[FFmpeg-devel] [PATCH]Mention in the documentation that fieldmatch needs cfr input

tim nicholson nichot20 at yahoo.com
Fri Oct 24 11:05:29 CEST 2014

On 23/10/14 17:06, Dave Rice wrote:
> On Oct 23, 2014, at 4:05 AM, Clément Bœsch <u at pkh.me> wrote:
>> On Tue, Oct 21, 2014 at 09:32:39AM +0200, Carl Eugen Hoyos wrote:
>>> Hi!
>>> It appears to me that we all know that fieldmatch needs cfr input, 
>>> but it isn't mentioned in the documentation.
>>> Related to ticket #3968.
>>> Please comment, Carl Eugen
>>> diff --git a/doc/filters.texi b/doc/filters.texi
>>> index c70ddf3..bc77623 100644
>>> --- a/doc/filters.texi
>>> +++ b/doc/filters.texi
>>> @@ -4447,6 +4447,9 @@ and VIVTC/VFM (VapourSynth project). The later is a light clone of TFM from
>>> which @code{fieldmatch} is based on. While the semantic and usage are very
>>> close, some behaviour and options names can differ.
>>> +The filter only works for strictly constant frame rate input. If your input
>>> +has variable frame rate, try the @ref{pullup} filter.
>>> +
>> Well... isn't telecined content supposed to be CFR anyway? I would say
>> it's supposed to be undefined behaviour in any ivtc filter but well. No
>> opinion really.
> This would be ideal, but some common production applications use VFR. For instance when digitizing with the Log and Capture interface of Final Cut Pro the first and last frame usually have custom durations although most of the rest of the file is 2997/100. There’s another setting in the Final Cut to compensate for dropped frames by extending the duration of the last good captured frame.

Not only FCP. Pipelines do it too.

> Dave Rice
> _[..]
Key Fingerprint 38CF DB09 3ED0 F607 8B67 6CED 0C0B FC44 8B0B FC83

More information about the ffmpeg-devel mailing list