Preview window problems in Mux

[gconnery]

I'm in the mux panel.  I add an h.264 elementary stream.  The preview panel will now allow me to watch the video if I want.  However, if I hit play, the slider will run into the right hand side quite quickly, in about 14 seconds.  The sample is 15:00 minutes long though.  When it gets to the right side the video will continue playing (I think it stops in the remux window oddly enough).  Although the slider is already at the right end, you can reposition it and it will correctly position the player in the stream, and allow you to play from there, however it will hit the right hand side again in about 14 more seconds.

[Vovs]

There is several methods for determination duration of AVC streams. Our demultiplexer uses BitRate[ SchedSelIdx ] value for that. The BitRate[ SchedSelIdx ] is calculated like ( bit_rate_value_minus1[ SchedSelIdx ] + 1 ) * 2(6 + bit_rate_scale) ((E-37) - t-rec-h.264-200503-p).

Streams that you sent us has bit_rate_value_minus1[0] = 712512 and bit_rate_scale = 1. Calculated average bitrate is 91 200 664 therefore demultiplexer is "mistaking" circa 50 times.

Perhaps more correct method of determination stream duration is present but I don't know what that is. May be anybody can give me hint about that?

[gconnery]

The sample I gave you is a transport stream.  Transport streams have PCRs.   If the sample is CBR, then the PCRs will quickly converge on a rate if you just calculate bytes between PCRs / time between PCRs.  If it is a CBR stream, then you can just look at how long it is in bytes and convert that to time.  Easy.  Of course this is transport time rather than play time, but these will generally not differ by much and should be sufficient for the slider.

If the sample isn't CBR, then you could use the PCR delta to figure out the length in most cases, though some samples might have PCR discontinuities in them.  Here you could simply look at the first PCR in the sample, and the last PCR in the sample.  The delta will be some number of 90kHz ticks.  Since PCRs are 33 bits they will only roll over after 26 hours or so, so a simple delta should work fine.  Again, this should work most of the time.

If however the sample is not CBR and has some PCR discontinuities, then...

[Vovs]

Unfortunately our demultiplexer not uses the PCR for navigation at all. Your idea looks attractive and we need to discuss it with demultiplexer developers.