mirror of
https://github.com/moonlight-stream/moonlight-qt.git
synced 2026-06-18 22:50:57 +00:00
Reduce buffering for MMAL to avoid requiring 128 MB VRAM split
This commit is contained in:
@@ -22,9 +22,11 @@ MmalRenderer::~MmalRenderer()
|
|||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
bool MmalRenderer::prepareDecoderContext(AVCodecContext*, AVDictionary**)
|
bool MmalRenderer::prepareDecoderContext(AVCodecContext*, AVDictionary** options)
|
||||||
{
|
{
|
||||||
/* Nothing to do */
|
// FFmpeg defaults this to 10 which is too large to fit in the default 64 MB VRAM split.
|
||||||
|
// Reducing to 2 seems to work fine for our bitstreams (max of 1 buffered frame needed).
|
||||||
|
av_dict_set_int(options, "extra_buffers", 2, 0);
|
||||||
|
|
||||||
SDL_LogInfo(SDL_LOG_CATEGORY_APPLICATION,
|
SDL_LogInfo(SDL_LOG_CATEGORY_APPLICATION,
|
||||||
"Using MMAL renderer");
|
"Using MMAL renderer");
|
||||||
|
|||||||
Reference in New Issue
Block a user