[LAU] Audio Interface for live Processing

Saul Rayson saul at atonic.com
Mon Jan 28 22:24:52 UTC 2013


Hi James/Len,

>Did you try 48k / 3 periods/buffer?

Just got it down to 4ms on the internal card!

often I'm processing a live and recorded input, which is mixed back with
the live performance, so latency is of some importance, but the numbers you
are both getting are very workable.

Thanks again for all the info.

/S




On 27 January 2013 23:26, James Stone <jamesmstone at gmail.com> wrote:

> On Sun, Jan 27, 2013 at 10:32 PM, Len Ovens <len at ovenwerks.net> wrote:
> >
> > On Sun, January 27, 2013 12:20 pm, James Stone wrote:
> >> On Sun, Jan 27, 2013 at 4:05 PM, Saul Rayson <saul at atonic.com> wrote:
> >
> >>>>For live work, Guitar effects, soft synths, etc. USB 1.1 at 16bit/48k
> >>>> may
> >>>>work just fine
> >>>
> >>> My experience with USB audio is limited, but USB 1 on mac, for live
> >>> processing, I have found to have unusable latency.
> >>>
> >>
> >> Latency in 1.1 vs usb 2 shouldn't be significantly different as I
> >> understand it. usb 2 just allows higher bandwidth - so more channels
> >> or 24bit vs 16bit. Just depends on whether the interface is properly
> >> supported in linux.
> >
> > I have used a USB 1.1 IF on my netbook running at 800Mhz (half speed) and
> > was still able to run the IF at -p 64 -n 2 (2.67 ms Latency) without
> > xruns. I did have to back the latency off if I ran too many softsynths or
> > too many effects, but even double that is not bad. -p 64 with USB 2
> > devices seems to be the lowest too. I run into memory problems before CPU
> > use problems in general with only 1G memory :)
>
> .. actually on the Scarlett 2i4, I can start jack with -p 16 -n 2..
> but it's not really usable on my ageing PC (Athlon 2500XP 1G RAM)..
> lots of xruns.. It will run xrun free at 48k with -p32. push it up to
> -p64, and it is pretty solidly stable at 96000 sample rate. Note 2.67
> is round trip latency - from sound in to sound out. For midi control
> of softsynths, it is half that.
>
> With very short sample sizes, linuxsampler complains that the audio
> fragment size is too small though! I think linuxsampler (at least the
> version shipped with ubuntu) requires a period of 128 to run
> correctly..
>
> James
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxaudio.org/pipermail/linux-audio-user/attachments/20130128/026e784a/attachment-0001.html>


More information about the Linux-audio-user mailing list