Speed During Encoding Avigreenspun.com : LUSENET : Video CD : One Thread
Hi, I am using the LIGOS lsx demo version encoder and also avi2vcd from www.mpeg.org (I have the source code too). My problem is about the speed of encoding. For say 2 sec. avi file LIGOS is taking 11 seconds and avi2vcd is taking 40 sec. Imaging the results for big AVI files. I want to know WHAT ARE TEH PARAMETERS DURING ENCODING OR DURING MULTIPLEXING WHICH EFFECTS SPEED OF ENCODING. Or other ways, how can this problem be rectified apart from skipping B frames concept.
-- Rahul Saxena (firstname.lastname@example.org), April 08, 2000
Speed of encoding is determined by a few things. For one, the encoding algorithm of the software encoder makes a difference. Some encoders are slow to encode regardless. And speed of your processor is key when doing software encoding. As far as settings in the encoder, there should be some that would speed up encoding, but quality is compromised.
-- MrVCD (email@example.com), April 08, 2000.
I know how you feel Rahul. I tried Xing's and Ligos software and they were just too slow for me. I don't know how fast Panasonic's is. I just finally broke down and spent the extra bucks for a hardware assisted encoder. I had checked them out for months and had decided to get either the Broadway 4.0 or the Vitec RT-6. The vitec was hard to get info on so I went with the Broadway which cost me $720.The Vitec was around $600 and then there is also the MPEGator which is supposed to be very good but I think it is over $1000. With the Broadway you will get a 2:1 ratio if you have a Pentium II and a 1:1 ratio with a Pentium III
-- Al McCraw (firstname.lastname@example.org), April 08, 2000.
The hardware encoding setups produce good-quality MPEG files but not nearly as good as that produced by a carefully tweaked system using, say Panasonic encoder. It is true software encoding is slow (in fact on some systems glacial): a QCIF(352x240) MJPEG-coded AVI clip with full 44k/16b/stereo audio about 22 minutes long (which is the 2Gb filesize limit captured with Marvel G200) takes about 2 hours 50 minutes to encode with my former system PII, 400MHz, 64Mb RAM. I have since upgraded to PIII, 600MHz and now encoding takes 1 hour 30-40 minutes for the same clip: less painful. Nonetheless, the bottomline, for me, that you might consider too, is that something with high quality that can last for a hundred years or more should not make one fret too much about the few hours or so it would take to encode. It is always daunting at first when presented with the prospect of having to wait for hours for encoding to finish. But surely one can put the time to do other worthwhile things?
-- EMartinez (email@example.com), April 09, 2000.