View Single Post
Old 22nd February 2006, 21:39   #13  |  Link
SadaraX
Registered User
 
SadaraX's Avatar
 
Join Date: Mar 2004
Posts: 81
Quote:
Originally Posted by Teegedeck
Windows for encoding, Linux (SUSE) for the rest. But I'm a dumb GUI-user, couldn't compile my own shorts. Any chance it comes as an rpm?
What, Avidemux? Probably. In fact, I'm sure about it. www.avidemux.org, pick a mirror and download what you want. If not, we can probably help you compile. I compile from the SVN archives fresh almost daily.

Quote:
Thanks again. Mainly I was about the 1st pass in Avidemux currently using the same settings as the 2nd pass (or so it seems to me); that's very good in a way and known as the 'full quality first pass' option in XviD's VfW config. But for people who use quite high-quality settings that results in a very slow first pass. It's really not necessary for the 1st pass to use VHQ=4 for example. Someone who uses all the HQ options can get a speedup of about 80 percent for the first pass if that one runs without GMC, QPel and chroma ME, perhaps engages the 'turbo' switch even (using faster algos), lowers ME precision to '5' and, most importantly, never sets VHQ mode higher than '1' (mode decision). The filesize discrepancy should still not be that big. Quantizer distribution in 2nd pass will still be good, and very good (almost constant quantizer if you don't use curve compression) if you've used a constant quantizer in first pass that is quite close to what you could expect for average quant in second pass. Have a look at Didée's posts in this thread for an explanation.
I will have a look at that, since I think I understand a little better, but not fully. If it is something so important, hopefully it can be addressed in the XviD and h264 update that will be coming soon.
SadaraX is offline   Reply With Quote