bug-vcdimager
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[VCDImager Bugs/Devel] MPEG end codes in 0.6.2 (mjpegtools mplex)


From: Andrew Stevens
Subject: [VCDImager Bugs/Devel] MPEG end codes in 0.6.2 (mjpegtools mplex)
Date: Wed, 16 May 2001 19:24:19 +0200

Hi Herbert,

> > On Sun, 13 May 2001, Herbert Valerio Riedel wrote:
> >
> > I've identified the problem,
> >
> > lavtools (you mean mjpegtools, don't you?) seems to create the last
> > packet without an pack header... it starts directly with an program end
> > marker... or as a alternate interpretation, the last packet is 2328 bytes
> > long, which vcdimager can't handle... (the older vcdimager had a bug,
> > when allowing it...)
> >

I'm trying to figure out the issue with the end-codes in a program stream.
I just checked my ISO 13818-1 and the syntax for an MPEG-2 program
stream is given as:

MPEG2_program_stream() {
        do {
        pack();
        } while(nextbits == pack_start_code );
        MPEG_program_end_code;
}

I.e. a program end code (or system stream end code) should definitively sit
after the end of the last Pack and not inside it.

Currently mplex writes the end-code into the start of the last sector
and then packs the rest of the sector with 0's.   Any idea what the correct
S(VCD) conformant solution is.  My guess would be a padding packet
4 bytes short of the sector size followed by the end code...

        Andrew



reply via email to

[Prev in Thread] Current Thread [Next in Thread]