[FFmpeg-devel] [PATCH] telecine filter

Paul B Mahol onemda at gmail.com
Mon Apr 8 20:25:51 CEST 2013

On 4/8/13, Derek Buitenhuis <derek.buitenhuis at gmail.com> wrote:
> On 2013-04-08 1:58 PM, Paul B Mahol wrote:
>>>> + at item pattern
>>>> +String representing for how many fields a frame is to be displayed.
>>>> +The default value is @code{23}.
>>>> + at end table
>>> This is a woefully inadequate explanation. Not only does it not explain
>>> the format the string should be it, it breaks from the standard ordering
>>> convention (which is calling it, e.g. "3:2 Pulldown" ). Examples are not
>>> a substitute for an actual explanation.
>>> Also, it is a grammatically invalid English sentence.
>> I will need hand for this because I'm not that guy to write valid English
>> sentence in this case.
> "A string of numbers representing the pulldown pattern you wish to apply.
>  The default value is @code{23}."
> As for why 23 is used instead of the normal notation of 32, I guess only
> the original plugin authors knows that. Maybe he just wanted to be special.
>>>> +        tc->frame[nout]->pts = AV_NOPTS_VALUE;
>>> This seems broken behavior at best.
>> Yes, that why there is TODO, note that there is no perfect solution
>> possible with current lavfi API.
>> (FYI original code does not have this issue)
>> So I need fedback in that area.
> For CFR content, it's pretty simple: calculate then new framerate and
> generate PTS from it.

By setting outlink time_base and increase pts by one for each output frame?

> For VFR content, the only thing I can think of is to set the time base to
> some multiple of what it used to be (calculated from the pattern), in order
> to use finer grained PTS (since there are new frames now). But yeah, this
> is ugly. On the other hand, if you are trying to apply telecine to VFR
> content, you're likely insane anyway.
> - Derek
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

More information about the ffmpeg-devel mailing list