Hi, I've just tried Ogg/Vorbis for the first time, and I'm quite impressed with the results. I haven't heard any artifacts in my tests so far, even at -b 112 (but I'm not a trained listener). Thanks for the great work! Now, some (mostly minor) points: When building libvorbis on either a Linux/libc5 or a Solaris system, both with gcc, it failed with undefined references to logf() etc. I worked around by setting NO_FLOAT_MATH_LIB in lib/os.h unconditionally. Of course, it would be better if the library would detect this automatically. How about an autoconf check for logf()? When building the libs on a foreign system without installing them in the system directories, I tried to configure libvorbis --with-ogg-prefix, but this didn't quite work. The Ogg test program compiled and linked, but didn't run because LD_LIBRARY_PATH wasn't set. To solve the problem for now, I inserted LD_LIBRARY_PATH="$ogg_prefix/lib:$LD_LIBRARY_PATH"; export LD_LIBRARY_PATH in configure (after OGG_LIBS="-L$ogg_prefix/lib" libvorbis). -- I know that it should actually be changed in aclocal.m4, but I didn't have (a current version of) autoconf on that system, so I didn't try this. (The same issue occurred in vorbis-tools for libao, libogg and libvorbis.) oggenc(1) says: : -b n, --bitrate=n : Sets encoding to the bitrate closest to n (in : kb/s). Defaults to 160 kb/s for stereo. In fact, the default seems to be 128. ogg123(1) says: : -v, --verbose : Increase verbosity. (Currently not implemented.) Well, apparently it is implemented. :-) The bit rate that is displayed with `-v', is this the rate of the current frame or the average of all frames played so far? If it's the former, it would be nice to also see the latter. When encoding a WAV file with broken header (wrong size information), oggenc used the wrong size to compute the percentage and remaining time, so they were quite wrong. Of course, that's not a bug, but since broken WAV files are not uncommon, it might be a good idea to base the calculation also on the file size. (It's not really important to me, just thought I'd mention it...) ogg123(1) says: : solaris : Solaris audio driver. Note that this driver has : not been tested! So, I just wanted to test it... but I couldn't find it. It doesn't seem to be contained in libao-0.6.0.tar.gz. Where can I get that code? Finally a question about encoding speed. Segher mentioned a 3-4x speedup recently. Is this incorporated in beta4 already, or can I look forward to it? (Beta4 is the first version I've used, so I can't compare.) Currently, oggenc is doing slightly more than 50% on my "ancient" home system (K6/233). Can I hope for realtime encoding on this machine in the future, or is this not realistic? Frank -- Frank Heckenbach, frank@g-n-u.de http://fjf.gnu.de/ PGP and GPG keys: http://fjf.gnu.de/plan --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
Frank Heckenbach wrote:> Finally a question about encoding speed. Segher mentioned a 3-4x > speedup recently. Is this incorporated in beta4 already, or can IThis is the gazillionth time people misquote me on this. A statement like "3-4x speedup" is meaningless without mentioning between which versions.> look forward to it? (Beta4 is the first version I've used, so ILook forward to it. My recent speedups are not in CVS yet, I think. They need some cleanup before Monty will accept them.> can't compare.) Currently, oggenc is doing slightly more than 50% on > my "ancient" home system (K6/233). Can I hope for realtime encoding > on this machine in the future, or is this not realistic?Look at it this way: I'm not satisfied until it runs at least 1.2x realtime on my K5 100MHz. Cheers, Segher --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
----- Forwarded message from owner-vorbis@xiph.org ----- To: Segher Boessenkool <vorbis@xiph.org> Subject: Re: [vorbis] Some comments about beta4 In-reply-To: <3AACB4A5.9B87843A@chello.nl> References: <721270A5.20010304205201.FOO-4A5E.frank@g-n-u.de> <3AACB4A5.9B87843A@chello.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-MDRemoteIP: 192.168.0.9 X-Return-Path: mat@eurogamer.net X-MDaemon-Deliver-To: vorbis@xiph.org Heya Segher, SB> Look at it this way: I'm not satisfied until it runs at least 1.2x SB> realtime on my K5 100MHz. Correct me if I'm wrong but that would make it faster than MP3 encoding? I would have thought that would be unrealistic with a more modern, more capable encoding scheme? If you can do 10X on my Athlon 1300MHz then I'll be happy :) -- Mat Bettinson - EuroGamer's Gaming Evangelist with a Goatee http://www.eurogamer.net | http://www.eurogamer-network.com ----- End forwarded message ----- --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.
> Look forward to it. My recent speedups are not in CVS yet, I think. > They need some cleanup before Monty will accept them.They're not there only due to my own business. I'm not holding them back due to unworthiness, but hosedness and a desire to be careful with one or two of them. BTW, Segher, I like that IEEE float LSP lookup hack more every time I look at it. I think I might actually make that basic technique the authoritative method of doing the last stage of computation (I'll have to choose the specific lookup table parameters carefully though ;-) Monty --- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a message to 'vorbis-request@xiph.org' containing only the word 'unsubscribe' in the body. No subject is needed. Unsubscribe messages sent to the list will be ignored/filtered.