Hello,
The comment for AVCodecContext::time_base seems to indicate this variable can
be interpreted two ways, depending if the content is fixed-fps or not. However,
I don't see in the code how this is accounted for, for example, in
mpegtssenc.c, where it wants to determine the pcr_packet_period, it appears to
rely on this value being set to the 1/fps, but it does not do any explicit
checks if the content is fixed-fps or not. Should this value really just be
1/fps always? A previously developer set time_base to 1/1000 in our
application to indicate milliseconds, but this caused the mpeg2-ts output to
have a PCR value every 100 packets which didn't meet the once every 100ms
MPEG2-TS specification. Changing the time_base to 1/fps, which was 1/25 in this
case, allowed a PCR at least once every 100ms. I'm worried about the
side-effects this may cause and want to understand the true meaning for
AVCodecContext::time_base. It there a way to tell if a codec is fixed-fps?
Thank you,
Will
avcodec.h:
typedef struct AVCodecContext {
...
/**
* This is the fundamental unit of time (in seconds) in terms
* of which frame timestamps are represented. For fixed-fps content,
* timebase should be 1/framerate and timestamp increments should be
* identically 1.
* - encoding: MUST be set by user.
* - decoding: Set by libavcodec.
*/
AVRational time_base;
...
mpegtsenc.c:
..
} else {
// max delta PCR 0.1s
service->pcr_packet_period =
pcr_st->codec->time_base.den/(10*pcr_st->codec->time_base.num);
}
...
_______________________________________________
Libav-user mailing list
[email protected]
http://ffmpeg.org/mailman/listinfo/libav-user
