Dan Conti
2002-Feb-20 14:55 UTC
[vorbis-dev] Re: [iPAQ] Ogg Vorbis fixed-point math anyone ?
Actually, we have a integerized version of the vorbis libraries that are available, royalty free, with a simple attribution requirement. These libraries are 100% C code and run faster than real time on a cirrus 7312-74mhz processor (arm core). My guess is they should run at about 20-30% cpu utilization on a strongarm/200. Monty has this same code, but his version doesn't play b4 content real time. For more information, feel free to contact me off-list.> -----Original Message----- > From: owner-vorbis-dev@xiph.org [mailto:owner-vorbis-dev@xiph.org]On > Behalf Of Ian Walters > Sent: Wednesday, February 20, 2002 2:33 PM > To: Gregor Hoffleit; ipaq@handhelds.org > Cc: vorbis-dev@xiph.org > Subject: [vorbis-dev] Re: [iPAQ] Ogg Vorbis fixed-point math anyone ? > > > Yes and no. > > I tried it but not seriously. I only went as far as floating point > emulation, not trying to understand the algorithms and turn them to > fixed-point/integer. I have been watching the ogg-vorbis > lists though, and > it would seem to be only one part of it that remains to be properly > integerized. Unfortunately from the sounds of things it > would take someone > who really knew what they were doing to transform that last > bit. I don't > remember what it was, just that when I Googled for the > definition what I > found made very little sense to me. > > The emulation proved insufficient as the quality was way, > way, way down. And > the normal floating point emulation (e.g. the automatic > stuff) wasn't fast > enough. > > Ian. > > On Thu, 21 Feb 2002 05:38 am, Gregor Hoffleit wrote: > > A few months ago, there was some discussion about adding fixed-point > > support for iPAQ's StrongARM to Ogg Vorbis (whether and how it's > > possible, that is). Has anybody tried this yet ? > > > > Gregor > > _______________________________________________ > > Open Source Software for Compaq iPAQ Handhelds. > > iPAQ mailing list > > iPAQ@handhelds.org > > http://handhelds.org/mailman/listinfo/ipaq > > irc://irc.openprojects.net #ipaq > > --- >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-dev-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. >--- >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-dev-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.
Ian Walters
2002-Feb-20 16:44 UTC
[vorbis-dev] Re: [iPAQ] Ogg Vorbis fixed-point math anyone ?
Really? I would be very interested in this. Does this include encoding as well as decoding? Any and all information would be greatly appriciated, I would like to see this added to Trolltech's Qtopia. Ian. On Thu, 21 Feb 2002 8:55am, you wrote:> Actually, we have a integerized version of the vorbis libraries that are > available, royalty free, with a simple attribution requirement. These > libraries are 100% C code and run faster than real time on a cirrus > 7312-74mhz processor (arm core). My guess is they should run at about > 20-30% cpu utilization on a strongarm/200. Monty has this same code, but > his version doesn't play b4 content real time. > > For more information, feel free to contact me off-list. > > > -----Original Message----- > > From: owner-vorbis-dev@xiph.org [mailto:owner-vorbis-dev@xiph.org]On > > Behalf Of Ian Walters > > Sent: Wednesday, February 20, 2002 2:33 PM > > To: Gregor Hoffleit; ipaq@handhelds.org > > Cc: vorbis-dev@xiph.org > > Subject: [vorbis-dev] Re: [iPAQ] Ogg Vorbis fixed-point math anyone ? > > > > > > Yes and no. > > > > I tried it but not seriously. I only went as far as floating point > > emulation, not trying to understand the algorithms and turn them to > > fixed-point/integer. I have been watching the ogg-vorbis > > lists though, and > > it would seem to be only one part of it that remains to be properly > > integerized. Unfortunately from the sounds of things it > > would take someone > > who really knew what they were doing to transform that last > > bit. I don't > > remember what it was, just that when I Googled for the > > definition what I > > found made very little sense to me. > > > > The emulation proved insufficient as the quality was way, > > way, way down. And > > the normal floating point emulation (e.g. the automatic > > stuff) wasn't fast > > enough. > > > > Ian. > > > > On Thu, 21 Feb 2002 05:38 am, Gregor Hoffleit wrote: > > > A few months ago, there was some discussion about adding fixed-point > > > support for iPAQ's StrongARM to Ogg Vorbis (whether and how it's > > > possible, that is). Has anybody tried this yet ? > > > > > > Gregor > > > _______________________________________________ > > > Open Source Software for Compaq iPAQ Handhelds. > > > iPAQ mailing list > > > iPAQ@handhelds.org > > > http://handhelds.org/mailman/listinfo/ipaq > > > irc://irc.openprojects.net #ipaq > > > > --- >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-dev-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. > > --- >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-dev-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.--- >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-dev-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.