(Dust) Would this be faster?

[HyperYagami]

I'm trying to do PixieDust on a TV capture (hyffyyuv AVI) with 5-pass CCE (/w DoCCE4U + Robshot) here, here is the script:

--------------------------------------------------------------
SetMemoryMax(256)
SegmentedAVISource("capture_2.avi")
Telecide(order=1)
Decimate(cycle=5)
Trim(1878,7176)+Trim(12219,18094)+Trim(22174,36466)+Trim(40788,44314)
LanczosResize(720,480)
LetterBox(6, 0)
PixieDust(5)
--------------------------------------------------------------

Not trying to do anything heavy-duty here. Now if I feed this AVS into VirtualDub and save a non-compressed (or hyffyyuv'd) AVI out if it (HD space isn't a concern for me), THEN feed that AVI (*not* the original AVS) into CCE...would that be faster than feeding the original AVS into CCE? Since I figured with multiple passes, you're essentially PixieDust'ing the thing multiple times and waste time.

Also, as a seperate problem, the reason why I wanted to do that is for some reason CCE (2.5) becomes very very very slow when I feed that AVS into it, the 1st pass was OK, then at the 2nd pass it goes to a virtual halt at about 34%. Not a complete halt though, just very very very slow, like more than a whole minute just for 12 frames (about 10s for 12 frames before). I'm not sure if that's a CCE or an AVISynth (2.5.4 here, with LoadPluginEx) problem. Anyone have seen problem?

Thanks for any suggestion!

[stickboy]

Originally posted by HyperYagami
Now if I feed this AVS into VirtualDub and save a non-compressed (or hyffyyuv'd) AVI out if it (HD space isn't a concern for me), THEN feed that AVI (*not* the original AVS) into CCE...would that be faster than feeding the original AVS into CCE?Yes. Telecide+Decimate+LanczosResize+PixieDust certainly should be much slower than the overhead from Huffyuv compression+decompression.

Make sure you use Fast Recompress mode in VirtualDub/VirtualDubMod, too.

[HyperYagami]

Thx for the quick respond!

What's the benefit of using "Fast Recompress" rather than, say "Full Processing"? Is that because both the original and output are in YUV2 (while the other ones involve *->RGB)?

Also, if I use "Direct Stream Copy" instead, I assume the output is gonna be uncompressed RGB, not something that I want?

[stickboy]

Originally posted by HyperYagami
What's the benefit of using "Fast Recompress" rather than, say "Full Processing"? Is that because both the original and output are in YUV2 (while the other ones involve *->RGB)Right.
Also, if I use "Direct Stream Copy" instead, I assume the output is gonna be uncompressed RGB, not something that I want?In your case, you'd end up with uncompressed YUY2 video.

[HyperYagami]

thank you so much for all your help.