Convolution3D--What am I doing wrong?

[wwaag]

I am trying to use the convolution3D filter with Avisynth. Whenever I attempt to load the AVS file into CCE Basic, I get the following message "Frame size 936x56 is not supported. Supported frame size is up to 720x576". (Same problem with MainConcept as well--it also shows a 936 x 56 frame size.) I get the same message even if I simply load the plug-in and nothing else (no colorspace conversions or setting the parameters). Here is a copy of my very simple script.

LoadPlugin("C:\Program Files\AviSynth 2.5\plugins\Convolution3d.dll")
SegmentedAVISource("G:\Tape 1 Archive\tape01.avi")

If I delete the "loadplugin" line, the script works fine.

Any assistance would be greatly appreciated. Thanks in advance.

wwaag

[troyvcd1]

I dont know about segmented avisource but I use avisource with no problem. I need to change my avi fourcc to cdvc with the canopus codec first though.

[FulciLives]

Here is the script I use with a full D1 capture of 712x480 (which might sound odd but that is how my capture card works).

LoadPlugin("C:\PROGRA~1\GORDIA~1\mpeg2dec3.dll")
LoadPlugin("convolution3d.dll")
AviSource("D:\capture.avi")
Trim(717,11507)
crop(6,2,-2,-6)
AddBorders(8,4,8,4)
SeparateFields() 
even = SelectEven(last).Convolution3D (0, 6, 10, 6, 8, 2.8, 0)
odd = SelectOdd(last).Convolution3D (0, 6, 10, 6, 8, 2.8, 0) 
Interleave(even, odd) 
weave() 
AssumeFPS(29.970, true)

The Trim command is imported into my script using VirtualDubMod and it's built-in AviSynth editing script function. I actually load the AVI first and get that value then manually copy and paste it into my script then load the AVI into VirtualDubMod with the script just to make sure all is A-OK.

The crop and AddBorders thing is done to adjust from 712x480 to 720x480 and is more-or-less specific to my video capture card. I crop a bit to get rid of black on the edges and to shift the image down and to the left since my card captures a bit high and to the right.

The AssumeFPS line is there because I use the AUDIO stream to keep sync so I usually end up with something like 29.966 or 29.964 or something like that so the AssumeFPS brings me back up to standard 29.970fps ... nice trick for those that use a capture card that only keeps sync when using AUDIO as the master stream.

In this example I'm using the MOVIELQ setting of Convolution3D

- John "FulciLives" Coleman

P.S.
This script works well with CCE but for audio I run it through TMPGEnc Plus and use SSRC to adjust the sampling rate which is what happens when you adjust the FPS as I did in the script. This then gives me a PCM WAV audio file that I then normalize and convert to AC-3. So CCE for VIDEO and TMPGEnc Plus for AUDIO. Again I'm only using TMPGEnc Plus for audio because of the FPS change which results in a sampling rate change (usually something like 48.5k to 48.0k).

[wwaag]

FucliLives,

Thanks for the reply. I discovered that I was using the wrong version of Convolution3D for Avisynth 2.5. I downloaded the beta version and it seems to work OK.

Since I am using DV and not a capture card, a lot of the script would not be applicable, but thanks for the explanation. One question I do have about your script though. E.g. in "SelectEven(last)", what does "last" refer to? I don't see the clip "last" defined anywhere. Just curious.

Again, thanks for your reply.

wwaag

[FulciLives]

FucliLives,

Thanks for the reply. I discovered that I was using the wrong version of Convolution3D for Avisynth 2.5. I downloaded the beta version and it seems to work OK.

Since I am using DV and not a capture card, a lot of the script would not be applicable, but thanks for the explanation. One question I do have about your script though. E.g. in "SelectEven(last)", what does "last" refer to? I don't see the clip "last" defined anywhere. Just curious.

Again, thanks for your reply.

wwaag

My script is set up for an interlaced source. So if you are using DV you would want to use it ... at least this part of it:

SeparateFields() 
even = SelectEven(last).Convolution3D (0, 6, 10, 6, 8, 2.8, 0) 
odd = SelectOdd(last).Convolution3D (0, 6, 10, 6, 8, 2.8, 0) 
Interleave(even, odd) 
weave()

Here I am using the MovieLQ mode of Convolution3D ... you can easily use this script but put in whatever values you wish for Convolution3D.

I got this bit of code ... if I recall correctly ... from the AviSynth website.

- John "FulciLives" Coleman