.value.p = &buf;
-    
-    // TODO: RT_FORMAT with 10 bit/hdr, VA_RT_FORMAT_RGB32_10
-    // TODO: Max size same as source_size
-    va_status = vaCreateSurfaces(va_dpy, VA_RT_FORMAT_RGB32, size.x, size.y, &input_surface_id, 1, attribs, 2);
-    if(va_status != VA_STATUS_SUCCESS) {
-        fprintf(stderr, "gsr error: vaapi_copy_drm_planes_to_video_surface: vaCreateSurfaces failed, error: %s\n", vaErrorStr(va_status));
-        success = false;
-        goto done;
-    }
-
-    const VARectangle source_region = {
-        .x = source_pos.x,
-        .y = source_pos.y,
-        .width = source_size.x,
-        .height = source_size.y
-    };
-
-    const VARectangle output_region = {
-        .x = dest_pos.x,
-        .y = dest_pos.y,
-        .width = dest_size.x,
-        .height = dest_size.y
-    };
-
-    const bool scaled = dest_size.x != source_size.x || dest_size.y != source_size.y;
-
-    // Copying a surface to another surface will automatically perform the color conversion. Thanks vaapi!
-    VAProcPipelineParameterBuffer params = {0};
-    params.surface = input_surface_id;
-    params.surface_region = NULL;
-    params.surface_region = &source_region;
-    params.output_region = &output_region;
-    params.output_background_color = 0;
-    params.filter_flags = scaled ? (VA_FILTER_SCALING_HQ | VA_FILTER_INTERPOLATION_BILINEAR) : 0;
-    params.pipeline_flags = VA_PROC_PIPELINE_FAST;
-
-    params.input_color_properties.colour_primaries = 1;
-    params.input_color_properties.transfer_characteristics = 1;
-    params.input_color_properties.matrix_coefficients = 1;
-    params.surface_color_standard = VAProcColorStandardBT709; // TODO:
-    params.input_color_properties.color_range = video_frame->color_range == AVCOL_RANGE_JPEG ? VA_SOURCE_RANGE_FULL : VA_SOURCE_RANGE_REDUCED;
-
-    params.output_color_properties.colour_primaries = 1;
-    params.output_color_properties.transfer_characteristics = 1;
-    params.output_color_properties.matrix_coefficients = 1;
-    params.output_color_standard = VAProcColorStandardBT709; // TODO:
-    params.output_color_properties.color_range = video_frame->color_range == AVCOL_RANGE_JPEG ? VA_SOURCE_RANGE_FULL : VA_SOURCE_RANGE_REDUCED;
-
-    params.processing_mode = VAProcPerformanceMode;
-
-    // VAProcPipelineCaps pipeline_caps = {0};
-    // va_status = vaQueryVideoProcPipelineCaps(self->va_dpy,
-    //                                    self->context_id,
-    //                                    NULL, 0,
-    //                                    &pipeline_caps);
-    // if(va_status == VA_STATUS_SUCCESS) {
-    //     fprintf(stderr, "pipeline_caps: %u, %u\n", (unsigned int)pipeline_caps.rotation_flags, pipeline_caps.blend_flags);
-    // }
-
-    // TODO: params.output_hdr_metadata
-
-    // TODO:
-    // if (first surface to render)
-    //     pipeline_param->output_background_color = 0xff000000; // black
-
-    va_status = vaCreateBuffer(va_dpy, context_id, VAProcPipelineParameterBufferType, sizeof(params), 1, &params, &buffer_id);
-    if(va_status != VA_STATUS_SUCCESS) {
-        fprintf(stderr, "gsr error: vaapi_copy_drm_planes_to_video_surface: vaCreateBuffer failed, error: %d\n", va_status);
-        success = false;
-        goto done;
-    }
-
-    va_status = vaBeginPicture(va_dpy, context_id, output_surface_id);
-    if(va_status != VA_STATUS_SUCCESS) {
-        fprintf(stderr, "gsr error: vaapi_copy_drm_planes_to_video_surface: vaBeginPicture failed, error: %d\n", va_status);
-        success = false;
-        goto done;
-    }
-
-    va_status = vaRenderPicture(va_dpy, context_id, &buffer_id, 1);
-    if(va_status != VA_STATUS_SUCCESS) {
-        vaEndPicture(va_dpy, context_id);
-        fprintf(stderr, "gsr error: vaapi_copy_drm_planes_to_video_surface: vaRenderPicture failed, error: %d\n", va_status);
-        success = false;
-        goto done;
-    }
-
-    va_status = vaEndPicture(va_dpy, context_id);
-    if(va_status != VA_STATUS_SUCCESS) {
-        fprintf(stderr, "gsr error: vaapi_copy_drm_planes_to_video_surface: vaEndPicture failed, error: %d\n", va_status);
-        success = false;
-        goto done;
-    }
-
-    // vaSyncBuffer(va_dpy, buffer_id, 1000 * 1000 * 1000);
-    // vaSyncSurface(va_dpy, input_surface_id);
-    // vaSyncSurface(va_dpy, output_surface_id);
-
-    done:
-    if(buffer_id)
-        vaDestroyBuffer(va_dpy, buffer_id);
-
-    if(input_surface_id)
-        vaDestroySurfaces(va_dpy, &input_surface_id, 1);
-
-    if(context_id)
-        vaDestroyContext(va_dpy, context_id);
-
-    if(config_id)
-        vaDestroyConfig(va_dpy, config_id);
-
-    return success;
-}
-
-bool vaapi_copy_egl_image_to_video_surface(gsr_egl *egl, EGLImage image, vec2i source_pos, vec2i source_size, vec2i dest_pos, vec2i dest_size, AVCodecContext *video_codec_context, AVFrame *video_frame) {
-    if(!image)
-        return false;
-
-    int texture_fourcc = 0;
-    int texture_num_planes = 0;
-    uint64_t texture_modifiers = 0;
-    if(!egl->eglExportDMABUFImageQueryMESA(egl->egl_display, image, &texture_fourcc, &texture_num_planes, &texture_modifiers)) {
-        fprintf(stderr, "gsr error: gsr_capture_xcomposite_vaapi_tick: eglExportDMABUFImageQueryMESA failed\n");
-        return false;
-    }
-
-    if(texture_num_planes <= 0 || texture_num_planes > 8) {
-        fprintf(stderr, "gsr error: gsr_capture_xcomposite_vaapi_tick: expected planes size to be 0<planes<=8 for drm buf, got %d planes\n", texture_num_planes);
-        return false;
-    }
-
-    int texture_fds[8];
-    int32_t texture_strides[8];
-    int32_t texture_offsets[8];
-
-    while(egl->eglGetError() != EGL_SUCCESS){}
-    if(!egl->eglExportDMABUFImageMESA(egl->egl_display, image, texture_fds, texture_strides, texture_offsets)) {
-        fprintf(stderr, "gsr error: gsr_capture_xcomposite_vaapi_tick: eglExportDMABUFImageMESA failed, error: %d\n", egl->eglGetError());
-        return false;
-    }
-
-    int fds[8];
-    uint32_t offsets[8];
-    uint32_t pitches[8];
-    uint64_t modifiers[8];
-    for(int i = 0; i < texture_num_planes; ++i) {
-        fds[i] = texture_fds[i];
-        offsets[i] = texture_offsets[i];
-        pitches[i] = texture_strides[i];
-        modifiers[i] = texture_modifiers;
-
-        if(fds[i] == -1)
-            texture_num_planes = i;
-    }
-    const bool success = texture_num_planes > 0 && vaapi_copy_drm_planes_to_video_surface(video_codec_context, video_frame, source_pos, source_size, dest_pos, dest_size, texture_fourcc, source_size, fds, offsets, pitches, modifiers, texture_num_planes);
-
-    for(int i = 0; i < texture_num_planes; ++i) {
-        if(texture_fds[i] > 0) {
-            close(texture_fds[i]);
-            texture_fds[i] = -1;
-        }
-    }
-
-    return success;
-}
-
 vec2i scale_keep_aspect_ratio(vec2i from, vec2i to) {
     if(from.x == 0 || from.y == 0)
         return (vec2i){0, 0};
@@ -895,3 +669,20 @@ vec2i scale_keep_aspect_ratio(vec2i from, vec2i to) {
 
