ffdshow raw video filter doesn't accept 10-bit input?

[heyer]

Hey there, I would like to use some of ffdshow's post-proccesing settings, while still using LAV filters for decoding. The weird thing is: LAV filters seem to dither the content down to 8-bit before sending to to ffdshow. However, when I use ffdshow alone, it seems to handle the 10-bit just fine and outputs it as 10-bit to MadVR. Does ffdshow not support 10 bit if it isn't the decoder? Or am I doing something wrong here?

[Vyral]

No reply ?
I also try to use the last version of ffdshow with LAV for post-proccesing and I noticed the same problem.

[paradoxical]

Did you uncheck everything but 10-bit colorspaces to see what happened?

[Vyral]

Did you uncheck everything but 10-bit colorspaces to see what happened?
If I force 10 bits output in LAV, it will bypass ffdshow. It seems that ffdshow raw filter only accept 8 bits input.
This "bug" has been reported in ffdshow bug tracker but never resolved.

[paradoxical]

Then that seems to be a pretty conclusive answer to the OP.

[clsid]

It does not accept 10-bit input because none of its processing filters supports working with such colorspaces. That isn't a bug, just a limitation. Even the ffdshow decoder will convert 10-bit to 8-bit internally when doing processing. So if you want to use a fully 10-bit playback chain, you should not use ffdshow.

Don't bother suggesting that 10-bit support should be added. That isn't going to happen. There are no developers interested in working on ffdshow. It is a dead project.