[FFmpeg-user] Determining how FFMPEG calculates start time for MPEG-2 files

[Rytis Urnezius]

On Sat, May 27, 2017 at 12:03 AM, Harold Tessmann <htessmann at control-tec.com
> wrote:

> Hi all,
>
> I'm trying to understand how ffmpeg calculates the start time for an
> MPEG-2 file containing one video track and one audio track. For
> example, when I run ffprobe, I get this output:
>
> Input #0, mpeg, from 'file.mpg':
>   Duration: 00:00:11.06, start: 579.376000, bitrate: 10704 kb/s
>
> As a guess, I looked at the system clock reference in the first pack
> header and the PTS/DTS values in the first PES header. When I
> calculate those values by hand, or examine the same file in Wireshark,
> I get a value that differs from the one calculated by ffmpeg. The
> first pack header gives me a system clock reference of 578.4, while
> the first video PES header gives PTS = DTS = 579.4. If I look at the
> first audio PES header, I get 579.375655555556, which is so close that
> it may be just a matter of doing math in the wrong mode (integer vs.
> floating-point). From looking at those values, I would guess that
> ffmpeg takes the minimum value from all the PES headers, but I'm not
> confident about that.
>
> I see in av_dump_format() the lines that calculate & output seconds
> (ic->start_time / AV_TIME_BASE) and microseconds (% instead of /), and
> am working backwards from there. Any information or pointers on where
> to look would be appreciated.
>
> Thanks,
> Harold


Hi,

it looks like in ffmpeg.c:

static void do_streamcopy(InputStream *ist, OutputStream *ost, const
AVPacket *pkt) {
....
if ((!ost->frame_number && !(pkt->flags & AV_PKT_FLAG_KEY)) &&
!ost->copy_initial_nonkeyframes)
return;

skips all input packets till first one with PTS/DTS.