Re: harmonic extraction (David Smith )


Subject: Re: harmonic extraction
From:    David Smith  <smithd@xxxxxxxx>
Date:    Sat, 28 Mar 2009 09:54:26 -0500
List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>

This is a multi-part message in MIME format. --_----------=_1238252066304512 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="windows-1252" Here is an article mentioning intermodulation tones created by the cochlea which may be relevant. http://bmb.oxfordjournals.org/cgi/content/full/63/1/223 Is the effect you're looking at dependent on stimulus amplitude? Dave Smith ----- Original Message ----- From: "Al Bregman" To: AUDITORY@xxxxxxxx Subject: Re: [AUDITORY] harmonic extraction Date: Fri, 27 Mar 2009 19:27:37 -0400 Dear Jim, Imagine your experiment in "reverse". Alternate the all-harmonic tone with an ODD-harmonic inducer. Then if the inducer captures the odd harmonics from the all-harmonic tone, the residual will be the set of even harmonics, and their fundamental is 2f not f. So you ought to hear a continuous tone at frequency f, accompanied by a pulsing tone an octave higher. I am quite sure that this will occur if there is no silence between the tones. In fact I'll turn in my badge if it doesn't come out that way. This method makes it easier to determine whether the residual harmonics are used to make their own tone, because it requires a pitch judgment rather than the more difficult timbre judgment that you reported. If the outcome is as I expect, there will be no need to evoke an explanation in terms of duplex perception for this odd-full-odd-full stimulus pattern. But even so, this wouldn't prove that there had been no duplex perception in the example that YOU reported (even-full-even-full). It may be that the partitioning of the spectrum is more ambiguous with your stimulus, and that this could lead to duplex perception. It is possible that duplex perception is a top-down process that tries to make sense of an ambiguous input. Best, Al ------------------------------------------------------------------- Albert S. Bregman, Emeritus Professor Psychology Department, McGill University 1205 Docteur Penfield Avenue Montreal, QC, Canada H3A 1B1. Office: Phone: (514) 398-6103 Fax: (514) 398-4896 ------------------------------------------------------------------- On Fri, Mar 27, 2009 at 1:08 PM, Pdivenyi wrote: > David, > Your explanation loses it's punch when you realize that it is theidd > harmonics that add up into a square wave. Not mentioning the probably better > explanation by.Blb Carlyon. > Pierre > Sent from my iPhone > On Mar 27, 2009, at 9:15 AM, David Smith wrote: > > > > Please pardon the brevity of my previous post. > > I also noticed, while using an "exciter" effect, which adds harmonics to > a signal that, in the presence of a signal with many even harmonics, > I can't hear the addition of odd harmonics. My knowledge in this area > is not extensive but I can think of a possible explanation: > > Successive in-phase addition of even harmonics produces a square > wave. Successive in-phase addition of odd harmonics yields a > triangle wave. A square wave has much more power than a triangle > wave of the same peak amplitude so you have more signal. > > Absorptive and dispersion losses are greater at higher pressure, > not higher power. So triangle waves, composed of odd harmonics, > lose a greater portion of their power to non-resonant or damped > systems, cavities and membranes with partially absorptive