On Wed, Sep 17, 2003 at 03:08:25PM -0700, Steven M. Schultz wrote: > > On Wed, 17 Sep 2003, Richard Ellis wrote: > > > Has anyone else noticed that the -Q parameter to mpeg2enc v 1.6.1.90 > > seems to have much less effect than it did in version 1.6.1? I've > > ... > If you look at the CVS info: > > http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/mjpeg/mjpeg_play/mpeg2enc/ratectl.cc > > and look at the difference between the 1.4 version and > the previous version (1.3) you'll see the section that was > rewritten back in January 2003.
Thanks for the pointer, yes, I looked at the diffs and the code that is driven by Q (in that section at least) was reworked quite a bit in January. That would account for the difference I'm seeing in the new mpeg2enc's Q that does not seem to be having much effect at times. > Ideally -X should be documented of course ;) Actually, if no one else wants to, I can do some documentation patching and submit a diff to the list of the changes. Of course, for some things, like -X, I'd have to know what to write, otherwise, all I'd have to go on is "-X sets a parameter that controls xyz variable in file pdq...". > Do the files compare identical (with 'cmp' for instance)? If so > then -Q had no effect at all. If not then it had a slight effect > on a few macroblocks here and there but not enough to make a file > size difference. The differ only after 413,364,142 bytes (about 1/2 of the way into the files). So it appears that, the new -Q had zero effect for that pair of encode runs. The total file sizes differ by only 4 bytes. > Perhaps -N <num> (num =0.0 to 2.0) would work as well. In 1.6.1 > and earlier -N was hardwired to what is now called 1.5 but in > the CVS/1.90 version -N takes a parameter. I've found -N 1.5 to > be a bit too "aggressive" in some cases but 1.0 works well and > preserves more of the quality. For fine tuning I've found that > 0.5 to 0.9 to be of use (when I cared about the size ... That is a good suggestion, thanks. The new adjustable -N value should be able to accomplish much the same level of "fine tuning" that I was using (abusing??) the old -Q setting to achieve. The tunable -N may even be a better fine tuning system than the old -Q setting was as well. I'll give this a try and see how it works out. ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Mjpeg-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/mjpeg-users
