[FFmpeg-user] How is the gmtime option used for pts?

[Matthew Mah]

On 02/05/2016 04:37 AM, Moritz Barsnick wrote:
> I think the options gmtime and localtime use the time *at the time of
> filtering*, not the creation time. (Yes, that may be the same time if
> you are capturing from a device. ffmpeg can also use the wallclock for
> timestamping the captured frames though, if that's of help for you.)
>
Based on the documentation, I understand the gmtime and localtime for 
drawtext are for the time of filtering, but for pts, gmtime and 
localtime should be referenced to the start of the file.

Could you please elaborate on what you mean by using the wallclock for 
timestamping frames? I am postprocessing an already captured file, so I 
am not sure whether this would apply.
> The reason for your error: Your ffmpeg version does not match with the
> documentation you are using. The feature you are referring to was
> introduced to git master on October 12, 2015:
> http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=cf28490e564d91c89b4c56e605c6f391bddb9ba9
> and hasn't become part of a release as of today. You need to compile
> (or grab a version which was compiled) from git master to have this
> feature. Or wait for 2.9, if you need it in a release. (No need to
> stick to releases, git master is considered stable.)
Thanks, that's exactly what I needed to know. I thought that because the 
commit was on the master branch, and my release was from a later date, 
that this would be included.

> BTW, while I didn't get your error message, the displayed timestamp in
> my experiment didn't seem correct. But that may be a different issue or
> just me.
I think this can be solved with the pts offset argument.