[FFmpeg-user] ffmpeg architecture question #2

Carl Eugen Hoyos ceffmpeg at gmail.com
Sat Apr 25 12:37:14 EEST 2020


Am Sa., 25. Apr. 2020 um 04:20 Uhr schrieb Mark Filipak
<markfilipak.windows+ffmpeg at gmail.com>:
>
> On 04/24/2020 01:22 PM, Carl Eugen Hoyos wrote:
> >
> >
> >> Am 24.04.2020 um 11:10 schrieb Mark Filipak <markfilipak.windows+ffmpeg at gmail.com>:
> >>
> >> I've been told that, for soft telecined video the decoder is fully compliant
> >> and therefore outputs 30fps
> >
> > (“fps” is highly ambiguous in this sentence.)

The decoder is "compliant" in the sense that the stream it outputs for
soft-telecined
input has a time base of 30000/1001, the output has 24000/1001 "frames per
second" though.

> > This is not correct.
> > I believe I told you some time ago that this is not how the decoder behaves.
>
> I beg your pardon, Carl Eugen. I thought you said that the decoders are
> fully compliant and therefore produce interlaced fields.

I am quite sure I wrote the opposite several times as replies to your mails.
Note that FFmpeg cannot produce "fields" because it cannot deal with devices
that know what a "field" is.
(Just as there are "interlacing" filters that you would call differently if it
were your decision, we also decided to name some of the filters that
deal with frames "field"-filters because this allows understanding the filters'
purpose for everybody except broadcast and video engineers like you.)

> > I believe such a behaviour would not make sense for FFmpeg (because
> > you cannot connect FFmpeg’s output to an NTSC CRT). The telecine filter
> > would not work at all if above were the case.
> > Or in other words: FFmpeg outputs approximately 24 frames per second for
> > typical soft-telecined program streams.

> > The only thing FFmpeg does to be “compliant” is to forward the correct time base.
>
> By "correct time base" you mean 24/1.001, correct?

lol, no

Carl Eugen


More information about the ffmpeg-user mailing list