As it’s run its short commercial life, Xiph.org have freely released the integer decoder for Ogg Vorbis – that is, the version that will be usable by the processors in most portable compressed music players. If the manufacturer adds support.
You, the consumer, have an important part to play: write to the manufacturer of your MP3 player, asking for Ogg Vorbis support. I’m looking at you, iPod owners.
I thought most MP3 players didn’t use microprocessor algorithms to do the decoding, but instead used hardware-based MPEG-decoding chips to do the work using less current. Then again, I suspect that things like Windows Media rely on microprocessor code, so there may be a niche for Ogg there (not counting RIAA harrassment, of course). (Are there actual Windows-Media-in-hardware chips?)
The ‘hardware-based MPEG-decoding chips’ are generally a DSP with its own ROM, which may or may not be flashable. Plenty of hardware players use the StrongARM (e.g. the iPod), which the Vorbis integer decoder apparently runs very nicely on.
How much difference is there between code for a DSP and integer C code? I was under the impression that (a) DSPs did floating-point quite easily, and (b) DSP code was quite different from imperative machine code in its organisation, to the point where you couldn’t just make it out of a C library. Though perhaps that’s why Vorbis have made the offer of free engineer time.
my cat’s breath smells like cat food
OK, OK :-) A more comprehensible reason for going Ogg is on Slashdot. In listening tests, many people picked the 128kbps Ogg as the original sample …