Hi Guys,
I have been looking at adding separate post processing filters depending on the source media type. The reason I would like to do this is simple. I want to use ffdshow tryouts for Live TV SD content as it has very good deinterlace/sharpening filters (among other things).
Now I would have liked to have used ffdshow for the MPEG 2 codec, but for some reason it will not work, so I set the normal MPV Video codec and use the ffdshow raw filter within post processing filters. Now the problem with this is that I use PowerDVD 8 for H264 content so that I can utilize DXVA, but with post processing filters in place, DXVA is no longer possible.
I was looking at the source for TSReaderPlayer.cs and was thinking of adding conditions in there (or somewhere similar) on adding the post processing filters depending on if the source is MPEG 2 or H264. I am unfamiliar with DirectX programming, but from the looks of things is doesn't look like you can tell what the source video is before the graph is started. Is this assumption correct, and if so what do you think the best way would be to achieve this.
Once I have a general idea, I will get stuck in with the dirty work as supply a patch hopefully.
Thanks
I have been looking at adding separate post processing filters depending on the source media type. The reason I would like to do this is simple. I want to use ffdshow tryouts for Live TV SD content as it has very good deinterlace/sharpening filters (among other things).
Now I would have liked to have used ffdshow for the MPEG 2 codec, but for some reason it will not work, so I set the normal MPV Video codec and use the ffdshow raw filter within post processing filters. Now the problem with this is that I use PowerDVD 8 for H264 content so that I can utilize DXVA, but with post processing filters in place, DXVA is no longer possible.
I was looking at the source for TSReaderPlayer.cs and was thinking of adding conditions in there (or somewhere similar) on adding the post processing filters depending on if the source is MPEG 2 or H264. I am unfamiliar with DirectX programming, but from the looks of things is doesn't look like you can tell what the source video is before the graph is started. Is this assumption correct, and if so what do you think the best way would be to achieve this.
Once I have a general idea, I will get stuck in with the dirty work as supply a patch hopefully.
Thanks