[FFmpeg-user] Pullup old MPEG-2

Carl Eugen Hoyos cehoyos at ag.or.at
Wed Dec 23 10:35:25 CET 2015

llee782 <llee040 <at> sbcglobal.net> writes:

> It appears that the simplest way I've found to 
> adequately handle this conversion then is to use 
> something like:
> ffmpeg -i 2024_telecine_source.mpg -c:v libx264 
> -c:a copy -r 24000/1001 2024_telecine_source_film.m4v

I don't know if reencoding is really necessary.
(It may be.)

> Does that work mainly because it removes the soft 
> telecine flag, as Rens suggested?

I don't think "removing" is the right word: FFmpeg 
doesn't know about the soft telecine flag and 
therefore cannot write it to the output file.
(Is it really possible to write a telecine flag to a 
m4v file? I thought it is a vob-only thing.)

> Here's another question that has arisen through 
> serendipity. When I first tested the above, I 
> simply failed to notice that the -r option existed
> twice, and the command I ran was actually:
> ffmpeg -r 24 -i 2024_telecine_source.mpg -c:v libx264 
> -c:a  copy -r 24000/1001 2024_telecine_source_film.m4v

(Unrelated: It is probably not a good idea to once use 
24 and once 23.9)

> I discovered that running it with just one "-r" value 
> specified (either "-r 24", or "-r 24000/1001") 
> resulted in many repeated instances of "Past
> duration 0.xxxxxx too large" in the shell output.

I think what you write here is not correct.

But it is true that warnings are shown if you 
overwrite the output frame rate and frame duration is 
not constant.
I don't know if any players will show issues for such 

> The video output appears the same.

It is definitely not the same, depending if you 
overwrite the input frame rate or not.

It appears that FFmpeg thinks that for your use-case 
it is more correct to overwrite the input frame-rate 
than the output frame-rate.

Carl Eugen

More information about the ffmpeg-user mailing list