"UnsteadyKen" wrote in message
m
An interesting new approach to signal analysis is
(sketchily) described on the following pages, it appears
to be an attempt to apply FFT techniques to analysis of
music signals.

http://www.quantumqrt.com/submenu.as...3&ContentID=24

Seems to be based on signal subtraction which unfortunately conflates a
number of differt kinds of distortion and in general makes equipment signal
accuracy seem far poorer than it actually is.

For an example, I can take an audio signal and apply some phase shift, and
drastically change its maximum amplitude and also create a large difference
between the input and output. Without knowing a priori how the signal was
changes, what does this apparenly large difference tell me?

Far more to the point would be to calaculate the transfer function of the
device by use of the FFT.

http://www.vertexaq.com/content/view/36/1/

More of the usual signal subtraction BS.

Any comments?

Strictly bush-league stuff.