jyapayne/moonlight-qt
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Increase the number of decoder surfaces for DXVA2/D3D11VA

Browse source 
We forgot to take into account frames buffered within Moonlight itself
This commit is contained in:
Cameron Gutman 2022-10-14 21:20:35 -05:00
commit 02c781b8f4
3 changed files with 4 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

3
app/streaming/video/ffmpeg-renderers/d3d11va.cpp
View file

@ -458,7 +458,8 @@ bool D3D11VARenderer::initialize(PDECODER_PARAMETERS params)
framesContext->width = FFALIGN(params->width, m_TextureAlignment);
framesContext->height = FFALIGN(params->height, m_TextureAlignment);
// We can have up to 16 reference frames plus a working surface
// We can have up to 16 reference frames plus a working surface plus the maximum
// number of frames in our various pacing and rendering queues.
framesContext->initial_pool_size = DECODER_BUFFER_POOL_SIZE;
AVD3D11VAFramesContext* d3d11vaFramesContext = (AVD3D11VAFramesContext*)framesContext->hwctx;