> >
> > but it is very inflexible as i must have one program driving both input
> > and output.
>
> The AWE64 is capable of hardware mixing. Not entirely sure on this one,

you are speaking of something else - while it's very interesting too.
no mixing is needed. just allowing /dev/sound or /dev/audio to be opened
once for reading and once for writing and to work full-duplex this kase.

> but I remember reading that the EMU8000 in the AWE32 can handle up to 32
> simultaneous voices. You might be interested in doing something similar
> to what I did recently to the esa(4) driver -- call audio_attach_mi()
> multiple times, one for each "voice". Then, /dev/sound[0-n] are
> available at your disposal.
>
> Jared
>