components > or boundaries. Speakers, headphones, ears, and air all contain > partially absorptive components. > > If you lop off the top of a triangle wave you end up with something > closer to a square wave, a signal which sounds like it has been built > with in-phase even harmonics. > > Signals composed of components with partial phase differences > would be more complicated. I have not considered them. > > Dave Smith > > ----- Original Message ----- > From: "Al Bregman" > To: AUDITORY@xxxxxxxx > Subject: Re: [AUDITORY] harmonic extraction > Date: Thu, 26 Mar 2009 11:51:32 -0400 > > > On Thu, Mar 26, 2009 at 9:58 AM, David Smith wrote: > >> Given the stimuli, I think you would be hard pressed to find a mechanical >> system, >> the atmosphere and ear included, which did not exhibit response at 200Hz. > > > Dear David, > > Please expand on this. It's not clear how you intend this to explain > Jim Bashford's observations. > > - Al > ---------------------------------------------------------- > > > > >> ----- Original Message ----- >> From: "Al Bregman" >> To: AUDITORY@xxxxxxxx >> Subject: Re: [AUDITORY] harmonic extraction >> Date: Wed, 25 Mar 2009 12:46:37 -0400 >> >> >> Dear James, >> >> This example of duplex perception, outside the domain of speech, and >> clearly not involving two distinct mental "modules", is very >> interesting, throws a different light on duplex perception of speech, >> and is certainly worth further development and publication.. >> >> Best, >> Al >> ------------------------------------------------------------------- >> Albert S. Bregman, Emeritus Professor >> Psychology Department, McGill University >> 1205 Docteur Penfield Avenue >> Montreal, QC, Canada H3A 1B1. >> Office: Phone: (514) 398-6103 >> Fax: (514) 398-4896 >> ------------------------------------------------------------------- >> >> On Wed, Mar 25, 2009 at 10:35 AM, James Bashford wrote: >>> Dear Xueliang Zhang, >>> >>> I was intrigued by the interchange between you, Yoshitaka Nakajima, >>> and >>> Al Bregman, and did some listening, diotically through headphones, to >>> stimuli consisting of the first 10 harmonics of 100 Hz alternating with >>> the >>> 5 even harmonics of that stimulus (all harmonics were presented at the >>> same >>> level). With on/off times for each complex matched at either 0.5 sec or >>> 1.0 >>> sec (10 ms rise/fall), I hear a fully continous 200-Hz tone that matches >>> the >>> intensity and timbre of the isolated even-harmonic complex. More >>> interesting, however, is that the 10-component =93all-harmonic=94 stimulus, >>> which is heard intermittently, has not only the 100-Hz pitch of that >>> complex >>> but also retains the loudness and timbre of the all-harmonic stimulus as >>> heard when it is presented in isolation. Were the even harmonics of the >>> all-harmonic complex exclusively allocated to support perception of the >>> continuous 200-Hz complex tone -- leaving only the odd-harmonics to >>> support >>> perception of the 100 Hz tone -- we would expect both a reduction in >>> loudness and a clear shift in timbre to the =93hollow=94 quality >>> characteristic >>> of odd-harmonic signals. This suggests that the =93priming=94 or =93capture=94 >>> effect observed with complex tones provides an example of duplex >>> perception >>> that requires neither a verbal stimulus nor dichotic presentation. As I >>> recall, Al Bregman has previously suggested that such an effect might >>> occur >>> when nonverbal stimulus