YV12 Info

Hi there, Any information coming my way on how to get GordianKnot to work seamlessly with the new YV12 tools, please know you'll be appreciated. I didn't have success in my initial attempt. How should I go about setting it up? I load my AVS script in VDub directly instead of letting GKnot handle it. Thank you, John

I'm not sure what you're asking for. Have you edited the avs-file in a proper way (as I understand it it's just to use the new mpeg2dec3.dll instead of the normal one)? Or is that the problem?

You might want to make sure that the VirtualDub called by GKnot is the Modified version and not the original, possibly by renaming the modified to have the original name.

Basically the thing to take from this is that currently, there is no way to perform a compressibility test. :(

SelectEvery(280,1,2,3,4,5,6,7,8,9,10,11,12,13,14) AssumeFPS(24.976) Put that at the end of your avs script, it should give about the same result as SelectRangeEvery(280,14) plus or minus a frame or two. Naturally you'd have to adjust your target fps as needed and manually load that in virtualdub. At least that's what I use and there may well be a more simple way to do it. AssumeFPS may not even be needed but I use it to check the effect of different filters.

@Fenris: Those filters are not YV12 compatible yet. At least not as far as I know. So JohnMK is correct, no pure YV12 compressibility test is possible yet.

@Acaila Ok, I saw them listed as "Untested, but should work" and didn't recall if anyone had had a problem with them.

Until mpeg2dec3.dll or avisynth 2.5 are re-written to incorporate a function that GordianKnot relies upon for its compressibility testing (SelectRangeEvery), YV12 will remain rather out of our reach. I suggest anyone wanting their GordianKnot encodes to go approximately 25%-35% faster contribute to the threads discussing these issues in the Avisynth Forum. Let them know that we're at their mercy . . . I think they like that. :D But don't ever, ever demand anything. It's certainly not our place . . .

@JohnMK: Try with this: