26 May
2005
26 May
'05
10:48 a.m.
On Thu, 2005-05-19 at 14:44 +0200, Tim Goetze wrote:
[Jens M Andreasen]
Sorry for the delay (no pun intended!)
Jitter accumulates. So iff we say one ms(==USB) is clearly acceptable,
then every 0.1 ms above that value is drifting us away from the
acceptable range, no?
...
...
Moving from 64 to 32 samples gave me a penalty of about 10% efficiency.
The next step, going to 16 samples, raised the bar with 30%. So this is
where my application reached the intractable memory barrier. (And yes, I
have tried to manually preload, but Gcc/Intel Hardware is ahead of me
and already done that)
A "synthesist" (that's my
scenario), who can't hear *anything* before
the end of the pipe, would be very sensitive to jitter (the deviation,
or "sloppyness", between triggers and actual sound) and would therefore
require the smallest of small buffers.
Trading jitter for constant latency can work well for some, too.
Unfortunately the increasingly popular MIDI over USB connections
already introduce a millisecond of jitter, so if you're already
working at a blocksize of 64 at 44.1 or faster it's not worth the
effort.
Cheers, Tim
--