Re: [ecasound] Loops and latency

From: S. Massy <lists@email-addr-hidden>
Date: Sun Apr 15 2012 - 20:33:04 EEST

On Sun, Apr 15, 2012 at 09:53:58AM +0300, Kai Vehmanen wrote:
> Hi,
> On Sat, 14 Apr 2012, S. Massy wrote:
> >>! each hop from a loop output to its input, causes an
> >>! additional delay of one engine iteration (equivalent to
> >To its own input, yes, but what about to another object's input?
> Joel is right, there's always the delay from loop output to its input.
> This follows from the fixed routing logic: 1) read inputs, 2)
> process chains, 3) write to outputs. So when we've reached (3) and
> wrote to a loop object, it'll have to wait until next engine
> iteration to read it in phase (1) again.
Interesting. So this means that the audio output of a signal processed
through two chains with a loop in the middle would be delayed by about
2.7 ms (-b:128 -r:48000) in comparison to a signal processed only in one
chain. The next question is: how can this be compensated for? That much
delay could cause nasty, unwanted side-effects in audio. I guess the
most obvious way to solve the problem would be to delay processing for
audio with shorter paths (through, say, playat).


For Developers, A Lot Can Happen In A Second.
Boundary is the first to Know...and Tell You.
Monitor Your Applications in Ultra-Fine Resolution. Try it FREE!
Ecasound-list mailing list
Received on Sun Apr 15 20:15:03 2012

This archive was generated by hypermail 2.1.8 : Sun Apr 15 2012 - 20:15:04 EEST