[M4IF Technotes] MPEG 4 Timing Questions

[Mariano Padilla]

Richard,
	Thank you for your comments, however, they do introduce one more
concern in our quest.  As you stated, SMPTE is potentially considering a
"variable frame rate" this would throw off completely the synchronization of
our subtitles to the video.  For example, if a subtitle is to appear at
00:23:25.10, and the variable frame rate is different, say the actual
subtitle should appear at frame 20 because the variable frame difference has
changed the count of the frame to when the actual actor speaks in the movie,
it would appear to the audience that the subtitle is shown too early.
	On the other hand if the actual support for MPEG-4 in 23.976 is to
the frame, we can then compensate for a prolonged loss of synch on 10+
minutes or so.  As far as the Europeans, well then we would just have to
work similarly to NTSC vs. PAL and do offsets accordingly.
	Lastly, just to confirm, if SMPTE agrees to use 23.976, I can
synchronize external pictures or text with the MPEG-4 movie, down to the
frame?
Again, thank you all for your comments.
Mariano J. Padilla
SDI Media USA
Director of IT/R&D
Tel: 323-602-5406
Fax: 323-602-5450
mpadilla   sdi-media-usa.com
www.sdi-media-usa.com
-----Original Message-----
From: richard mizer [mailto:ramizer   wmr.com]
Sent: Wed, October 24, 2023 9:07 AM
To: Gary Sullivan
Cc: Mariano Padilla; technotes   lists.m4if.org; Scott Rose
Subject: Re: [M4IF Technotes] MPEG 4 Timing Questions
Digital Cinema, which is what started this thread, will be 23.976 fps
non-drop time code in its initial implementation, and eventually will
add 47.95, 25 and 50 (since Europe doesn't seem to have the same
electrical interferance problems) that will also be non-drop.  Higher
frame rates will also be considered.  There is even some potential for
variable frame rate, much like the new Panasonic camera supports.
Also regarding 12M SMPTE timecode...the SMPTE S22 committee is working
on a new standard that will replace 12M at some point in the future
commonly called Universal Time Code, based on time since a start point
defined in the 1950s.  This will allow the creation time of any piece of
content to be compared.
Hopefully MPEG4 Studio Profile is tracking with these plans.
> An important note: In Studio profile, the time representation
> corresponds directly to SMPTE 12M.  All of my other remarks
> only apply to other profiles.  In fact I believe you're talking about
> Studio profile, so none of my other remarks apply in that context.
> And I think you should be happy with the time representation
> in the Studio profile, since it is just conventional SMPTE 12M
> and that seems to be what you want.  I guess the Studio Profile
> designers knew what the studio environment users liked.
>
> There are at least two flavors of conventional 29.97 fps timecode in
> common use: Drop and Non-Drop.  My understanding is that both
> are commonly used (some people prefer one and some prefer the
> other).
>
> Drop frame timing has about 0.86 seconds of drift per day
> and non-drop has about 86 seconds of drift per day.  Neither is
> exactly drift free.  And neither carries any assurance that the
> amount of time between HH:MM:SS:FF and HH:MM:SS+N:FF
> is exactly N seconds.  In fact, in both counting methods the
> precise time intervals are difficult to calculate and in drop frame
> timing the computation of the precise duration of time intervals
> becomes more difficult because the counting process is more irregular.
>
Gary Sullivan wrote:
>    Part 1.1.1    Type: Plain Text (text/plain)
>              Encoding: quoted-printable
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/mp4-tech/attachments/20011024/c71643b2/attachment.html