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

Re: spectral shape + perceptual distance

Dear Andy,

A further investigation was done in the following contribution:

Waeltermann, M., Raake, A., Moeller, S. (2008). "The Sound Character Space of Spectrally Distorted Telephone Speech and its Impact on Quality". In: Proc. 124th AES Convention, 17-20 May, NL-Amsterdam, Paper No. 7464.

There, the perceptual space of linearly distorted telephone speech has been revealed by Multidimensional Scaling (MDS). It has been shown that the two dimensions highly correlate with two parameters that can be extracted from the gain function of the channel. A model for the resulting subjective quality is given.

Significant work in this direction has also been done by Moore and co-workers:

Moore, B. C. J., and Tan, C. T. (2003). "Perceived naturalness of spectrally distorted speech and music," J. Acoust. Soc. Am. 114, 408-419.

Moore, B. C. J., and Tan, C. T. (2004). "Development and validation of a method for predicting the perceived naturalness of sounds subjected to spectral distortion," J. Audio Eng. Soc. 52, 900-914.

They developed sophisticated models for naturalness prediction of spectrally distorted signals.

In all these papers, comprehensive lists of related literature are given.

Best regards,
Nicolas Côté

-----Original Message-----
From: AUDITORY - Research in Auditory Perception [mailto:AUDITORY@xxxxxxxxxxxxxxx] On Behalf Of Bernhard Laback
Sent: jeudi 23 octobre 2008 09:28
To: AUDITORY@xxxxxxxxxxxxxxx
Subject: Re: [AUDITORY] spectral shape + perceptual distance

Hi Andy,

Look at:
Rao P, van Dinther R, Veldhuis R, Kohlrausch A. (2001). "A measure for 
predicting audibility discrimination thresholds for spectral envelope 
distortions in vowel sounds," J Acoust Soc Am. 109, 2085-2097.


Andrew Sabin wrote:
> Hi All,
> Is anyone aware of a paper investigating a way to model the perceptual
> distance between 2 sounds that differ in spectral shape, but are
> otherwise identical? The ideal investigation would compare a single
> recorded sound that has been processed with 2 different equalization
> (frequency vs gain) curves. I am aware of the recent work by Terasawa
> et al., but I am most interested in real (recorded) sounds.
> Many thanks
> Andy Sabin
> Graduate Student
> Northwestern University

Bernhard Laback, Ph.D.
Experimental Audiology Group
Acoustics Research Institute (ARI)
Austrian Academy of Sciences
Wohllebengasse 12-14
A-1040 Wien
Tel.: +43 1 51581 2514
Fax.: +43 1 51581 2530