On Fri, 2008-02-22 at 23:57 +0100, Christian Ebert wrote: > I tested again with dv source material and: > > $ qttoy4m test.mov | mpeg2enc -f 8 -o /dev/null > > INFO: [mpeg2enc] Pass2 5379 5379( 0) I q=62.00 RECODED > INFO: [mpeg2enc] Pass2 5380 5380( 1) P q=62.00 RECODED > INFO: [mpeg2enc] Pass2 5381 5381( 2) P q=15.41 RECODED > INFO: [mpeg2enc] Pass2 5382 5382( 3) P q=62.00 RECODED > INFO: [mpeg2enc] Pass2 5383 5383( 4) P q=15.41 RECODED > INFO: [mpeg2enc] Pass2 5384 5384( 5) P q=62.00 RECODED > INFO: [mpeg2enc] Pass2 5385 5385( 6) P q=62.00 RECODED > INFO: [mpeg2enc] Pass2 5386 5386( 7) P q=15.41 RECODED > INFO: [mpeg2enc] Pass2 5387 5387( 8) P q=15.85 RECODED > INFO: [mpeg2enc] Pass2 5388 5388( 9) P q=62.00 RECODED
Looks like the quantization ran wild for a bit. I saw that sort of thing before when it was really broken. Or perhaps it's still broke. I use a script file like this to process a raw DV capture from satellite TV after I have edited it with Kino. smil2yuv -a snowmonster.wav -i 2 snowmonster.smil | y4mshift -b 8,0,704,476 | y4mscaler -v 0 -I active=704x480+8+0 -O sar=src -O chromass=420_mpeg2 | mpeg2enc -f 8 -D 10 -E -10 -4 1 -2 1 -c -q 3 -o snowmonster.m2v mp2enc -o snowmonster.mp2 < snowmonster.wav This worked rather well until the over compressing problem with mpeg2enc cropped up a year ago. I'll try a current CVS to see if I get the same results as you are getting. If so, I guess it's back to the older version. Stan ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Mjpeg-users mailing list Mjpeg-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mjpeg-users
