Script to remove telecine/interlacing from DVD rips

[UtahJohnnyMontana]

I am in the process of writing a script that prepares MakeMKV DVD rips for upscaling. It is able to figure out whether a rip is soft or hard telecined, interlaced, or a combination of these things. If more than one method is used, it splits the video into segments and then figures out where the transitions are. It is mostly working for me at this point, so I thought I would give some more people a chance to try it and hopefully spot all the things that I am doing wrong. You can find it here:

https://github.com/utahjohnnymontana/DVD-Rip-Prep

[hello_hello]

UtahJohnnyMontana,
I haven't tried your script yet, but I probably will when I get a chance. After reading through the info on GitHub though, I have a few thoughts...

PAL DVD content shot on film in a PAL country and originally intended to be broadcast there is almost always shot at 25fps. There's quite a lot of it. There's also a very common scenario for PAL DVDs containing progressive video or film where the fields are out of phase, or misaligned, so while it might appear to be interlaced it's actually not. Something like:

A B C D E F
B C D E F G

Using Avisynth Plus, the fields can easily be matched up again using TFM from the IVTC plugin. Sorry if your script can actually detect that sort of thing, but I'll confess I don't understand a lot of it yet.

For a long time when film at 24fps was sped up to 25fps, the audio was sped up to match, but without correcting the pitch.
These days it's not uncommon for the audio to be pitch corrected when it's sped up, so pitch correcting it again when slowing it down is the correct thing to do, but it's hard to know which way to handle it unless you have an NTSC version with audio at the correct pitch to compare it to. I'd never use MKVToolNix to stretch audio to the correct length as it can produce glitches (pops and crackles) on playback.

A thought on video with multiple frame rates... mainly NTSC with sections containing film and sections containing interlaced video etc...
By far, my preferred method for handling that sort of thing is to output variable frame rate video. Keep the film sections at 23.976 and de-interlace the video sections to 29.97 fps etc. Handbrake does that sort of thing fairly well and for Avisynth Plus, once again there's the TIVTC plugin. Unfortunately for Avisynth, there's no plugin that will automatically do combinations of 23.976fps and 59.94fps, although it's usually not an issue as NTSC tends to be predominately one or the other. I'd even prefer to convert to a common frame rate using frame blending when there's only small sections of one type. Changing the frame rate by duplicating frames can often make the video look jittery, which is worse than a bit of blending, at least to me.

Cheers.

Edit: for the record I realise this thread is in the Linux sub-forum, but Avisynth Plus runs perfectly in Wine.
For both Windows and Linux, VapourSynth can generally do anything that can be done with Avisynth. They have many plugins in common, although I prefer the Avisynth scripting language. And there's native Windows and Linux flavours of Handbrake.