input is strongly ambiguous. >>> >>> This use of the even-harmonic components to support two >>> simultaneous >>> percepts (that of both the intermittent =93all-harmonic=94 tone and the >>> continuous even-harmonic tone) contrasts sharply with the processing >>> underlying the general phenomenon of illusory continuity that is observed >>> when one sound alternates with a higher-intensity, potential masking >>> sound. The latter effect, which has been called auditory induction >>> (Warren, >>> 1972), occurs with a wide variety of signals, such as tones alternating >>> with >>> other tones, noise alternating with higher intensity noise, or speech >>> interrupted by noise (phonemic restoration). This type of continuity, in >>> which there are no exactly matching components to be found between the >>> alternating signals, does appear to involve subtractive or exclusive >>> allocation. For interrupted tones, noise, or speech, continuity is >>> obtained >>> at the expense of the interrupting signal, which is reduced in loudness >>> by >>> an amount proportional to the extent the illusion (Warren et al., 1994). >>> >>> Warren, R. M., Obusek, C. and Ackroff, J. M. (1972). Auditory induction: >>> Perceptual synthesis of absent sounds. Science, 176, 1149-1151. >>> >>> Warren, R. M., Bashford, J. A., Jr., Healy, E. W., and Brubaker, B. S. >>> (1994). Auditory induction: Reciprocal changes in alternating sounds. >>> Perception & Psychophysics, 55, 313-322. >>> >>> >>> >>> James Bashford >>> >>> On Mar 19, 2009, at 5:24 AM, xlzhang wrote: >>> >>> Dear list, >>> A pure tone can extract corresponding harmonic from complex sound when >>> appearing alternatively. I wonder if a harmonic sound can do the same >>> job? >>> For example, a complex sound with F0=3D200Hz appears with a complex sound >>> with F0=3D100 Hz, can we get a continuous perception for F0=3D200Hz? >>> Thank you for your answers in advance. >>> >>> Xueliang Zhang >>> - >> >> -- >> Be Yourself @xxxxxxxx mail.com! >> Choose From 200+ Email Addresses >> Get a Free Account at www.mail.com! > > > -- --=20 Be Yourself @xxxxxxxx mail.com! Choose From 200+ Email Addresses Get a Free Account at www.mail.com --_----------=_1238252066304512 Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Content-Type: text/html; charset="windows-1252" <div> <br>Here is an article mentioning intermodulation <br>tones created by the = cochlea which may be relevant.<br><br>http://bmb.oxfordjournals.org/cgi/con= tent/full/63/1/223<br><br>Is the effect you're looking at&nbsp; dependent o= n stimulus amplitude?<br><br>Dave Smith<br> <br> <blockquote style=3D"border-left: 2px solid rgb(16, 16, 255); margin-left: = 5px; padding-left: 5px;">----- Original Message -----<br> From: "Al Bregman" <al.bregman@xxxxxxxx><br> To: AUDITORY@xxxxxxxx<br> Subject: Re: [AUDITORY] harmonic extraction<br> Date: Fri, 27 Mar 2009 19:27:37 -0400<br> <br> <br> Dear Jim,<br> <br> Imagine your experiment in "reverse". Alternate the all-harmonic tone<br> with an ODD-harmonic inducer. Then if the inducer captures the odd<br> harmonics from the all-harmonic tone, the residual will be the set of<br> even harmonics, and their fundamental is 2f not f. So you ought to<br> hear a continuous tone at frequency f, accompanied by a pulsing tone<br> an octave higher. I am quite sure that this will occur if there is no<br> silence between the tones. In fact I'll turn in my badge if it<br> doesn't come out that way.