Is this too much filtering?

[Colyptic]

I'm just getting into using external avisynth 2.5 plugins, used to use only Gordian Knot default scripts. Here is the script:

LoadPlugin("C:\PROGRA~1\GORDIA~1\mpeg2dec3.dll")
LoadPlugin("C:\PROGRA~1\GORDIA~1\decomb.dll")
LoadPlugin("C:\PROGRA~1\GORDIA~1\Convolution3DYV12.dll")
LoadPlugin("C:\PROGRA~1\GORDIA~1\Blockbuster.dll")
mpeg2source("D:\movie.d2v", cpu=6)
Telecide().Decimate(5)
crop(4,2,714,474)
Convolution3d (preset="movieHQ")
LanczosResize(512,272)
Blockbuster(method="sharpen",block_size=3,luma_offset=-2,luma_threshold=30,detail_min=1, detail_max=30)

Using XviD 170203 (Koepi's build) with 3 Bframes 150/100, VHQ 1, Chroma Motion, motion search - ultra high, MPEG quants...The quality looks great, but the encoding took over 10 hours for a 100 min. film on an XP 1800+ takes 6 hours with the same script with the post processing left out. I would like to speed it up a bit without losing much (if any) quality.

I think I might be doing about the same thing with blockbuster and post-processing. Would it be better to use just the post-processing and forget about the convolution3D and blockbuster, vice versa, or does this script look alright to you?
TIA

[Manao]

An easy way to speed up the encode without losing too much in quality : in mpeg2dec, use cpu =4 instead of cpu = 6 ( which I think blurs too much the picture anyway ), and put convolution3d after the resize.

[Evil Andy]

Try turning off Chroma Motion as it's exceptionally slow.

[Colyptic]

OK I put Convolution3D after the resize, which is how I did for a few rips in the past. I just heard that if you put it before the resize you'll see a noticable difference in quality, which I can't see but I'm not very picky. I ran a few tests with a 2:34 clip at quality 100, all other XviD parameters unchanged, and this is what I found out:

Total time with original script: 6:20

with C3D after everything else unchanged; 13 sec. faster, 1.2MB bigger

C3D after, cpu=4; 14 sec. faster, 1.6MB larger

C3D after, no PP; 18 sec. faster 2.7MB larger

C3D before, CPU=6, no BlockBuster; 47 sec. faster and 1.1MB smaller

C3D after, CPU=6, no BlockBuster; 56 sec. faster, .7MB smaller

C3D after, no PP, no Blockbuster; 1m 3 sec. faster, .1MB larger

It's just what I expected from the file sizes but it seems BlockBuster adds alot of time to the encode, where PP isn't slowing it down much at all, surprised me, I've been using it BlockBuster for awhile and I notice its a little slower, 5-10fps but never 5-10fps, which I was getting with my first try at PP. Maybe it was just that certain movie I was encoding, seemed to take alot longer than normal.

Looks like I had a good combo to begin with averages everything out, I might take advantage of the little time saved on moving C3D after. I will mess around with the XviD settings when I find some time...