Hi guys,

I have a question about my PotP config, currently it's configured as following:

Splitter: LAV Splitter
Audio Decoder: LAV Audio Decoder
Video Decoder: Built In FFMpeg/LAV Video Decoder
Video Filter: ffdshow raw video filter
Video Renderer: madVR
Audio Renderer: ReClock

I have an issue with CUDA, whenever I try to select CUDA as the video decoder under built in filter settings or within LAV Video Decoder, I can't play a movie, the video will start playing and then freeze.

So, I tried using standard DXVA: If I try using Native DXVA in PP, it doesn't work, and the codec being used is the built in FFMPpeg codec:



If I enable DXVA Copy Back mode, it then works:



This is with the built in FFMpeg decoder.
I tried using LAV Video Decoder and selected Native DXVA, it seems to be working, GPU load is higher, and this is what I got:





However, with LAV as the Video Decoder there is no indication of DXVA in the Video Codec, it just says "LAV Video Decoder". With the built in decoder it says "DXVA".
Also, my input is YUY2 and output is NV12 12 bits, whereas with the built in decoder the input is AVC1 and the output is YUY2 16 bits. Is there an actual difference in quality between these two?

My other question is about ffdshow, I set it as the video filter according to Maken's guide for SVP, but I stopped using SVP due to the very high CPU load, and also because with DXVA I get similar results with less performance hit to the CPU.

I was told that ffdshow is a post processing filter for software decoding, and that if I use DXVA, then ffdshow is not needed.

Is this true? Can I disable ffdshow and just leave LAV Splitter, LAV Audio Decoder and LAV Video Decoder?

Please if you can explain where ffdshow should and need to be used, it would be a great help to me, I'm just trying to find the best settings for playing my movies.

Thanks