Subject: Re: [linux-audio-dev] Measured latency
From: Karl MacMillan (karlmac_AT_peabody.jhu.edu)
Date: Tue Jan 30 2001 - 07:33:56 EET
Hi Dave!
These measurements were made with some software that I had written
(basically a c++ abstraction for soundcard io that works with ALSA or
OSS). My goal was to use the most basic code possible to limit variables.
It was using real time scheduling but nothing fancy other than that - just
a loop that calls poll on the soundcard fd and does the io when poll
returns. If I remember correctly about the latency tests that you were
doing it was time from hitting a note on a midi keyboard to sound coming
out of cmix - is this right? I haven't measured anything dealing with
midi, just audio through. Overall things seem to be getting better, but I
would like to hear any experiences from others as well.
Karl
On Tue, 30 Jan 2001, David J. Topper wrote:
> Hey Karl,
>
> A while ago I did a similar test and ended up with measured latencies of
> around 300-500 samples. I tried mucking with kernel priorities and
> whatnot, but didn't see a big change. I also don't have any kernel
> patches for latency applied. I also mucked with fragments and buffer
> sizes too. No real effect.
>
> What were you using on Computer A? PD? What audio driver? How about
> MIDI? There's latency there too probably. We'll get it eventually ...
> I'm curious about what others have found.
>
> DT
> --
> Technical Director - Virginia Center for Computer Music
> http://www.people.virginia.edu/~djt7p
>
_____________________________________________________
| Karl W. MacMillan |
| Computer Music Department |
| Peabody Institute of the Johns Hopkins University |
| karlmac_AT_peabody.jhu.edu |
| www.peabody.jhu.edu/~karlmac |
-----------------------------------------------------
This archive was generated by hypermail 2b28 : Tue Jan 30 2001 - 17:52:17 EET