Re: Sound Quality metrics for ABS auto products

[Khalid <khalid@xxxxxxxxxxxxxxxxxxx>]

Dear Jim
First thanks for your reply.  Now, I did read many of the postings in this
group.
Most of the members are like yourself, i.e. interested in music.  However,
laws of
nature that apply to music also apply to auto systems.  I have recently browsed
through a paper that studied how sound is generated in a violin.  It came to the
conclusion that breif periods of structural instability is reached in order
to obtain
the required high levels of vibration.  This is exactly the same mechanism that
makes some vehicles squeal when you touch your brake early in the morning
or at a light stop.  I mention this example to invite you to look at other
problems
including mine, may be whatever tool developed for one application may serve
the other.
In engineering we always think of degrees of freedom and spaces to define our
problems.  If there exists a number of sound quality metrics that measure
independent
aspects of sound quality, then there is a larger possibility that a better
definition
of sound quality can be found in the space defined by these metrics.  Then comes
the natural step of trying to project this space on a one diemensional
space, where
a unique mapping can project all these metrics onto your newly invented yard
scale.
To my understanding, unlike spectral analysis, human perception has a memory
element.
That is where I have a difficulty with the spectral analysis tools that I am
used to.
May be with your help and that of others members in this group progress can
be made.
I have literally hundreds of sound recordings for ABS noise, and Booster
noise, that I
can make available for trying out ideas.  I use MATLAB for most of my analysis.
Thanks again for your intesest.

Khalid

>Dear Khalid,
>
>I am interested in sound quality metrics for musical instrument sounds.
>Do we have any overlap?
>
>Jim Beauchamp
>beaucham@ircam.fr
>after May 11th: j-beauch@uiuc.edu
>
>
----------------------------------
Khalid@dialup.francenet.fr