On Wed, 7 Sep 2005, Florin Andrei wrote:
> Well, it could be a brief bitrate spike or something like that. mplex
> would not complain, yet the stream would be non-compliant.
Sure it would complain - at the _least_ you'd see the numbers in the
Average/Peak summary at the end. If the peak is less than 9800 you're
ok. You'd get a buffer underrun if the rate was too high for the
buffering. The only time I've had any problems is when mplex gives
any "warning" messages - they're not warnings, they're fatal as it turns
out. No message, no problem.
> I guess I was lucky, mine plays everything I throw at it (including old
> SVCDs made years ago with mjpegtools).
"everything"? Then why are we having this discussion? :)
I maintain that if the streams work in most players and not in
one particular player then the problem could be in that player.
> But some older mjpegtools versions never had this issue, even at 8500,
No, they had the problem. I recall various postings with "8500" and
the advice was the same then as now ;)
Besides, nothing's been done in ages in the encoder except to disable
the threading, fix DualPrime, and use fixed GOP sizes (the variable
gop size logic is inoperative). I'll check the cvs logs but the
rate control hasn't been rewritten/redone in a long time is what I
seem to recall.
I may not have seen the problem though since my problem is getting
the bitrate HIGH enough these days (-q 3 and the average still came
out around 4100Kb/s). Rather embarassing to have 500MB left on a
DVD for a 2hr movie - haven't needed to get anywhere near 8000+ in a
long time. Most DVDs I've looked at aren't that high either (not
too surprising really since they're coming from noise free master
film).
> forum, but you'll figure out anyway... Let me just say that you'll have
> to trust me when I say that I tested material generated with commercial
> software on the same cheap DVD-Rs and the problem never occurs.
Which commercial software? I gave up on "Compressor" since it added
(maybe fixed now) the NTSC setup when it shouldn't - washed out the
blacks rather nicely. And you'll just need to take my word for it
that the commercial DVD authoring software I use accepts mpeg2enc's
output without complaint and the discs run fine (but then then Philips
players *actually do* play anything without regard for mpeg2enc or
not ;)).
There are a lot of variables involved - encoder on thru the player.
> Not high bitrates generated by commercial tools. Those go right up to
> 9800 and have no issues with the same "recordable media" on my rig.
Hmmm, the forums I follow show all manner of problems with >8000 and
even using the 2pass encoding the "Commercial" encoders still have
rate spikes at times. And the authoring software I use cautions
against rates greater than 8000Kb/s in several places - player
compatibility is the cited reason.
> mjpegtools is unique among the open source projects I'm following
> closely, with regard to the extremely low frequency of releases. Often,
> a bug is fixed only in CVS and it sits there idle. Meanwhile, most users
My turn for "sighing". Folks can't spell 'c', 'v', 's' then they'll
just have to use old software. ffmpeg, mplayer, etc don't have
releases as such (or so it seems) - you're expected to use cvs.
And as I mentioned or at least tried to imply - it's important to get
this release out this weekend and see if the curse of the extremely
low frequency releases can be broken. THEN and IF the problem can
be reproduced it can be fixed - at the moment we've a sample size of
1 and 1 (one person with one difficult player) and that's just not
enough to work with...
> use packages, but package repositories only follow "stable" releases
That's their problem to a degree. I think my feelings about package
systems are well enough known to not need repeating.
> Too bad, since it's the best open source MPEG2 encoder I'm aware of.
Thanks! :) If you think there's rate control issues with mpeg2enc
you should try ffmpeg sometime <grin>
But have you tried any bitstream verifiers and/or other DVD authoring
software to see if it complains about the files? If you're using
(or have access to ) a commercial encoder then how about some
commercial DVD authoring software?? If _that_ declares a file invalid
I'd give that a lot more weight than a single brand of DVD player
having a playback issue. If a bitstream verifier says "stream invalid
because XXXXX" then that's a BIG HELP and something that can be
reproduced and investigated.
The main problem is that we really don't have someone to deal with
subtle problems in the encoder. Build/compile problems yes, but
I ended up stubbing out the threading logic because there's a race
condition buried down in the C++ code and some folks have an allergy
to C++ (no loss though since the threading didn't gain much anyhow).
And with that I think I'm going to bed :)
Cheer,
Steven Schultz
-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Mjpeg-users mailing list
Mjpeg-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mjpeg-users
