[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: MDS distances

Dear Malcolm and list,
I think the main problem in the classical MDS approach to the study of
timbre is that dimensions were interpreted qualitatively, and not
quantitatively. That's why, as you point out, it might be hard to understand what their MDS solutions mean.

A quantitative studies of MDS dimensions are found at least starting from McAdams et al. (1995). This is a good approach, unless one forgets that it's of an exploratory nature.

The confirmatory approach to the study of timbre dimensions has been instead pursued at least by Caclin et al. (2005).

A good approach to the study of psychological domains of unknown
dimensionality, as is the case of timbre, is thus to alternate between
exploratory and confirmatory studies: before if timbre is indeed a function
of perceptual dimensions (acoustical features) A and B (confirmatory
approach) you might actually be willing to find which, among perceptual
dimensions/acoustical features {A,B,C,D} explain your data (exploratory


McAdams S, Winsberg S, Donnadieu S, De Soete G, Krimphoff J.  Perceptual
scaling of synthesized musical timbres: common dimensions, specificities,
and latent subject classes.Psychol Res. 1995;58(3):177-92.

Caclin A, McAdams S, Smith BK, Winsberg S. Acoustic correlates of timbre
space dimensions: a confirmatory study using synthetic tones. J Acoust Soc
Am. 2005 Jul;118(1):471-82.

----- Original Message ----- From: "Malcolm Slaney" <malcolm@xxxxxxxx>
To: <AUDITORY@xxxxxxxxxxxxxxx>
Sent: Thursday, June 22, 2006 3:36 PM
Subject: Re: MDS distances

On Jun 21, 2006, at 2:54 AM, Olivier Tache wrote:

I have read a number of "classical" papers about MDS and auditory
dissimilarity (by Gordon&Grey, Grey&Moorer, Wessel) (and was  wondering
if such experiments were still carried out).

I think the Gray/Wessel approach has failed.. it's too hard to figure  out
what the results mean.  (Just trying to be blunt to get your  attention.
;-)  You start with convenient sounds, measure perception  and then try to
figure out what the MDS dimensions mean.  That hasn't  worked.  I think
that is why people have not been pushing on it very  hard lately.

Hiroko Terasawa and I have been taking an opposite approach.  We're
*starting* with the dimensions, synthesizing sounds and then  measuring
the stress between human perception and the pre-ordained  model.  Several
papers describing our initial results are online at
Sounds like Jim is doing something in between the two extremes.

- Malcolm