AMVIVTC - new function for easy ivtc

[Zarxrax]

I just wrote a new avisynth function tonight, which is intended to condense THIS huge page of info into an extremely easy to use function.

AMVIVTC can be downloaded here: http://amvhell.com/stuff/AMVIVTC.avsi

It requires the following plugins: TIVTC, Tdeint, TMM, NNEDI
Those plugins can all be found on the following two pages:
http://forum.doom9.org/showthread.php?t=82264
http://forum.doom9.org/showthread.php?t=129953


AMVIVTC is designed to be super amazingly simple. As such, it only has 2 arguments.
The first argument is mode, which has 4 settings, ranging from 0-3.
mode=1 is the setting for when you plan to encode your script to a lossless file for editing.
mode=2 is for when you want to edit your avs scripts directly.
mode=3 is for when you want to do bait and switch editing with proxy files.
mode=0 is a fast mode that does not do any postprocessing.

The other argument is decimate. It can be either true or false. When set to true, it will decimate the footage, and when set to false it wont. It will select the proper decimation mode automatically, based on the mode that you choose.

The default settings are mode=1 and decimate=true.

My plan is to vastly simplify the ivtc section of the guide by referring to this function rather than the complex scripts that are in there now. Please let me know your thoughts and comments on this.

[The Origonal Head Hunter]

I think I may love you.

[LantisEscudo]

This looks to be a good simplifying function, but it uses TFM, which gives me some really ugly combing artifacts on some cuts or high-motion scenes.

These caps were done with AMVIVTC(mode=1, decimate=true), TDeint 1.1, TIVTC 1.0 (TIVTC 1.0.5 has the same behavior), TMM 1.0, and NNEDI 1.3:





Is it just my system, or can there be a way to use the Telecide/Decimate combo from Decomb as another option, for sources that exhibit this ugliness?

[Pwolf]

Whats the difference between modes 1 and 2?

[Zarxrax]

Lantis: if you try using TFM on its own, have you tried messing with various modes and things to see if anything helps it?
Pwolf: mode 2 is much faster because it uses the default deinterlacing in tfm, rather than tdeint+nnedi. It also uses a different decimation mode which is frame accurate. Mode 3 is identical to mode one, but uses the frame accurate decimation mode.

I can add more settings if it's necessary for some sources, but at some point down that road the purpose of the function would be lost.

[The Origonal Head Hunter]

this worked beautifully for my Riding Bean footage. Many thanks Zarx.

[Krisqo]

I'll give this a try when I begin work on my new project. But I have a question. When using the Fake AVIs, should I use mode 2 since I am KINDA working with AVS files or go with 1 or 0?

[The Origonal Head Hunter]

It says use 3 for bait-and-switch.

[Krisqo]

It says use 3 for bait-and-switch.

Isn't bait and switch a whole other concept? I'm not making low quality clips and then subbing in another clip. I'm updating the AVS files as explained in the guide under "Editing AVS files Directly" but running them through VFAPIConv since it loads the scripts into Premiere more stably and they work in After Effects.

[Zarxrax]

Krisqo: Modes 2 or 3 both support random frame access, and Mode 2 is faster but lower quality than mode 3, those are basically the differences.

I just updated the function, and it behaves somewhat differently now. Mode is the only setting that you really need, but I have added some other settings for people who like the ability to tweak things a little.
Changes:
- mode 0 is removed
- decimate parameter is changed to allow a value from 0-2. 0=disabled, 1=best method but doesn't support random access, 2=supports random access
- tfmmode parameter basically lets you set TFM's "mode" parameter. Don't mess with it unless you know why you are messing with it.
- pp lets you specify a postprocessing mode. 0=disabled, 1=high quality and slow, 2=an alternate high quality slow method, 3=fast and crappy

I have also added a new function called AMVDeint(), which basically lets you try out a variety of deinterlacing methods (NOT IVTC!) by simply changing a single parameter.
It doesn't let you tweak settings at all, but it does give you a way of easily testing a variety of deinterlacing methods without having to remember a bunch of different commands.
AMVDeint(mode=0) is equal to LeakKernelDeint(threshold=10)
AMVDeint(mode=1) is equal to SangNom(aa=32)
AMVDeint(mode=2) is equal to Tdeint()
AMVDeint(mode=3) is equal to TDeint(mode=2, mtnmode=3, blim=100)
AMVDeint(mode=4) is equal to TDeint(edeint=c.nnedi(),emask=c.TMM())