Re: [Jack-Devel] latency: reducing buffer allocation

On 03/18/2014 06:52 PM, Leonardo Gabrielli wrote:
>               If  so, JACK will use the smallest possible buffer
> containing at
>               least --nperiods, but the playback latency does not
> increase."
> 
> Also: what do you mean exactly by the latency does not increase? 

I'm not sure about JACK, but PulseAudio often uses the strategy to
allocate a big buffer, but not to fill it up completely. E g, if you
have a buffer of say 400 ms with 4 periods, you can choose to fill that
buffer completely up and then start playback, or just write 200 ms and
start playback. (And in either case, write 100 ms more every time you
get a period interrupt.) In the latter case, the latency will be lower,
even though the hardware buffer size is equal.



-- 
David Henningsson, Canonical Ltd.
https://launchpad.net/~diwic

1395394527.3340_0.ltw:2,a <532C07D7.7090903 at canonical dot com>