VFR for Fansub Encoders - how, why, WTF?

[Nicholi]

60fps (59.940) is also becoming used quite a bit more in Japanese raws, from the ending text (usually scrolling) being bobbed from 30fps. So it's not just 24/30 so much now in this 120fps junk.

[Quarkboy]

Originally Posted by Mentar

With all due respect, but unless this filter dramatically improved its performance in the last 6 months, it is (excuse me) useless crap. Let's face it, people, there is _no_ truly reliable way to properly determine VFR other than _using your eyes_ and making the proper call afterwards. I would strongly DISCOURAGE inexperienced encoders from using those semi-automatic "fire and forget" filters which pretend to mysteriously detect VFR parts, and afterwards even decimate to a VFRAC video with timecode file.

Well, you're quite right, I've never actually used tfm for that purpose other than experimentation. However I've encountered extremely hybrid shows where it's not simply the OP or ED, but 50% of the zooms, pans, and digital compositing effects that are in 30 fps, making manual tweaking intensively time-consuming. Plus, with very careful settings and the use of an override file using tdecimate can, in theory, save a lot of time from doing the entire thing manually. (I'm not expert in yatta, but that might also be just as fast).
And anyway, doing vfr AT ALL is not something I'd recommend for an inexperienced encoder to begin with, and the original question seemed to ask for whether there were any automated or semi-automated solutions, so I gave him one (albeit, as you say, it's far from perfect).

Actually one of the better solutions is to create your own 120 fps avi with skip frames using whatever IVTC method you prefer, then dedup, and THEN use tdecimate on the special 120->vfr timecode mode, which produces much better results than using the other direct modes.

[Yumi`]

Won't tdecimate mode=6 after a dedup pass go haywire? It does expect an (undecimated) 120fps stream afaik. At the very least the mkvOut timecodes should be way off, tdec doesn't have an input option for the timecodes file that would be produced by dedup.
I don't see why you would need dedup anyway, tdec can handle this on its own, mode 6 is made exactly for this purpose.

[Quarkboy]

Originally Posted by Yumi`

Won't tdecimate mode=6 after a dedup pass go haywire? It does expect an (undecimated) 120fps stream afaik. At the very least the mkvOut timecodes should be way off, tdec doesn't have an input option for the timecodes file that would be produced by dedup.
I don't see why you would need dedup anyway, tdec can handle this on its own, mode 6 is made exactly for this purpose.

Sorry, I misspoke, I meant "dup"... the original one that simply replaces frames that are just a TAD bit different with bit for bit identical ones. not the more powerful dedup.

[TheFluff]

Well, regarding DVD's: if the studios actually encoded progressive material as progressive and set the RFF flags properly things would be a lot easier (it could be done automatically by reading the .d2v or the VOB/MPEG2 bitstream directly). As it stands, with progressive content usually coded as interlaced with "hard" telecining, the only remotely reliable way to tell what's what is manually with eyes, keyboard and YATTA (which does have a few semi-automatic ways of marking 30fps sections).

I intentionally didn't cover DVD VFR in the original post, partly because it's hard and really annoying, partly because I think it's pretty rare for fansub encoders to deal with, partly because I didn't want to get trolled by the ethics crew, but mostly because I didn't want to write a YATTA manual ().

[D404]

Originally Posted by TheFluff

... but mostly because I didn't want to write a YATTA manual ().

Someone (or some group of people) really SHOULD write a complete manual someday soon though :/

Besides maybe lack of knowledge of theory, simply not knowing how to use YATTA itself is what is keeping various encoders from not using it. As it stands now, pretty much teh only way to learn YATTA is to have someone who knows explain it a bit (as the way i originally learned), which is kinda a pita.

[martino]

Originally Posted by D404

As it stands now, pretty much teh only way to learn YATTA is to have someone who knows explain it a bit (as the way i originally learned), which is kinda a pita.

I disagree (partially) on that point. I've learned how to use it on my own (the basics anyway) + a couple of darkhold logs and the little documentation that there is and so far it got me around pretty well -- I think. So it's not really the only way in my opinion, though albeit probably the best and most efficient. So in the end, the point that you are trying to make is still pretty valid, and I do agree with what you are trying to convey since as it stands, it is probably the most common cause.

/me thinks the phrasing of that is weird but doesn't know how to put it better

[Unearthly]

Besides maybe lack of knowledge of theory, simply not knowing how to use YATTA itself is what is keeping various encoders from not using it. As it stands now, pretty much teh only way to learn YATTA is to have someone who knows explain it a bit (as the way i originally learned), which is kinda a pita.

I will say that I don't use YATTA for this reason and would totally give it another shot if a manual was made.

[TheFluff]

Originally Posted by Unearthly

I will say that I don't use YATTA for this reason and would totally give it another shot if a manual was made.

I guess I'll have to do it sooner or later then. Myrsloik obviously never will, Nicholi and Mentar are too busy with other stuff (and Mentar thinks tutoring is the Right Thing to do anyway), Kintaro doesn't seem very active and I can't think of anyone else who actually understands YATTA. I don't understand it either, at least not all parts of it, but eh...

[Nicholi]

Pshaw...YATTA Guide?

Part 1:
If you don't know what IVTC is, you can't use YATTA. Obvious point is obvious. How do you plan to use a tool that makes manual IVTC easier...if you have no idea what IVTC is?

Part 2:
Combine theory of IVTC with buttons that have obvious names. Done.

That is pretty much it. Use right click mouse button and look at all the features. The ancient help file nmap made will tell you what all the obvious buttons are if you are that dumb.

