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

Re: Definition of masking

"Richard J. Fabbri" wrote:
> Chris,
> >The problem then is that to call something "masking" we need to be sure
> >of the underlying process.
>         ...Yes.
>         ... I have worked on masking along a new physiologic route for
>         several years and am convinced it is an interactive consequence
>         occuring DURING the interaction.
> >For example, we shouldn't call the increase in threshold for a tone
> >following a masker "forward masking" unless we can be sure that
> >the process is one of swamping rather than, say, adaptation.
>         ... Yes.
>         ... Masking is a "swamping" interaction and, as such, requires
>         all signal elements to be present during the same time period.

Here is my take on all this:

I do not like the term "swamping." This sounds like I am in a swamp;-).

"Obscure" is the first term used, and a better term. The obscuring
is caused by the stochastic representation of the signal. If it is not
such noise that limits the detection, then it is not masking. Turning off
the radio, or reducing the volume, is not masking, as it is not stochastic.

An important question is: What is the physical mechanism of the masking
noise. Is it external noise (i.e., due to roving the signal), or internal
noise (i.e., neural noise).

In a recent paper (JASA, Dec. 1997, pp 3628-3646), we argue that under
many conditions the noise is internal, and due to neural noise. At least
it is "Poisson like" since the variance of the signal is equal to the signal
mean. In the case of signals like tones, that are not stochastic, the
ONLY noise is internal noise. When working with wideband noise signals, it
is less clear where the noise comes from. Is it due to the internal representation
(yes, we say), or is it due to the stochastic nature of the stimulus?
This would apply for the JND of wideband noise, for example (Miller, 1947)

Finally, it does not require the masker to be present at the same time.
If the noise is internal, and the variance is equal to the mean, then it
depends on how large the internal signal is, and that will determine the
masking, assuming a threshold at d'=1. If the internal signal decays
instantly after the external signal is removed, then one might believe
that the masker must be present at the same time. We believe we showed
that the masking is due to "loudness noise," namely the internal representation
noise of the signal.

>         ... An effect succeeding a stimulus (not during the stimulus)
>         reveals a different process and may, in fact, be an adaptation,
>         possibly a consequence of middle ear (AGC?) time constants:
>         I agree with this masking qualification.
> Regards,
> Rich


Jont B. Allen
AT&T Labs-Research, Shannon Laboratory
180 Park Ave., Room E161, Florham Park NJ, 07932-0971
973/360-8545voice, x7111fax, http://www.research.att.com/info/jba

 -My favorite URL: www-history.mcs.st-and.ac.uk/history

 -Imagination is more important than knowledge.  --Albert Einstein