Re: [Mjpeg-users] mpeg2enc CVS version

2004-01-13 Thread Steven M. Schultz

On Mon, 12 Jan 2004, James Finnall wrote:

> Thank you for all your help Steven, have CVS version built 
> and installed now.

Hurrah!

> Built and installed the SDL, after I googled to find out what it is.

AH, so my (strong) suspicion that the sdl.m4 file was the root
cause of the problems turned out to be a good guess.   The SDL
should be optional (but if you want to run yuvplay I think the SDL
becomes required - so it is just as well that you have it installed
now).

> Still had the quote warnings though from the other packages but it
> didn't stop it.

Those I think are automake-1.8 being more verbose/picky - I've never
seen those before with 1.7.9 or earlier.

> dvdauthor.  Perhaps you can clarify if what I am doing is correct before
> I post to the dvdauthor list.


> I created 10 photo scans and saved them as jpeg files and resized etc.
> Used jpeg2yuv to create the video stream and piped to mpeg2enc.  I used
> -f 8 for the format code code on the output.  Projects for me would be 
> NTSC so gave proper arguments for that.  Each of the streams should
> be 6 seconds in length, cat them together for a total of 60 seconds.
> 
> Copied the 60 second clip into a second file and cat'd the copy into the
> output file 29 times, for a total of 30 minutes of video.  Then used the same

I do not believe  that 'cat'ing mpeg files together is the correct thing
to do.Better method would be to replicate the source frames and
feed a large stream of input into mpeg2enc.

Instead of creating 6 second clips and replicating 10 times it may
be necessary to create a single 60 second movie.

> Created an xml file for just a simple DVD, no menus or anything. Tried to 
> use dvdauthor with the -x project.xml file and it keeps reporting:
> 
> dvdauthor: relocation error: dvdauthor: undefined symbol: xmlNewTextReaderFilename

That looks like a libxml2 version problem - I do not know what
the minimum version of libxml2 is required by dvdauthor but I suspect
a newer version of libxml2 that contains xmlNewTextReaderFilename()
is needed.

> So I changed to;
> dvdauthor -o test -f dvd/video.mpg
> 
> It creates the DVD structure dir in test and the a zero byte VOB file.
> Then it keeps reporting the same message over and over never ending.
> 
> WARN: System header found, but PCI/DSI information is not where expected
> (make sure your system header is 18 bytes!)
> WARN: Skipping sector, waiting for first VOBU...
> WARN: Skipping sector, waiting for first VOBU...
> WARN: Skipping sector, waiting for first VOBU...

That's the type of error I think comes from cat'ing mpeg files
together.

> the program is waiting for.  At this point I am wondering if the problem is
> in using mp3 files that are 44 khz.  I read somewhere that 48 khz is required
> for DVD's.  But since mplex accepted the files, would it have resampled as 

Quite true - the audio needs to be 48000 for DVDs.   While there are
several video sizes that can be used the audio has to be at 48000.
I don't think that's the the problem in this case though.

> If this is a dvdauthor type of problem and everything is OK with mjepgtools,
> then I will post there.

At this point it looks like a dvdauthor issue.   I'm not sure how
to create a slideshow with dvdauthor - but if it's possible then 
dvdauthor probably has rules/guidelines on how to do it.   Perhaps
creating mpeg still images and crafting some xml to say how long
to display each image, etc.   

Looks as if the mjpegtools-cvs configuration battle has ended in
victory though - and thats good!

Cheers,
Steven Schultz
I thought involved



---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users


Re: [Mjpeg-users] SVCD production trouble

2004-01-13 Thread Martin Collins
On Mon, 12 Jan 2004 20:12:07 -0800 (PST)
"Steven M. Schultz" <[EMAIL PROTECTED]> wrote:

> On 12 Jan 2004, Craig Lawson wrote:
> 
> > Can someone help me out with a command line recipe for SVCD
> > production? I am following the MJPEG Tools instructions, and I end
> > up with an SVCD that works great with Xine on my system, but is
> > unrecognized by both Macs and Windows, even when I directly open
> > the mpeg file on the SVCD.
> 
>   I have no idea if windows knows what a SVCD is or not,

Go to www.doom9.org for this info. I understand SVCD is similar to XCD
which needs the Riff CD/XA filter. I set this up yesterday on Win98 and
it works fine for XCDs with Zoom Player.

Martin


---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users


Re: [Mjpeg-users] kvcd or tmpgenc better with same filesize

2004-01-13 Thread Andrew Stevens
On Tuesday 13 January 2004 06:27, Steven M. Schultz wrote:
> On Mon, 12 Jan 2004, Al Bogner wrote:
> > What do you think makes a better encoding from excellent sources: the
> > kvcd or tmpgenc matrix if the result should have the _same_ filesize.
>
>   tmpgenc of course.   the kvcd tables were designed to reduce the bitrate
>   to allow for longer playing time - once detail/contrast/sharpness is
>   lost it can not be recovered by raising the bitrate.
>
> > With kvcd you can use higher bitrates. I have to encode a movie with
>
>   Yes, and defeat the purpose of using the kvcd tables in the first
>   place ;)   Depending on the material you can see the effect of the
>   kvcd tables - slight softening or less sharpness at times.

