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

Re: By any other name...

So, when a tree falls in the forest, with a helicopter is flying over your
head, does the tree ...

It seems that the answer to the question of whether the signal is present
or absent when it is replaced by noise will depend on our level of
analysis.  As far as we can trust signal processing methods for 'objective'
judgments, it seems that different signal processing methods may classify a
signal as either present or absent as well.

Also, I recall that the ability to judge the signal as being replaced vs
masked by noise was used to study the strength of restoration by Samuel,
1981.  (Not that there isn't a psychological trick for the investigator
studying this phenomenon as well). On the other hand, knowing under what
conditions the signal (or rather associated percept) is reported as being
present can tell us something about the analysis process, making the
metaphysical aspects of 'objective' reality of the signal somewhat

For example, we did a small experiment (sorry for self promotion) that
seemed to indicated that the same element of an acoustic signal (i.e. white
noise) can be perceived as either hiding fricative noise of speech or as
nonspeech noise depending on whether it is integrated with the speech or
noise components of a mixed signal.  Since it was always the same acoustic
segment (which was not speech per se), the objective reality of it seemed
less relevant than the perceptual one.  (The details of the study can be
found here: http://dx.doi.org/10.1007/s10936-007-9054-y )


Valeriy Shafiro
Communication Disorders and Sciences
Rush University Medical Center
Chicago, IL

office (312) 942 - 3298
lab    (312) 942 - 3316
email: valeriy_shafiro@xxxxxxxx

-----AUDITORY - Research in Auditory Perception <AUDITORY@xxxxxxxxxxxxxxx>
wrote: -----

To: AUDITORY@xxxxxxxxxxxxxxx
From: Bruno Repp <repp@xxxxxxxxxxxxxxxx>
Sent by: AUDITORY - Research in Auditory Perception
Date: 03/22/2007 12:38PM
Subject: Re: By any other name...

Thanks, Dan, Daniel, and Yoshitaka, for your excellent comments. I
agree with you,
of course, yet there is still something that bothers me. If it does
not really matter
whether a signal is present or absent, why do researchers make the effort
put a gap in the signal? Why not just mask a continuous signal instead?
If the masker is strong enough, it should not matter. However, the finding
perceived continuity will seem much LESS SURPRISING when the signal
was actually present than when it was absent. So, the actual introduction
a signal absence seems like a psychological trick on the part of the
This does not apply, of course, in cases like the one described by
Yoshitaka, where
some percept is synthesized out of nothing, as it were.

>On 3/22/07, Bruno Repp <repp@xxxxxxxxxxxxxxxx> wrote:
>>If objective methods cannot prove the absence of the signal,
>>then I would argue that the signal is in fact present. Is an
>>objective proof of signal absence typically presented in studies of
>>the auditory continuity effect?
>I don't think the objective presence or absence is very interesting; there
>is a range of circumstances in which a more optimally configured
>detector might be able to detect the absence of a perceptually restored
>tone (although those circumstances may be surprisingly narrow).
>What is more interesting is that even in genuinely undecidable
>when, as Yokashita puts it, the signal is objectively "either present
>or absent",
>the perceptual system does not report that ambiguity but instead returns a
>confident answer.  Moreover, in the case of continuity, that answer is not
>locally simplest answer (no spectral peaks = no perceived tones), but
>is the "simplest" answer on a much broader scale (continuous tone more
>than tone with a gap synchronized with noise burst).
>Maybe the objection is that *of course* the perceptual system will do the
>reasonable thing of assuming continuity when there is no counter-evidence.
>But the computational implementation of a system that can capture and
>apply this kind of definition of "reasonableness" is much more complex
>a lay person might expect from the auditory system - and a majore
>for those of us interested in modeling perceptual sound analysis.
>>If objective methods cannot prove the absence of the signal,
>>then I would argue that the signal is in fact present.
>This reminds me of the discussion we had a few years ago about the
>WW2 aircrews who could conjure up the illusory experience of listening
>to favorite pieces of music in among the earsplitting drone of the
>engines during long missions.  Since no objective measure can distinguish
>the presence or absence of Beethoven's 5th at 20 dB below the air
>noise in my office, why am I not perceiving it (or only that one, and not
>infinity of other unmeasurably-quiet signals that are also "present")?
>  DAn.

Bruno H. Repp
Haskins Laboratories
300 George Street
New Haven, CT 06511-6624
Tel. (203) 865-6163, ext. 236
Fax (203) 865-8963

NOTE: I am at Rutgers University, Newark, two days each week,
usually Wednesday and Friday, and don't read my
Haskins e-mail on those days. To reach me at Rutgers, send
e-mail to <repp@xxxxxxxxxxxxxxxxxxxxxx>.