[vlc-devel] IS0/IEC13818-1 standard compliance of MPEG2 TS stream.
Battula Kishor Babu
contact.battula at gmail.com
Fri Jan 22 13:03:24 CET 2010
I happened to see the streamed content of MPEG2 TS through a file based
analyser and want to make few comments about the generated TS stream and its
compliance to the MPEG2TS (ISO/IEC 13818-Part1) standard.
1. It appears that VLC tries to maintain decode delay (Diff of DTS and PCR)
of both audio and video to be the same. Stream independent 'dts-delay' also
confirms me that same delay is forced upon all the elementary streams
present in the TS (of multiple programs). Such implementatation is bound to
overflow / underflow buffers just before the decoders (D1, D2 .....Dn in the
T-STD standard notation as shown in Figure2-1 of IS0/IEC-13818-1. All
references will be made with this document ) as audio and video buffers size
vary greatly depending on number of channels for audio and level and
profiles for video. Please refer to section 2.4.2 for description on buffers
and the constraints on buffers management . Why such a design decision is
taken which is not compliant to the standard?
2. The 'shaping' param is simply superb to maintain the line bitrate.
Similarly the configurability for PMT PID and PCR intervals.
3. Is there a choice to choose a stream carrying PCR information? It appears
that PCR is always carried with video elementary stream.
4. I see that the interval between successive PCRs is going beyond 100ms. Is
there a tight control over PCR intervals and table scheduling intervals?
at the end, i can Update/program TS mux to include above mentioned features
once i realize the motives behind choosing to violate standard and help is
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the vlc-devel