Quote:
Originally Posted by nst7
Thank you Matt.
So if Carbon is 17ms at a buffer of 256, what would it be for 64? Not sure exactly how the math works. Or does Carbon actually go down to 64, or even 32?
|
I've not tested 64. Others have stated 8ms.
The math is simply 128 samples for buffers in (64) and out (64), plus ~20 samples for conversion, plus protocol and driver (which I haven't tried to test). Take whatever that sum is and divide it by the samplerate (48) and you get your latency in ms. For the 8ms to be correct that last number (protocol and driver) needs to be 4-5ms alone.
~5ms would seem to be correct according to what I know of 256 (256+256+20+250=782 divide by 48=16.3ms, so pretty close).