AVISynth Read Error: AVISynth: Stack Overflow

[fasttimes]

I ran the AVI through Telecide, then opened the new AVI, and run it through Decimate, to create the AVS script file. Then I added the correct line to the beginning of the AVS script, and open it with VDub. The file loads fine, but if I try moving along the time line in any way, and I receive the error ""AVISynth Read Error: AVISynth: Stack Overflow", and the preview screens show no picture.

The AVI is a HuffYUV RGB compressed AVI of about 72GB, and I'm using AVISynth version 2.05, and VDub v1.4.12.

Also tried with the newest versions of AVISynth and VDub.

Any ideas?

[Guest]

Try just serving the AVI without the filters:

avisource("your_file.avi")

Does it fail that way? If not, then add the filters on-by-one and try. Tell us what you find.

[fasttimes]

Merry Christmas, Donald! Hope you and yours, have a good one.

No, AVISynth works correctly, with just "avisource("f:\telecided.avi")".

I also checked further... It will also work if I cut down the AVS file size. The current AVS created by Decimate is about 700KB, for a 1:33:00 movie. I tried chopping the AVS in half, but it still had the same problem. Then tried 1/4 the size (~175KB), and AVISynth would then work without the stack overflow!

Looks like some sort of bug in AVISynth? Lucky me, for finding it! :(

[sh0dan]

How can an AVS-file be 770K?

I really don't get this ;)

[fasttimes]

The script was produced by the Decimate filter in VDub, on a 1:33:18 movie that consists of just under 168,000 frames. The AVS is over 33,000 lines long. :eek:

This is the first time I've had a problem with AVISynth, but it's also the first time I have tried to IVTC a full length movie with Telecide/Decimate combo. Up to now, it's mainly been TV shows, ect. I have a GB of RAM, so I don't think it's a memory issue. :p

If someone can tell me the "pure" AVISynth way to do Decimate this AVI, and I'll be happily try it! I got this 70+GB file just sitting there, and I can't go any further with it, until I can get decimate to do its thing. :mad:

Happy Holidays, all!:cool:

[sh0dan]

I think neuron has just the right answer waiting for you :)

[fasttimes]

Originally posted by sh0dan
I think neuron has just the right answer waiting for you :)
Hmm... I don't like that sound of this reply... hope you mean good news!:scared: Hehe.

[Guest]

Do not use the VirtualDub versions of Telecide and Decimate with Avisynth! Use only the functions in the Decomb.dll. Make an AVS file like this:

avisource("your_file.avi")
Telecide()
Decimate()

[fasttimes]

I'd like to try the all-AVISynth route, anyway. However, your reply is confusing. Sorry for the misunderstanding. I'll give it another try:

1. I used VDub's Telecide in VirtualDub, not AVISynth!
2. I used VDub's Decimate in VirtualDub, [b]not[b] AVISynth, to generate the AVS script.
2. Then, I added the appropriate AVISynth line to the AVS file, and loaded it up in VDub.

As far as I know, I followed the directions for using VDub's Telecide/Decimate Combo. Did I not? I think you thought I an trying to load the VDub plugins via the AVS file, which is not the case.

Like I said, I want to try the all-AVISynth route, anyway, as it's a much cleaner solution. But, I'm still curious why I get the AVISynth stack overflow, using the VDub procedures.:)

[sh0dan]

7-800Kb of script is probably beyond the limits of AviSynth. Just use Donald excellent decomb package for AviSynth - it'll rock your world :)

[fasttimes]

Message deleted...

actually, decomb comes with very good docs. I just wish AVISynth's docs were a little better oginized! Thanks both of you, for your help!

[Guest]

Originally posted by fasttimes
I'd like to try the all-AVISynth route, anyway. However, your reply is confusing. Sorry, I see now what you were doing. sh0dan is quite correct about the script size!

The VirtualDub Telecide and Decimate were really just proofs of concept for a blind field matching with separate decimation approach I wanted to try for IVTC. I always realized it was cumbersome and awaited for a long time the capability for VirtualDub filters to add or subtract frames. That never came and so I finally decided to implement it all in Avisynth. That led to Decomb.

There's no good reason not to use the pure Avisynth approach with Decomb. You'd also then be able to try out other great tools for dealing with combing, like IVTC22, GreedyHMA, and TomsMoComp. And the great world of Avisynth filters gapes widely before you. :)