This one is bugging me.  I have scanned the archives and see many people
confused by this one but none of the proposed explanations seem clarify
things at all!

 

I have dumped the first few frames of an mpeg2 file and get:

F1 - dts 16501

F2 - dts 18001

F3 - dts 19501

F4 - dts 21001

Etc.

The codec context time_base for the stream has num=1, den=60 (60fps).

 

I want to convert the dts values into 'double seconds' but if I try
using the timebase converters (presumably incorrectly) I get crazy
numbers.

Clearly the frames are separated by 1500 - but how does this relate to
the 60fps (time_base)?

 

Also - while we are at it - has anyone noticed that if you av_seek_frame
to a timestamp and read off a few frames - the frame_number field of the
codecContext might report frames 1..n.  If I repeat the seek and read
off the same frames, the codecContext reports frame numbers n+1..n+N
(i.e. they continue from the last frame decoded).  Surely they should
read the same each time I decode the same frames?  This is really
annoying for me since I need to use the frame_number's for referencing
the image! 

 

Any help would be welcome?

Gary Smart 



 

_______________________________________________
libav-user mailing list
[email protected]
https://lists.mplayerhq.hu/mailman/listinfo/libav-user

Reply via email to