NTSC film (23.976 progrressive soft-telecine) DVD playback FPS and remux problem

[fireattack]

Hi everyone,

Recently I was encoding an NTSC DVD that has actual content of 100% FILM (23.976p, with pulldown tag for soft telecine). After playing with it for a while, I have two questions which I can't find answer anywhere.

1. why is this film DVD played back in 24 fps instead of 23.976 fps on computer?

When playing back this DVD directly on a computer, it is expected to play as-is instead of trying to soft-telecine it, since it's unnecessary. The actual FPS is indeed ~24 when playing. However, with *all* the players I tried, including VLC, MPC-hc, MPC-be, MPV, etc., I noticed that it is played back in 24fps, instead of 23.976fps, despite that's what it says in the meta tag (24000/1001).

I know it not because I can tell such small difference, but because the duration of the video is shown ~1:25:21. While if I encode it, or even just remux it with MakeMKV, it will be ~1:25:26. The difference of these two duration is exactly 1000/1001. You can also see this difference if you only play just one .vob file (vs. encoded version) as well, albeit the difference will be very small (<1s) for shorter videos.

On a side note, the audio file extracted by DGIndex is also ~1:25:26 (consistent with 23.976 duration). There is no audio desync either way when playing back, though.

2. why does remuxed video have inconsistent frame intervals?

I also remuxed the DVD using both MakeMKV and mkvtoolnix to a single MKV file. I then extracted its timecodes_v2 using mkvextract. The average FPS, is indeed 23.976. However I found another problem, the timecode interval is like this:

timecode Interval
0 33
33 50
83 34
117 50
167 33
200 50
250 34
284 50
334 33
367 50
417 33
450 51
501 33
534 50
584 33
617 50
667 34
701 50
751 33
784 50
834 34

...

As you can see, while the average interval is indeed ~41.71ms (1000/23.976), they're not evenly distributed, but have a 33/50/34/50 pattern. This doesn't happen in the encoded video (which would just be consistent 40/41).

So my question is:

1. why is it like this?
2. will it affect the playback? I can't tell from my eyes, but I assume having such weird interval will cause some abnormality in theory?


Thanks for your time.

[Logan9778]

.. Nevermind, I was thinking of the wrong thing.

[StainlessS]

Provide MediaInfo fot VIDEO_TS.IFO, and eg VTS_01_0.IFO, and VTS_01_1.VOB,
pretty sure that I have seen framerate differ in at least one DVD.

[fireattack]

Originally Posted by StainlessS

Provide MediaInfo fot VIDEO_TS.IFO, and eg VTS_01_0.IFO, and VTS_01_1.VOB,
pretty sure that I have seen framerate differ in at least one DVD.

Here it is: https://gist.github.com/fireattack/1...7f4ca4bcf3859e

[fireattack]

Just found that some other people noticed the FPS playback difference of DVD vs. encoded video, but in his case it's 60/59.94. So it looks like this 1001/1000 playback FPS difference is not limited to FILM DVD.

https://forum.doom9.org/showthread.php?t=170405

[StainlessS]

Complete name                            : D:\eveything\VIDEO_TS\VIDEO_TS.IFO
Frame rate                               : 29.970 (29970/1000) FPS

Complete name                            : D:\eveything\VIDEO_TS\VTS_01_0.IFO
Frame rate                               : 29.970 (29970/1000) FPS

CompleteName                             : D:\eveything\VIDEO_TS\VTS_01_1.VOB
FrameRate/String                         : 23.976 (24000/1001) FPS

Does above look right to you, (compared to other similar NTSC DVD, I'm in PAL land and dont have any NTSC DVD's handy at the moment).

Note, in your linked thread, FeB seems to have gotten his numbers just a bit wrong (shifted decimal point in wrong direction).

0.001, which is equal to
0.00001%

Should be 0.1%.

In other thread, FeB, asks if Sony is doing it wrong playing at 60.0FPS when everybody else does it 0.1% slower, perhaps Sony feels that
removing the artificial 0.1% slowdown is simply a correction back to recorded real time play rate.

[Logan9778]

Well, considering everyone has computers and flat screen TV's, does it really matter to slow it down .1% anymore? Probably going back to 30 and 60 FPS is the future of video.

For now, I would just be worried that the FPS matched the sound track length, and didn't go out of sync.

[manono]

The difference is due to the drop frame flag. DVD players use 30fps and something like DGIndex uses 29.97fps. The fps difference accounts for the player showing the film to be about 5 seconds shorter in length. You can look it up but the Wikipedia says this:

This meant that an "hour of timecode" at a nominal frame rate of 30 frame/s, when played back at 29.97 frame/s was longer than an hour of wall-clock time by 3.6 seconds, leading to an error of almost a minute and a half over a day.

To correct this, drop frame SMPTE timecode was invented. In spite of what the name implies, no video frames are dropped (skipped) using drop-frame timecode. Rather, some of the timecodes are dropped. In order to make an hour of timecode match an hour on the clock, drop-frame timecode skips frame numbers 0 and 1 of the first second of every minute, except when the number of minutes is divisible by ten (i.e. when minutes mod 10 equals zero). (Because editors making cuts must be aware of the difference in color subcarrier phase between even and odd frames, it is helpful to skip pairs of frame numbers.) This achieves an "easy-to-track" drop frame rate of 18 frames each ten minutes (18,000 frames @ 30 frame/s) and almost perfectly compensates for the difference in rate, leaving a residual timing error of only 1.0 ppm, roughly 2.6 frames (86.4 milliseconds) per day.

https://en.wikipedia.org/wiki/SMPTE_...frame_timecode

[fireattack]

Originally Posted by manono

The difference is due to the drop frame flag. DVD players use 30fps and something like DGIndex uses 29.97fps. The fps difference accounts for the player showing the film to be about 5 seconds shorter in length. You can look it up but the Wikipedia says this:



https://en.wikipedia.org/wiki/SMPTE_...frame_timecode

Thanks for the information. If I read it correctly, 29.97fps should be the correct playback speed, right?

[Stereodude]

The players play it back at the correct speed (24/1.001). The software players like MPC-HC just don't show the true time. They have a 1.001x scaling factor in the displayed time.

This goofy 1.001x scaling for DVDs makes creating chapter .xml files for mkv's and chapter files for x264/x265 encodes from DVDs a pain since many of the tools don't account for that. You have to set ChapterGrabber to 24.000fps to create the x264/x265 with the correct frame numbers for the chapter even though the content is 24/1.001. The .xml file will have the wrong times for the chapters and has to manually be edited to correct it.

[fireattack]

Yeah, I definitely noticed that in ChapterGrabber. Megui's OneClick Encoder has similar problem as well.

Thankfully in my testing, at least HandBrake and MakeMKV can automatically convert the chapter file to correct timing.