Subject: Re: "LinuRe: [linux-audio-dev] plugin idea
From: Iain Sandoe (iain_AT_sandoe.co.uk)
Date: Tue Mar 06 2001 - 04:29:49 EET
> On the other hand, for external gear, its much harder to discover the
> delay introduced. So presumably some analysis method of some sort
> would be required, preferably using a test tone.
probably needs to be broad-band... e.g. a shaped pulse of some sort.
The delay can then be found by a relatively straight forward
cross-correlation and peak-position detect. Can do the cross-correlation by
efficient process - e.g. FFT/cross-spectrum/FFT.
I'm assuming we don't do this every slice - but once to establish params...
The suggestion is assuming the out-board gear doesn't mangle things so much
that there is no longer any correlation (or multiple similar-sized peaks)...
as you say... if that happens it would get much more interesting...
Iain.
This archive was generated by hypermail 2b28 : Tue Mar 06 2001 - 04:54:13 EET