[pulseaudio-discuss] Low latency (was: PulseConf report)

[David Henningsson]

On 11/07/2012 09:59 PM, Pierre-Louis Bossart wrote:
>
>> For those who aren't following the planet, thought I'd like you know
>> that I've put up notes from PulseConf up at:
>>
>> http://arunraghavan.net/2012/11/pulseconf-2012-report/
>
> One comment on the low-latency case for desktop gaming with a 16ms
> latency. I imagine this means trouble when sending data to the HDaudio
> driver. With the PulseAudio sink architecture, you need the sink and
> ring buffer to be of equal size (feature, not bug), which means you need
> a ring buffer size of 8ms tops (neglecting the client-server buffer),

The ring buffer is still large (to be able to dynamically change to 
higher latency), but it is being almost empty. But maybe that is not a 
practical difference here.

Not sure what you mean with "sink" and "ring buffer". When mixing, data 
goes from the sink-input / "client-server" buffer into the DMA buffer 
directly.

I know Arun is experimenting with the latency calculations to improve 
low-latency scenarios, so better not go into details as they might be 
about to change :-)

> and events up to 4ms apart.  Has anyone tried the changes we pushed
> recently at the kernel level to properly handle the ring buffer pointer
> and delay? I believe some of the underruns may be due to the ~1ms
> inaccuracy that we had before these changes.  If your driver is already
> giving you a 25% precision error no wonder things are broken?

Right now we have bigger issues, such as why nobody is responding to 
messages such as this one [1] :-(


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

[1] https://lkml.org/lkml/2012/11/5/74