I'd go along with what the two Brians have said - you're vastly better off 
doing the mixing yourself.

Two extra (and largely architecture independent points):

1) If you're mixing 200 oscillators, then you'll need eight bits of headroom 
just to do the mixing; if you have a polyphonic synthesiser with a thousand or 
more oscillators per voice, you'll need even more.  Bearing in mind that you 
also want to generate the sine waves with as many bits of precision as you can, 
it's easier to manage (or at least debug) this stuff if you do it yourself.

2) Most machines these days are SIMD.  There are a variety of algorithms for 
producing efficient sine oscillators on such architectures, but the output may 
be time- or oscillator- interleaved in odd ways.  Again, if you're managing the 
mixing buffer yourself, this is easier to handle.

As you can see (www.fdsynthesis.com), I got bored with this before CoreAudio 
even reached phones:-), but some of it may still be relevant.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Ian Kemmish                         2 Sale Mews, Biggleswade, Beds SG18 0AY
[email protected]        Tel: +44 1767 318605     Mob: +44 7952 854387
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



 _______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list      ([email protected])
Help/Unsubscribe/Update your Subscription:
https://lists.apple.com/mailman/options/coreaudio-api/archive%40mail-archive.com

This email sent to [email protected]

Reply via email to