<br> <br> This method makes it easier to determine whether the residual<br> harmonics are used to make their own tone, because it requires a pitch<br> judgment rather than the more difficult timbre judgment that you<br> reported. If the outcome is as I expect, there will be no need to<br> evoke an explanation in terms of duplex perception for this<br> odd-full-odd-full stimulus pattern.<br> <br> But even so, this wouldn't prove that there had been no duplex<br> perception in the example that YOU reported (even-full-even-full). It<br> may be that the partitioning of the spectrum is more ambiguous with<br> your stimulus, and that this could lead to duplex perception. It is<br> possible that duplex perception is a top-down process that tries to<br> make sense of an ambiguous input.<br> <br> Best,<br> <br> Al<br> -------------------------------------------------------------------<br> Albert S. Bregman, Emeritus Professor<br> Psychology Department, McGill University<br> 1205 Docteur Penfield Avenue<br> Montreal, QC, Canada H3A 1B1.<br> Office: Phone: (514) 398-6103<br> Fax: (514) 398-4896<br> -------------------------------------------------------------------<br> <br> <br> On Fri, Mar 27, 2009 at 1:08 PM, Pdivenyi <pdivenyi@xxxxxxxx> wrote:<br> &gt; David,<br> &gt; Your explanation loses it's punch when you realize that it is theidd<b= r> &gt; harmonics that add up into a square wave. Not mentioning the probably = better<br> &gt; explanation by.Blb Carlyon.<br> &gt; Pierre<br> &gt; Sent from my iPhone<br> &gt; On Mar 27, 2009, at 9:15 AM, David Smith <smithd@xxxxxxxx> wrote:<br> &gt;<br> &gt;<br> &gt;<br> &gt; Please pardon the brevity of my previous post.<br> &gt;<br> &gt; I also noticed, while using an "exciter" effect, which adds harmonics = to<br> &gt; a signal that, in the presence of a signal with many even harmonics,<b= r> &gt; I can't hear the addition of odd harmonics.&nbsp; My knowledge in this= area<br> &gt; is not extensive but I can think of a possible explanation:<br> &gt;<br> &gt; Successive in-phase addition of even harmonics produces a square<br> &gt; wave.&nbsp; Successive in-phase addition of odd harmonics yields a<br> &gt; triangle wave. A square wave has much more power than a triangle<br> &gt; wave of the same peak amplitude so you have more signal.<br> &gt;<br> &gt; Absorptive and dispersion losses are greater at higher pressure,<br> &gt; not higher power. So triangle waves, composed of odd harmonics,<br> &gt; lose a greater portion of their power to non-resonant or damped<br> &gt; systems, cavities and membranes with partially absorptive components<b= r> &gt; or boundaries.&nbsp; Speakers, headphones, ears, and air all contain<b= r> &gt; partially absorptive components.<br> &gt;<br> &gt; If you lop off the top of a triangle wave you end up with something<br> &gt; closer to a square wave, a signal which sounds like it has been built<= br> &gt; with in-phase even harmonics.<br> &gt;<br> &gt; Signals composed of components with partial phase differences<br> &gt; would be more complicated.&nbsp; I have not considered them.<br> &gt;<br> &gt; Dave Smith<br> &gt;<br> &gt; ----- Original Message -----<br> &gt; From: "Al Bregman"<br> &gt; To: AUDITORY@xxxxxxxx<br> &gt; Subject: Re: [AUDITORY] harmonic extraction<br> &gt; Date: Thu, 26 Mar 2009 11:51:32 -0400<br> &gt;<br> &gt;<br> &gt; On Thu, Mar 26, 2009 at 9:58 AM, David Smith wrote:<br> &gt;<br> &gt;&gt; Given the stimuli, I think you would be hard pressed to find a mec= hanical<br> &gt;&gt; system,<br> &gt;&gt; the atmosphere and ear included, which did not exhibit response at= 200Hz.