Hi,

I'm writing an application using directsound that captures, mixes and plays
back sound. During the mixing phase I allocate a buffer using a standard
call to new, to use as temporary storage. I would like to know how i can
have this temporary buffer allocated on device memory so that i don't have
the overhead of data being captured on the sound card RAM, then being sent
to system RAM, mixed, then sent back to the device for playback. I assume
that it may have to do with overloading the new operator, but i'm not sure
how to specifiy the memory location to allocate from, etc...

Any help is appreciated. If this is not the right group to ask, can you
direct me somewhere else?

David Reid