Otto J. Makela wrote: > John Pilkington wrote: > >> Until recently I used only CentOS_5/i686 and saw no noticeable picture >> corruption. Now I also use fc10/x86_64, and on this system I do see >> corruption - small transient rectangles - when shrinkage has been applied. >> >> So it seems to me that tcrequant is *not* broken on all platforms and I >> would not like it to disappear globally unless, of course, there is a >> reliable and effective alternative. > > I've also had my share of problems with tcrequant. > > ... >> I thought it might be useful to add a note that the download from the >> company that posted the initial version of this requantiser now has a >> source date for 'main.c' of January 2008. I'm not confident that I could >> rebuild it immediately for my system, but others may not have seen it. It's >> GPL. http://metakine.com/products/dvdremaster/modules.html > > Compilation on a x86_64 system worked for me (with minor tweaks, > seems they aren't prepared for the system to be anything but a ppc or i386). > It does seem to work!
I spake too soon. Unfortunately both tcrequant and M2VRequantiser seem to work with _some_ files, but not others. Can anyone here run a test on the video clip of a few seconds found at: http://www.otto.net/~otto/requant-sample.tar.gz (3.7MiB) and see if tcrequant and/or M2VRequantiser is able to do the job for you? I've included pictures of the garbage I get with both of them. -- /* * * Otto J. Makela <o...@iki.fi> * * * * * * * * * * * * * * * */ /* Phone: +358 40 765 5772, FAX: +358 42 7655772, ICBM: 60N 25E */ /* Mail: Mechelininkatu 26 B 27, FI-00100 Helsinki, FINLAND */ /* * * Computers Rule 01001111 01001011 * * * * * * * * * * * * */