<br> &gt;<br> &gt;<br> &gt; Dear David,<br> &gt;<br> &gt; Please expand on this. It's not clear how you intend this to explain<b= r> &gt; Jim Bashford's observations.<br> &gt;<br> &gt; - Al<br> &gt; ----------------------------------------------------------<br> &gt;<br> &gt;<br> &gt;<br> &gt;<br> &gt;&gt; ----- Original Message -----<br> &gt;&gt; From: "Al Bregman"<br> &gt;&gt; To: AUDITORY@xxxxxxxx<br> &gt;&gt; Subject: Re: [AUDITORY] harmonic extraction<br> &gt;&gt; Date: Wed, 25 Mar 2009 12:46:37 -0400<br> &gt;&gt;<br> &gt;&gt;<br> &gt;&gt; Dear James,<br> &gt;&gt;<br> &gt;&gt; This example of duplex perception, outside the domain of speech, a= nd<br> &gt;&gt; clearly not involving two distinct mental "modules", is very<br> &gt;&gt; interesting, throws a different light on duplex perception of spee= ch,<br> &gt;&gt; and is certainly worth further development and publication..<br> &gt;&gt;<br> &gt;&gt; Best,<br> &gt;&gt; Al<br> &gt;&gt; ------------------------------------------------------------------= -<br> &gt;&gt; Albert S. Bregman, Emeritus Professor<br> &gt;&gt; Psychology Department, McGill University<br> &gt;&gt; 1205 Docteur Penfield Avenue<br> &gt;&gt; Montreal, QC, Canada H3A 1B1.<br> &gt;&gt; Office: Phone: (514) 398-6103<br> &gt;&gt; Fax: (514) 398-4896<br> &gt;&gt; ------------------------------------------------------------------= -<br> &gt;&gt;<br> &gt;&gt; On Wed, Mar 25, 2009 at 10:35 AM, James Bashford wrote:<br> &gt;&gt;&gt; Dear Xueliang Zhang,<br> &gt;&gt;&gt;<br> &gt;&gt;&gt; &nbsp;&nbsp;&nbsp;I was intrigued by the interchange between y= ou, Yoshitaka Nakajima,<br> &gt;&gt;&gt; and<br> &gt;&gt;&gt; Al Bregman, and did some listening, diotically through headpho= nes, to<br> &gt;&gt;&gt; stimuli consisting of the first 10 harmonics of 100 Hz alterna= ting with<br> &gt;&gt;&gt; the<br> &gt;&gt;&gt; 5 even harmonics of that stimulus (all harmonics were presente= d at the<br> &gt;&gt;&gt; same<br> &gt;&gt;&gt; level).&nbsp;&nbsp;With on/off times for each complex matched = at either 0.5 sec or<br> &gt;&gt;&gt; 1.0<br> &gt;&gt;&gt; sec (10 ms rise/fall), I hear a fully continous 200-Hz tone th= at matches<br> &gt;&gt;&gt; the<br> &gt;&gt;&gt; intensity and timbre of the isolated even-harmonic complex.&nb= sp;&nbsp;More<br> &gt;&gt;&gt; interesting, however, is that the 10-component =93all-harmonic= =94 stimulus,<br> &gt;&gt;&gt; which is heard intermittently, has not only the 100-Hz pitch o= f that<br> &gt;&gt;&gt; complex<br> &gt;&gt;&gt; but also retains the loudness and timbre of the all-harmonic s= timulus as<br> &gt;&gt;&gt; heard when it is presented in isolation.&nbsp;&nbsp;Were the e= ven harmonics of the<br> &gt;&gt;&gt; all-harmonic complex exclusively allocated to support percepti= on of the<br> &gt;&gt;&gt; continuous 200-Hz complex tone -- leaving only the odd-harmoni= cs to<br> &gt;&gt;&gt; support<br> &gt;&gt;&gt; perception of the 100 Hz tone -- we would expect both a reduct= ion in<br> &gt;&gt;&gt; loudness and a clear shift in timbre to the =93hollow=94 quali= ty<br> &gt;&gt;&gt; characteristic<br> &gt;&gt;&gt; of odd-harmonic signals.&nbsp;&nbsp;This suggests that the =93= priming=94 or =93capture=94<br> &gt;&gt;&gt; effect observed with complex tones provides an example of dupl= ex<br> &gt;&gt;&gt; perception<br> &gt;&gt;&gt; that requires neither a verbal stimulus nor dichotic presentat= ion.