[VCDImager Bugs/Devel] preliminary SVCD support in cvs!

[Herbert Valerio Riedel]

fyi, I've been working this weekend on finalizing svcd support; vcdimager
should create SVCD's now, except for an 'empty' /SVCD/SEARCH.DAT and a
/SVCD/TRACKS.SVD which misses the playing time (set to 00:00:00)

till now I've been able to write the videocd.bin in a linear way, that is,
w/o having to fseek() around while outputing this file -- I'd like to keep
it that way, in order to be able to pipe/redirect the output sometime...

but I fear, either I have to scan the mpeg's twice or I'll have to scan
them while mastering in one pass... :-/ none of this possibilities is
nice. I'd go for the first one, since it's the cleaner one, even if
inefficient...

On Sun, 19 Nov 2000, Mike Bernson wrote:
> I just try the code for burning a svcd with a vaild
> mpeg-2 stream and it worked but not time counter.
that is, you could actually play it on a DVD player? did it recognize it
as a SVCD? (I've lended my DVD player to someone, so I can't test myself
yet)

...if so, that'd be a great step forward! :-)

(btw, you had problems with VCD's some weeks ago, this should have been
fixed now, it was some bug in the PSD like I suspected)

> Is the next stage for me to parse the mpeg stream and
> build up an array of where each frame is. I can see keeping
> the display timestamp, decode timestamp, offset in stream,
> and frame type IBP. This should also tell how long the
> stream is.

> I should be able to detect if mpeg-1 or mpeg-2 stream.
> Does this.

btw, I've already asked andrew if he could investigate how to tell
apart/detect if a stream is mpeg1 or mpeg2 -- don't know if he has already
got time to look after this...

-- 
Herbert Valerio Riedel      /     Finger address@hidden for GnuPG Public Key
GnuPG Key Fingerprint: AC2A CD57 A5C8 A1CB 0A18  DA95 CB0B DB23 60B6 16F5