On Fri, 11 May 2007, Jonathan Woithe wrote:
> For a long time there have been known issues with encoding at low quality
> factors such as q=2 (ie: high quality encoding) on ix86. I understand this
Not -q again. Seems like just yesterday there was (yet another)
discussion about it.
> was due to something related to potential overflows in some parts of the MMX
> code. From memory the general recommendation from a year or so ago was that
"or so" ;) The mmx code was rewritten/updated/modified between
December 2004 and March 2005 to workaround/fix the issues with
overflow in the mmx routines.
> on ix86 one shouldn't go below a "q" value of 3.
That's still, if you understand what -q is and so on, good advice
most of the time ;)
> What is the current status on this - does mpeg2enc still suffer this
> limitation or has it been dealt with?
The current state, since approximately Feb 19, 2007, is that mpeg2enc
is BROKEN and should not be used. If you DO use it then -q of 4 or
higher MAY produced uncorrupted output.
Last known working mpeg2enc is RC1. RC2 is broken. Current CVS
has mpeg2enc effectively disabled (you can remove the error_exit call
if you want to use the program).
That is why I mentioned the BitVice encoder - not as flexible but the
output is fantastic (6Mb/s never looked so good!).
Cheers,
Steven Schultz
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Mjpeg-users mailing list
Mjpeg-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mjpeg-users