&nbsp;&nbsp;As I<br> &gt;&gt;&gt; recall, Al Bregman has previously suggested that such an effec= t might<br> &gt;&gt;&gt; occur<br> &gt;&gt;&gt; when nonverbal stimulus input is strongly ambiguous.<br> &gt;&gt;&gt;<br> &gt;&gt;&gt; &nbsp;&nbsp;&nbsp; &nbsp;&nbsp;This use of the even-harmonic c= omponents to support two<br> &gt;&gt;&gt; simultaneous<br> &gt;&gt;&gt; percepts (that of both the intermittent =93all-harmonic=94 ton= e and the<br> &gt;&gt;&gt; continuous even-harmonic tone) contrasts sharply with the proc= essing<br> &gt;&gt;&gt; underlying the general phenomenon of illusory continuity that = is observed<br> &gt;&gt;&gt; when one sound alternates with a higher-intensity, potential m= asking<br> &gt;&gt;&gt; sound.&nbsp;&nbsp;The latter effect, which has been called aud= itory induction<br> &gt;&gt;&gt; (Warren,<br> &gt;&gt;&gt; 1972), occurs with a wide variety of signals, such as tones al= ternating<br> &gt;&gt;&gt; with<br> &gt;&gt;&gt; other tones, noise alternating with higher intensity noise, or= speech<br> &gt;&gt;&gt; interrupted by noise (phonemic restoration).&nbsp;&nbsp;This t= ype of continuity, in<br> &gt;&gt;&gt; which there are no exactly matching components to be found bet= ween the<br> &gt;&gt;&gt; alternating signals, does appear to involve subtractive or exc= lusive<br> &gt;&gt;&gt; allocation.&nbsp;&nbsp;For interrupted tones, noise, or speech= , continuity is<br> &gt;&gt;&gt; obtained<br> &gt;&gt;&gt; at the expense of the interrupting signal, which is reduced in= loudness<br> &gt;&gt;&gt; by<br> &gt;&gt;&gt; an amount proportional to the extent the illusion (Warren et a= l., 1994).<br> &gt;&gt;&gt;<br> &gt;&gt;&gt; Warren, R. M., Obusek, C. and Ackroff, J. M. (1972). Auditory = induction:<br> &gt;&gt;&gt; Perceptual synthesis of absent sounds.&nbsp;Science, 176, 1149= -1151.<br> &gt;&gt;&gt;<br> &gt;&gt;&gt; Warren, R. M., Bashford, J. A., Jr., Healy, E. W., and Brubake= r, B. S.<br> &gt;&gt;&gt; (1994). Auditory induction: Reciprocal changes in alternating = sounds.<br> &gt;&gt;&gt; Perception &amp; Psychophysics, 55, 313-322.<br> &gt;&gt;&gt;<br> &gt;&gt;&gt;<br> &gt;&gt;&gt;<br> &gt;&gt;&gt; James Bashford<br> &gt;&gt;&gt;<br> &gt;&gt;&gt; On Mar 19, 2009, at 5:24 AM, xlzhang wrote:<br> &gt;&gt;&gt;<br> &gt;&gt;&gt; Dear list,<br> &gt;&gt;&gt; A pure tone can extract corresponding harmonic from complex so= und when<br> &gt;&gt;&gt; appearing alternatively. I wonder if a harmonic sound can do t= he same<br> &gt;&gt;&gt; job?<br> &gt;&gt;&gt; For example, a complex sound with F0=3D200Hz appears with a co= mplex sound<br> &gt;&gt;&gt; with F0=3D100 Hz, can we get a continuous perception for F0=3D= 200Hz?<br> &gt;&gt;&gt; Thank you for your answers in advance.<br> &gt;&gt;&gt;<br> &gt;&gt;&gt; Xueliang Zhang<br> &gt;&gt;&gt;<br> -<br> &gt;&gt;<br> &gt;&gt; --<br> &gt;&gt; Be Yourself @xxxxxxxx mail.com!<br> &gt;&gt; Choose From 200+ Email Addresses<br> &gt;&gt; Get a Free Account at www.mail.com!<br> &gt;<br> &gt;<br> &gt;<br> <br> <br> <br> --<br> </smithd@xxxxxxxx></pdivenyi@xxxxxxxx></al.bregman@xxxxxxxx></blockquote> </div> <BR> --=20 <div> Be Yourself @xxxxxxxx mail.com!<br> Choose From 200+ Email Addresses<br> Get a <b>Free</b> Account at <a href=3D"http://www.mail.com/Product.aspx" t= arget=3D"_blank">www.mail.com</a>!</div> --_----------=_1238252066304512--


This message came from the mail archive
http://www.auditory.org/postings/2009/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University