I think you'll have to consider the difference between popularizing a
scheme to get a "reasonable sounding" mix, probably by presuming certain
self-correlation and sync issues in your input signals, and proper
information theoretical theorems.
I mean taking a random number of channels with ac
On Mon, Dec 10, 2012 at 01:39:48PM +1100, Ross Bencina wrote:
> avoid any loss of precision due to truncation... etc. There is also
> arbitrary precision arithmetic if you don't want to throw any bits
> away.
This seemed most pertainent to Alessandro's requirement that N was unknown
and might be
Just wanted to say thanks to everybody who answered. I know the problem could
seem trivial, but since I need to try a lot of preprocessing setups (noise
gating, eq'ing, and some speech recognition too..) for this mix to sound
nicely, I wanted to be sure not to take a wrong approach in the very l