Seconded... the kvcd matrices discard fine detail very aggressively. For good 
quality material this is not needed and can make gradations in shadings more 
visible than necessary.

> > more than 2 hrs at a bitrate between 4000-4500kbps.
>
>   For clean material I'd start with "-q 6 -K tmpgenc -E -10" - that's
>   been a good combination of parameters for me.

Andrew



---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users


Re: [Mjpeg-users] Converting DVD from PAL to NTSC

2004-01-13 Thread Andrew Stevens

> > The final result was a smooth flowing image (on my DVD player) with a
> > bit less quality than the original - it's a bit blotchy in certain
> > scenes. The original Dolby Digital (2 channel) sound was preserved.
>
> There the -q option might help.

If it is just certain scenes then you need to boost peak bitrate (the scenes 
are running out of bits!).

Just for reference, the peak bitrate on commerical DVDs is often 8000Kbps or 
more!

Andrew



---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users


Re: [Mjpeg-users] BUZ Card Upgrading

2004-01-13 Thread Ronald Bultje
On Mon, 2004-01-12 at 04:00, Lance F. Squire wrote:
> I would prefer something Linux would 'find' instead of having to point 
> to it.
[..]
> I would like something that also included a SCSI controller. For a one 
> piece replacement.

You are weird... ;). The Buz really is one of the few cards that can do
this, and the only one that I know to be supported. (Note to Steven:
yes, the Buz has a built-in SCSI controller... That card is wacky.)

> If you're woundering, the upgrade is an AMD64 on an ASUS KV8 board. If I 
> can get it to work on that, that would really save me $. ;)

What isn't working?

Ronald

-- 
Ronald Bultje <[EMAIL PROTECTED]>
Linux Video/Multimedia developer



---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users


Re: [Mjpeg-users] kvcd or tmpgenc better with same filesize

2004-01-13 Thread Al Bogner
Am Dienstag, 13. Januar 2004 22:09 schrieb Andrew Stevens:

> > > With kvcd you can use higher bitrates. I have to encode a movie
> > > with
> >
> > Yes, and defeat the purpose of using the kvcd tables in the first
> > place ;)   Depending on the material you can see the effect of the
> > kvcd tables - slight softening or less sharpness at times.
>
> Seconded... the kvcd matrices discard fine detail very aggressively.
> For good quality material this is not needed and can make gradations
> in shadings more visible than necessary.

So, if size doesn't matter is it _always_ better to use tmpgenc, also 
with _bad_ sources?

I mean, when I encode an old VHS-movie with poor quality, could the 
result be _better_ with kvcd or does kvcd safe space only?

Al


---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users


Re: [Mjpeg-users] how to join m1v and mp2 files?

2004-01-13 Thread Andrew Stevens
On Monday 12 January 2004 14:44, Dragon_at_work wrote:
> How would I join multiple m1v and mp2 files in a way that would later lend
> itself to mplexing them.
>
> I tried cat. And it seemed to work for both types.
>
> But, mplex was unable to join them properly --ie the video worked, but the
> audio did not.

Thats very odd... both should work.  However, depending on if the m1v
has sequence end markers in it mplex may be prone to attempting to split
the multiplexed result at the joins.  There should be absolutely no problem
with audio (unless you changed encoding paramters mid-stream, but even then it
*should* work just fine).  How did the audio go wrong: any error messages?

Andrew
PS
If the source of the .mp2 is a demultiplex 'rip' from a Digital TV or 
SVCD/DVD/VCD stream you may have difficulties.  The audio frames don't align 
with the program stream packets which would mean the last one of each rip may
be cut-off in the middle.



---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users


Re: [Mjpeg-users] kvcd or tmpgenc better with same filesize

2004-01-13 Thread Steven M. Schultz

On Wed, 14 Jan 2004, Al Bogner wrote:

> So, if size doesn't matter is it _always_ better to use tmpgenc, also 
> with _bad_ sources?

Well, if size doesn't matter you can use '-K hi-res' :-)

What I use for high quality encoding is a custom table made up of
the hi-res Intra matrix with the tmpgenc nonIntra matrix and then
-K file=myfile

> I mean, when I encode an old VHS-movie with poor quality, could the 
> result be _better_ with kvcd or does kvcd save space only?

kvcd's claims to fame are 1) smaller filesizes (lower bitrate) and 2)
reduced splotchiness (pixelization I think is the term used) during
dark scenes.   The kvcd folks never made any claims about increased
quality.

For old VHS movies that have been run thru yuvdenoise and/or the
medianfilter the use of kvcd tables won't really hurt the picture too
much more.  A good reason to use the kvcd tables in this situation 
would be to reduce the blocks/splotches during dark scenes.

Experimentation is an important part of the process - try it and see
what happens.

Cheers,
Steven Schultz



---
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
___
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users