YATTA isn't complicated....rather IVTC'ing can be though :P. The best way to understand IVTC imo is read about how the Telecide/TFM filters work, since that is the rather rigorous part of IVTC and YATTA has based most of how it works off those filters (more Telecide than TFM...but Myrs is pushing that over slowly). If you know how those work, you know how to use YATTA. If you are completely in the dark as to what those filters are doing, well even if someone told you how to use YATTA you still would have no idea what you are doing.

[Mentar]

Originally Posted by TheFluff

I guess I'll have to do it sooner or later then. Myrsloik obviously never will, Nicholi and Mentar are too busy with other stuff (and Mentar thinks tutoring is the Right Thing to do anyway), Kintaro doesn't seem very active and I can't think of anyone else who actually understands YATTA. I don't understand it either, at least not all parts of it, but eh...

Well... my point is that YATTA is the result of metastasizing cancer. It initially started as a small tool to help create decomb override files, and more and more features were added wildly, until the final end result is a prime example on how NOT to make it usable. Therefore, in order to use it efficiently, you NEED someone who not only gives you a feature manual, but rather who tells you what to do how, but most importantly also WHY.

First, you need to have a detailed understanding of the IVTC process, what it does, and why it is done. As a followup, you need to understand what a ccnnc pattern is, and why it's important to work to achieve it whereever possible. And THEN you need to learn which YATTA tricks and shortcuts are the best way to achieve it. And I doubt that all of this can really be done by manual. The best and easiest way is to use a tutor to introduce someone.

Once you're a veteran and know the tool, it's extremely powerful. I can prepare a match-perfect Claymore episode including filtering, VFR and all the bells and whistles in around 5-10 minutes. But if you only learn and use the obvious ways, you can waste _hours_ to come up with something worse. This is my experience from introducing people who have already worked with yatta, thought to have mastered it, and then found out how much easier it could have been with a couple of tricks. As a self-check, the most important keystrokes in Yatta are Ctr-S and Ctr-Alt-G. If you don't know what exactly they do, you have barely scratched the outer layer of the tool.

And it's not like it's impossible to find a yatta veteran tutor. Off the bat, I'd also name Alizar (who recently ran some training for a Static-Subs bootcamp) among several others. If you really want to use it, you can also go to #darkhold on Deltaanime and ask for help there.

I mean, a good manual would ALWAYS be very helpful, so I don't wanna discourage Fluffy by any means. But yea - that's why I believe a tutor is better. You should have IVTC down cold before you touch YATTA.

[D404]

Mentar makes a very good point (rather, two) about both the buttons/options being fairly obvious and needing a "tutor" to some extent.

I still do think, however, it could benefit from a manual/documentation so peopel who are new to it don't have to just assume what some of the stuff in it for, or how it works.

[Kristen]

I'm running into a bit of a problem.

I have a x264 in an avi container. When I play it, it definately is showing a changing frame rate. When I put it into AVSP, it displays 120 fps. So, I downloaded cfr2tc. I used the GUI, since command line prompts and me don't get along.

So, in cfr2tc, I put the raw into the "Input AVI File", and then I said I want an output AVI file and timecodes. I asked for v1 timecodes and an AVI file on the bottom. Then I hit "Run"
The processing screen popped up and siad "Proccess running..." "Processing incomplete" "Exit Code (10): A problem was encountered"
The output screen said:
cfr2tc v 1.4 by tritial
Frame Parsing Progress 100.00% (139265)
Number of data chunks does not match frame count.

How would I go about fixing this?

[Dark Shikari]

Originally Posted by Kristen

I'm running into a bit of a problem.

I have a x264 in an avi container. When I play it, it definately is showing a changing frame rate. When I put it into AVSP, it displays 120 fps.

Is this where we take the person who encoded this file, nail him to the wall, and start distributing darts?

[edogawaconan]

Originally Posted by Dark Shikari

Is this where we take the person who encoded this file, nail him to the wall, and start distributing darts?

count me in

seriously, x264 in avi is a headache in itself. Then add vfr in it.

[Kristen]

Originally Posted by edogawaconan

count me in

seriously, x264 in avi is a headache in itself. Then add vfr in it.

It's a complete mess. But, I think I managed to get something.

I did the following script:

Loadplugin("C:\Documents and Settings\Administrator\Desktop\Filters\FFMpegSourc e.dll")
FFMPEGSource("C:\Documents and Settings\Administrator\Application Data\mIRC\downloads\raw.avi", timecodes="test.timecodes")
ASSUMEFPS(23.976)

And put it into VDubMod. It output a timecodes file and a lossless.avi (Unfiltered, but this is just for testing purposes).

I'm assuming with h.264 I encode it as normal, and then use MKVmerge to mux in the timecodes. But, for XviD, how do I apply the timecodes?

(By the way, I confused a few people. The RAW is x264 .avi VFR. I'm not trying to encode to it, but from it)

[edogawaconan]

I'd just screw the vfr, feed with directshowsource, changefps to 29.blah or 23.blah then output to lossless.

[Kristen]

Originally Posted by edogawaconan

I'd just screw the vfr, feed with directshowsource, changefps to 29.blah or 23.blah then output to lossless.

I definately would like to do that. We have for the past 5 months. XD. But with this show, a lot of people have started asking us to fix up the panoramic screens and stop them from being jumpy.

Is there no way to put timecodes into an XviD other than convert or change fps?

[edogawaconan]

you must be doing it wrong.
or the capper.

...

or both.

[TheFluff]

Originally Posted by Kristen

Is there no way to put timecodes into an XviD other than convert or change fps?

You mean into an AVI. Xvid is a codec, AVI is a container (and it doesn't support variable framerate). If you want VFR Xvid, mux it to mkv.