Deinterlacing/IVTCing Star Trek The Next Generation Episodes

[Mawazi]

Okay. For the past few weeks I've been attempting to deinterlace and/or inverse telecine some Star Trek : The Next Generation episodes. Some episodes work better than others.

Scenes that do not include any effects at all are simply 3:2 telecined. They are simple.

Some special effects sequences are even 3:2 telecined and are also simple.

On the other hand, many effects sequences are more like the included link. The warp starfield is 3:2 telecined, but the Starship Enterprise is true interlaced. Evil.

I'm guessing there is no way to make those sequences progressive while also retaining all of the possible detail. But, perhaps I'm wrong.

Anyone have a suggestion? Other than to cry and bang my head against the wall? :scared:

http://rapidshare.com/files/75064501/tngsegment.vob.html

[mahsah]

http://avisynth.org/mediawiki/VFR#encoding_to_vfr_.28mkv.29

Try that guide using tfm and tdecimate.

[Mawazi]

http://avisynth.org/mediawiki/VFR#encoding_to_vfr_.28mkv.29

Try that guide using tfm and tdecimate.

Thanks for the quick reply, Mahsah.

I have definitely looked over that guide before and found the idea of vfr encoding fascinating. However, I'm not convinced that these episodes require this. I have yet to find any sequences that appear entirely interlaced, and that would be the purpose of a vfr encode, I believe.

The problem I'm having is the combination of what appears to be telecined film content with interlaced content within the same frame. This must have occured during the editing process with the compositing of an apparently video-taped Enterprise (though that seems hard to believe) with a 'filmed' warp starfield.

I'm guessing that the filmed content was telecined on its own and then later combined with the video starship. But, perhaps something more insidious has taken place that has blurred the origin of these sequences.

[FredThompson]

ST is notorious for this kind of crap.

[FrozenDog]

I just started to backup my Star Trek Voyager collection to get rid of a few annoyances and to simplify the menus and ran into the same problem as you. After reading the article on VFR (http://avisynth.org/mediawiki/VFR), I looked into TIVTC (http://web.missouri.edu/~kes25c/) and took the example from the included tfm_tdeint_postprocessing.avs file.

Here is my script:

LoadPlugin("...DGDecode.dll")
LoadPlugin("...TIVTC.dll")
LoadPlugin("...TDeint.dll")
LoadPlugin("...degrainmedian.dll")
d2vpath = "...tngsegment.d2v" ### Honor Pulldown Flags ###
mpeg2source(d2vpath, idct=2)
deint = tdeint()
tfm(d2v=d2vpath,clip2=deint)
DeGrainMedian(limitY=2,limitUV=3,mode=1)
ConvertToYUY2()


I think it looks great on both the PC and on my 50 inch DLP. But of course, you'll have to decide that for yourself.

Also, if you're using CCE, don't forget to change from the usual 12 frame GOP's to 15.

Good Luck.

[Mawazi]

I just started to backup my Star Trek Voyager collection to get rid of a few annoyances and to simplify the menus and ran into the same problem as you. After reading the article on VFR (http://avisynth.org/mediawiki/VFR), I looked into TIVTC (http://web.missouri.edu/~kes25c/) and took the example from the included tfm_tdeint_postprocessing.avs file.

Here is my script:

LoadPlugin("...DGDecode.dll")
LoadPlugin("...TIVTC.dll")
LoadPlugin("...TDeint.dll")
LoadPlugin("...degrainmedian.dll")
d2vpath = "...tngsegment.d2v" ### Honor Pulldown Flags ###
mpeg2source(d2vpath, idct=2)
deint = tdeint()
tfm(d2v=d2vpath,clip2=deint)
DeGrainMedian(limitY=2,limitUV=3,mode=1)
ConvertToYUY2()


I think it looks great on both the PC and on my 50 inch DLP. But of course, you'll have to decide that for yourself.

Also, if you're using CCE, don't forget to change from the usual 12 frame GOP's to 15.

Good Luck.

Hi, FrozenDog. First, I'd like to say that I believe Voyager to be highly underrated. My wife and I thoroughly enjoyed the show when we recently watched it all the way through on DVD in the space of a few months. Anyhow, just a general thumbs up to Voyager.

I notice that in that script you are deinterlacing and inverse telecining, but not decimating. Seems like that would leave you with some noticeable stutter, especially in panning shots. Though I see you're encoding back into mpeg2, while I'm encoding to x264, so perhaps that makes some sort of difference.

Cheers.

[Zach]

ST is notorious for this kind of crap.

Why?


No, seriously. Over the years here, I have read countless threads from users about various troubles with various commercial DVDs. I've always wondered why.

It seems like a lot of trouble for the makers of the original DVD, so is this convoluted switching back and forth between interlacing schemes, in fact, an attempt to thwart piracy? (That would make sense, I guess.) Or is it just cheaper to do it that way for reasons I can't begin to understand?

[FrozenDog]

Aha, I didn't know you were encoding to x264. I have absolutely no experience with x264. I agree that Voyager is underrated. Outside of the original Shatner series, I think Voyager had the best writing and characters.

I notice that in that script you are deinterlacing and inverse telecining, but not decimating.

Exactly right. In my first attempt, by analysing the d2v file with a text editor, I could precisely dissect the telecined film portions from the interlaced parts and then process them individually. Then feed that to Scenarist as multi-file data (had to do that with Jim Carrey's Number 23). Problem was I found some weird things just like you did.

...the combination of what appears to be telecined film content with interlaced content within the same frame.

So, I gave up on that very tedious project and concluded that you just can't fix Star Trek. It is what it is. Frustrated, all I wanted was a way to reduce the filesize just like DVDShrink and Rejig does but use CCE instead. Tritical's filter gives me a smooth progressive 29.97 encode without combs. I do find panning stutter but only because it was already inherent in the original video. It's just the nature of the beast.