[PD] The old ASIO latency thing

matthew jones m.jones at signal.QinetiQ.com
Mon Sep 20 10:02:21 CEST 2004


Is there any reason why the minimum latency I get with Pd is more than twice
the minimum latency I get with maxMSP?
460 samples - Pd
216 samples - MaxMSP

I'm trying the demo of maxMSP on winXP for the first time, and writing the
same patch in both programs which simply records the [noise~] input to [dac~
1] and the return path from [adc~ 1] to wavfiles (from which I then do
interpolated cross-correlation and peak detection in matlab), the numbers I
get out are astounding.

I use the M-audio 1010 which has a 'latency' control in the 'control panel'
software of the driver.  Playing around with Pd's 'audiobuf' values seems to
have a compound effect with this 'latency' value, where unless you select
'64 samples' you cannot get anywhere near good latency.  This is the only
thing that effects latency in maxMSP, while in Pd I have the additional
parameter 'audiobuf' to play with, although even putting this to 1 or 2
still gives latency values far removed from the ones I get with max.....!
Is there something going on here I don't know about?  Just trying to work
out if/why Pd is underperforming.....thanks.



More information about the Pd-list mailing list