Subject: [linux-audio-dev] Fwd: [iiwusynth-devel] Iiwusynth on windows, latency measurements
From: Bob Ham (node_AT_users.sourceforge.net)
Date: Mon Jun 17 2002 - 22:11:51 EEST
Interesting benchmarks.. fjear teh lunix :)
-----Forwarded Message-----
> From: M. Nentwig <nentwig_AT_users.sourceforge.net>
> To: peter_AT_hanappe.com, jgreen_AT_users.sourceforge.net, iiwusynth-devel_AT_mail.freesoftware.fsf.org
> Subject: [iiwusynth-devel] Iiwusynth on windows, latency measurements
> Date: 17 Jun 2002 08:09:33 +0300
>
> Hello,
>
> iiwusynth_debug.exe works on Windows 98. Nice job, Peter!
> Now that was a good reason to do long-overdue latency testing:
>
> Midi data was sent from one keyboard into a sound module, MIDI thru
> connected to the MIDI input of iiwusynth.
> Both outputs were recorded, and the time difference measured. Four
> samples were taken per measurement.
>
> Windows, UA-1 USB audio interface: 60..63 ms
> Windows, on-board sound chip, Maestro3: 55 .. 57 ms
> Linux, UA-1 USB interface with kernel default audio.o driver: 30..33 ms
> Same with only one ? buffer (-c 1 option): 12..17 ms
> Linux, UA-100 with customized low-latency driver: 11..13 ms
>
> The Linux and Windows versions of iiwusynth differ in their default
> audio queuesize setting (iiwusynth.h). Is that setting used somewhere?
> Still, I wonder what makes the Linux version four to five times faster
> than its Windows counterpart:
>
> http://www.cakewalk.com/DevXchange/audio_i.htm
> "Unfortunately, this power comes at a price. Due to internal buffering
> KMixer nominally adds 30 msec of
> latency to audio playback streams. (At present, Microsoft does not
> provide a method to allow host
> applications to bypass KMixer.)"
> the same at
> http://linux.oreillynet.com/pub/a/linux/2000/11/17/low_latency.html
>
> Looks like Windows is unable to deliver sufficient performance on
> non-pro (i.e. lacking ASIO driver) hardware. Sequencing should work fine
> (win-iiwusynth is a lot faster than other software synths I've seen so
> far, with latencies up to 500 ms), but real time playing is still
> impossible at the moment.
-- Bob Ham: bob_AT_ham.org http://pkl.net/~node/My music: http://mp3.com/obelisk_uk GNU Hurd: http://hurd.gnu.org/
This archive was generated by hypermail 2b28 : Tue Jun 18 2002 - 14:31:44 EEST