If one follows the example example/api2-samples/api2-decode-encode-video.cpp, a VideoDecoderContext won't have a valid timebase, yet when decoding, each individual frames will be converted to the timebase of the videodecoder (default-constructed 0/1):
If one follows the example
example/api2-samples/api2-decode-encode-video.cpp
, aVideoDecoderContext
won't have a valid timebase, yet when decoding, each individual frames will be converted to the timebase of the videodecoder (default-constructed 0/1):https://github.com/h4tr3d/avcpp/blob/4aaaaa11d18f5f4d2e364796dffada105943ba48/src/codeccontext.cpp#L928
As a result, frames have invalid PTS. This is one of the problems of #113 - that is solved by the user code manually counting the frames.