     return from;
 }
+
+unsigned int gl_create_texture(gsr_egl *egl, int width, int height, int internal_format, unsigned int format, int filter) {
+    unsigned int texture_id = 0;
+    egl->glGenTextures(1, &texture_id);
+    egl->glBindTexture(GL_TEXTURE_2D, texture_id);
+    egl->glTexImage2D(GL_TEXTURE_2D, 0, internal_format, width, height, 0, format, GL_UNSIGNED_BYTE, NULL);
+
+    const float border_color[4] = { 0.0f, 0.0f, 0.0f, 0.0f };
+    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    egl->glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, border_color);
+    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, filter);
+    egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, filter);
+
+    egl->glBindTexture(GL_TEXTURE_2D, 0);
+    return texture_id;
+}

src/window/wayland.c

@@ -1,4 +1,4 @@
-#include "../../include/window/window_wayland.h"
+#include "../../include/window/wayland.h"
 
 #include "../../include/vec2.h"
 #include "../../include/defs.h"

src/window/x11.c

@@ -1,4 +1,4 @@
-#include "../../include/window/window_x11.h"
+#include "../../include/window/x11.h"
 
 #include "../../include/vec2.h"
 #include "../../include/defs.h"

src/window_texture.c

@@ -85,8 +85,10 @@ int window_texture_on_resize(WindowTexture *self) {
         texture_id = self->texture_id;
     }
 
-    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
-    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
+    const float border_color[4] = { 0.0f, 0.0f, 0.0f, 0.0f };
+    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_BORDER);
+    self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_BORDER);
+    self->egl->glTexParameterfv(GL_TEXTURE_2D, GL_TEXTURE_BORDER_COLOR, border_color);
     self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR);
     self->egl->glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR);