Re: AUDITORY Digest - 30 Aug 2006 to 31 Aug 2006 (#2006-197) (Ajith Kumar U )


Subject: Re: AUDITORY Digest - 30 Aug 2006 to 31 Aug 2006 (#2006-197)
From:    Ajith Kumar U  <ajithkumar18@xxxxxxxx>
Date:    Fri, 1 Sep 2006 09:14:53 -0500

------=_Part_72126_28863797.1157120093958 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Dear Dr.Sarah Hargus Ferguson, Probe tone levels depends on the manufacturer and frequency used. 226 Hz its85 dB SPL for most the manufacturer (GSI, Madsen). Additional information can be found in Silman and Silverman, Auditory diagnosis, ANSI std, ISO std. Ajith On 8/31/06, AUDITORY automatic digest system <LISTSERV@xxxxxxxx> wrote: > > There are 7 messages totalling 424 lines in this issue. > > Topics of the day: > > 1. Cochlear liquid-particle trajectories. > 2. tympanometry probe tone level? (4) > 3. STFT vs Power Spectral in Musical recognition system ? (2) > > ---------------------------------------------------------------------- > > Date: Thu, 31 Aug 2006 10:37:03 +0000 > From: "reinifrosch@xxxxxxxx" <reinifrosch@xxxxxxxx> > Subject: Cochlear liquid-particle trajectories. > > Dear List, > > The calculations on the "new" basilar-membrane stiffness > formula take longer than expected. If they give interesting > results, I shall communicate them in a week or two. > > Yesterday I received a cochlear-mechanics report on a > transmission-line model, containing a graph of liquid > streamlines which disagree with what I have learned. > > In the case of small-displacement surface waves on > the ocean it is helpful to change to a coordinate system > which moves along with the wave crests: > > x' = x - c * t > > (where x = longitudinal coordinate, t = time, c = phase > velocity). In the primed coordinate system, the liquid > particles make a stationary flow in the -x'-direction. > At the crests, the streamlines are far apart from each other, > so that the velocity v_x' of the particles is comparatively low. > > Transformation back to the lab system yields closed, > elliptical liquid-particle trajectories (circular for short > waves). > In long waves (e.g., tsunamis far from the coast) the particle > trajectories are oblong ellipses; at the ocean floor, the length > of the short ellipse axis is zero. > > In cochlear waves, the phase velocity c decreases if x > increases; the liquid-particle trajectories, however, are > similar (I believe) to those in ocean waves, i.e., closed > and approximately elliptical. > > The elliptical particle trajectories do not cross the basilar > membrane (BM). At the BM, the liquid particles move > along the BM. That is why the mass of the organ-of-Corti > cells (which in the idealistic models do not move along > the BM) plays a role even if their density is equal to > that of the liquid. > > A good introduction into cochlear waves was written > by the late G.K. Yates: Chapter 2 of the book "Hearing", > B.C.J. Moore, ed., Academic Press, San Diego, > 1995, Section II.B, pages 49-53. > > With best wishes, > > Reinhart Frosch. > > > Reinhart Frosch, > Dr. phil. nat., > r. PSI and ETH Zurich, > Sommerhaldenstr. 5B, > CH-5200 Brugg. > Phone: 0041 56 441 77 72. > Mobile: 0041 79 754 30 32. > E-mail: reinifrosch@xxxxxxxx . > > ------------------------------ > > Date: Thu, 31 Aug 2006 10:08:03 -0500 > From: "Ferguson, Sarah Hargus" <safergus@xxxxxxxx> > Subject: tympanometry probe tone level? > > Hello list! Students ask the darnedest questions. Today in Intro to > Audiology, one of the students asked what the presentation level is for > the probe tone used in tympanometry. I didn't find it in a quick Google > search (not even a manufacturers' specs I looked at) or during a quick > peek in a classic audiology textbook - perhaps someone out there knows > off-hand what it is?=20 > > ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~=20 > Sarah Hargus Ferguson, Ph.D., CCC-A > Assistant Professor > Department of Speech-Language-Hearing: Sciences and Disorders=20 > University of Kansas=20 > Dole Center=20 > 1000 Sunnyside Ave., Room 3001=20 > Lawrence, KS 66045 > office: (785)864-1116 > Speech Acoustics and Perception Lab: (785)864-0610=20 > http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html > > ------------------------------ > > Date: Thu, 31 Aug 2006 11:45:04 EDT > From: Harriet B Jacobster <Hjacobster@xxxxxxxx> > Subject: Re: tympanometry probe tone level? > > -------------------------------1157039103 > Content-Type: text/plain; charset="US-ASCII" > Content-Transfer-Encoding: 7bit > > It's around 85 dBSPL > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > Harriet Jacobster, Au.D., CCC-A, FAAA > Board Certified in Audiology > > -------------------------------1157039103 > Content-Type: text/html; charset="US-ASCII" > Content-Transfer-Encoding: quoted-printable > > <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> > <HTML><HEAD> > <META http-equiv=3DContent-Type content=3D"text/html; charset=3DUS-ASCII"> > <META content=3D"MSHTML 6.00.2900.2963" name=3DGENERATOR></HEAD> > <BODY id=3Drole_body style=3D"FONT-SIZE: 10pt; COLOR: #000000; > FONT-FAMILY:=20= > Arial"=20 > bottomMargin=3D7 leftMargin=3D7 topMargin=3D7 rightMargin=3D7><FONT > id=3Drol= > e_document=20 > face=3DArial color=3D#000000 size=3D2> > <DIV>It's around 85 dBSPL</DIV> > <DIV>&nbsp;</DIV> > <DIV>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~</DIV> > <DIV>Harriet Jacobster, Au.D., CCC-A, FAAA</DIV> > <DIV>Board Certified in Audiology</DIV></FONT></BODY></HTML> > > -------------------------------1157039103-- > > ------------------------------ > > Date: Thu, 31 Aug 2006 08:47:42 -0700 > From: Navid Shahnaz <nshahnaz@xxxxxxxx> > Subject: Re: tympanometry probe tone level? > > Hi > It depends on the probe tone frequency and it is usually set at a level > bel= > ow the level the elicit the aural reflex. It varies a bit between > manufactu= > rer; however, it is close the the following values. > 226 Hz @xxxxxxxx 85dB SPL =B1 1.5dB > 1000 Hz (a probe tone frequency of choice for infant assessment) @xxxxxxxx 75dB > SP= > L =B1 1.5dB > Hope this helps. > Best > Navid > ------------------------------- > Navid Shahnaz, Ph.D > Assistant Professor > School of Audiology & Speech Sciences > Faculty of Medicine > 5804 Fairview Ave. > Vancouver, B.C. V6t 1Z3 > Canada > Tel. 604-822-5953 > Fax. 604-822-6569 > E-mail: nshahnaz@xxxxxxxx > Website: http://www.audiospeech.ubc.ca/school/faculty/navid/ > > ---------- Original Message ---------------------------------- > From: "Ferguson, Sarah Hargus" <safergus@xxxxxxxx> > Reply-To: "Ferguson, Sarah Hargus" <safergus@xxxxxxxx> > Date: Thu, 31 Aug 2006 10:08:03 -0500 > > >Hello list! Students ask the darnedest questions. Today in Intro to > >Audiology, one of the students asked what the presentation level is for > >the probe tone used in tympanometry. I didn't find it in a quick Google > >search (not even a manufacturers' specs I looked at) or during a quick > >peek in a classic audiology textbook - perhaps someone out there knows > >off-hand what it is? > > > >~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ > >Sarah Hargus Ferguson, Ph.D., CCC-A > >Assistant Professor > >Department of Speech-Language-Hearing: Sciences and Disorders > >University of Kansas > >Dole Center > >1000 Sunnyside Ave., Room 3001 > >Lawrence, KS 66045 > >office: (785)864-1116 > >Speech Acoustics and Perception Lab: (785)864-0610 > >http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html > > > > > > > > ________________________________________________________________ > Sent via the WebMail system at audiospeech.ubc.ca > > > > > > ------------------------------ > > Date: Thu, 31 Aug 2006 11:20:43 -0500 > From: Jeremy Federman <jeremy.federman@xxxxxxxx> > Subject: Re: tympanometry probe tone level? > > Hi - > > I suspect it depends on the frequency being used (and maybe the equipment, > too), but I believe the probe tone level at the standard 226 Hz should be > 85 > dB SPL plus or minus 1.5 dB. For additional calibration info, you could > check the ANSI standard... > > With kind regards, > > Jeremy > > On 8/31/06 10:08 AM, "Ferguson, Sarah Hargus" <safergus@xxxxxxxx> wrote: > > > Hello list! Students ask the darnedest questions. Today in Intro to > > Audiology, one of the students asked what the presentation level is for > > the probe tone used in tympanometry. I didn't find it in a quick Google > > search (not even a manufacturers' specs I looked at) or during a quick > > peek in a classic audiology textbook - perhaps someone out there knows > > off-hand what it is? > > > > ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ > > Sarah Hargus Ferguson, Ph.D., CCC-A > > Assistant Professor > > Department of Speech-Language-Hearing: Sciences and Disorders > > University of Kansas > > Dole Center > > 1000 Sunnyside Ave., Room 3001 > > Lawrence, KS 66045 > > office: (785)864-1116 > > Speech Acoustics and Perception Lab: (785)864-0610 > > http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html > > > > -- > Jeremy Federman, MS, CCC-A > Vanderbilt University Medical Center > Dan Maddox Hearing Aid Research Lab > Department of Hearing and Speech Sciences > Medical Center East, South Tower > 1215 21st Ave. South, Room 8310 > Nashville, TN 37232-8242 > jeremy.federman@xxxxxxxx > > ------------------------------ > > Date: Thu, 31 Aug 2006 18:32:56 -0400 > From: Arturo Camacho <acamacho@xxxxxxxx> > Subject: Re: STFT vs Power Spectral in Musical recognition system ? > > One problem of the square-root compression is that its slope > approaches infinity as the magnitude M approaches zero. A more > appropriate approach may be to use log(1+KM), where K is a constant to > be determined. The response of this function is almost logarithmic for > high magnitudes and almost linear for low magnitudes. Of course, the > determination of the optimal value for K given an input is not > trivial. > > Arturo > -- > __________________________________________________ > > Arturo Camacho > PhD Candidate > Computer and Information Science and Engineering > University of Florida > > E-mail: acamacho@xxxxxxxx > Web page: www.cise.ufl.edu/~acamacho > __________________________________________________ > > On Fri, 25 Aug 2006, Richard F. Lyon wrote: > > > Edwin, > > > > A power spectral density is only defined for stationary signals, not > > music. The STFT generalizes it to short segments, if you use the > > squared magnitude. > > > > The difference between the absolute value, square, log, etc. are just > > point nonlinearities that do not change the information content, but > > do change the metric structure of the space a bit. Log is too > > compressed, leading to too much emphasis on near-silent segments, > > while the square (the power you ask about) is too expanded, leading > > to too much emphasis on the louder parts. A good compromise is > > around a square root or cube root of magnitude (roughly matching > > perceptual magnitude via Stevens's law), but the magnitude itself is > > also sometimes acceptable, depending on what you're doing. > > > > Dick > > > > At 7:12 AM -0700 8/25/06, Edwin Sianturi wrote: > > >Content-Type: text/html > > >X-MIME-Autoconverted: from 8bit to quoted-printable by > > >torrent.cc.mcgill.ca id k7PED6jh031610 > > > > > >Hello, > > > > > >I am just a master student, doing my internship. Right now, I am > > >building a musical instrument recognition system. I have read > > >several papers on it, and I am just curious: > > > > > >All the papers/journals that I have read use the STFT, a.k.a the > > >|X(t,f)| of a signal x(t), in order to extract several (spectral) > > >features to be used as the input to the recognition system. > > > > > >What are the reasons behind using the |X(t,f)| instead of using the > > >"power spectral" |X(t,f)|^2 ? > > >(technically speaking, a power spectral density is the expectation > > >of |X(f)|^2, i.e. E(|X(f)|^2) ) > > > > > >Thanks in advance, > > > > > >Edwin SIANTURI > > > > > > > > > ------------------------------ > > Date: Thu, 31 Aug 2006 16:00:31 -0700 > From: "Richard F. Lyon" <DickLyon@xxxxxxxx> > Subject: Re: STFT vs Power Spectral in Musical recognition system ? > > Arturo, I totally agree with the idea of using log(1 + KM), or log(M > + epsilon) as I usually do it. This kind of nonlinearity is > especially important in systems with an imprecisely known zero level > or a variable noise floor. On the other hand, a power law, though it > has an infinite slope at 0, is not half as bad as a plain log, and > lots of people use that anyway. A stabilized power law, like (M + > epsilon)^(1/3) is another good choice, probably more in line with > perception that letting it go log-like at high magnitudes. With any > of these, adjusting your parameters to accommodate a realistic range > of input signal levels becomes important; you can no longer ignore > scale factors and hope for algorithms to work fine on inputs varying > over many orders of magnitude in scale. > > Dick > > At 6:32 PM -0400 8/31/06, Arturo Camacho wrote: > >One problem of the square-root compression is that its slope > >approaches infinity as the magnitude M approaches zero. A more > >appropriate approach may be to use log(1+KM), where K is a constant to > >be determined. The response of this function is almost logarithmic for > >high magnitudes and almost linear for low magnitudes. Of course, the > >determination of the optimal value for K given an input is not > >trivial. > > > >Arturo > >-- > >__________________________________________________ > > > > Arturo Camacho > > PhD Candidate > > Computer and Information Science and Engineering > > University of Florida > > > > E-mail: acamacho@xxxxxxxx > > Web page: www.cise.ufl.edu/~acamacho > >__________________________________________________ > > > >On Fri, 25 Aug 2006, Richard F. Lyon wrote: > > > >> Edwin, > >> > >> A power spectral density is only defined for stationary signals, not > >> music. The STFT generalizes it to short segments, if you use the > >> squared magnitude. > >> > >> The difference between the absolute value, square, log, etc. are just > >> point nonlinearities that do not change the information content, but > >> do change the metric structure of the space a bit. Log is too > >> compressed, leading to too much emphasis on near-silent segments, > >> while the square (the power you ask about) is too expanded, leading > >> to too much emphasis on the louder parts. A good compromise is > >> around a square root or cube root of magnitude (roughly matching > >> perceptual magnitude via Stevens's law), but the magnitude itself is > >> also sometimes acceptable, depending on what you're doing. > >> > >> Dick > >> > >> At 7:12 AM -0700 8/25/06, Edwin Sianturi wrote: > >> >Content-Type: text/html > >> >X-MIME-Autoconverted: from 8bit to quoted-printable by > >> >torrent.cc.mcgill.ca id k7PED6jh031610 > >> > > >> >Hello, > >> > > >> >I am just a master student, doing my internship. Right now, I am > >> >building a musical instrument recognition system. I have read > >> >several papers on it, and I am just curious: > >> > > >> >All the papers/journals that I have read use the STFT, a.k.a the > >> >|X(t,f)| of a signal x(t), in order to extract several (spectral) > >> >features to be used as the input to the recognition system. > >> > > >> >What are the reasons behind using the |X(t,f)| instead of using the > >> >"power spectral" |X(t,f)|^2 ? > >> >(technically speaking, a power spectral density is the expectation > >> >of |X(f)|^2, i.e. E(|X(f)|^2) ) > >> > > >> >Thanks in advance, > >> > > >> >Edwin SIANTURI > >> > > >> > >> > > ------------------------------ > > End of AUDITORY Digest - 30 Aug 2006 to 31 Aug 2006 (#2006-197) > *************************************************************** > > -- ----------------------------------------------------------- Stand on the shoulders of giants Ajith Kumar Uppunda Post doctoral fellow, Department of communication sciences and disorders, Speech research laboratory, Northwestern University, Evanston, IL, 60208. Ph: 847 491 2430 Email a-uppunda@xxxxxxxx ------=_Part_72126_28863797.1157120093958 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline <div>Dear Dr.Sarah Hargus Ferguson,</div> <div>Probe tone levels depends on the manufacturer and frequency used. 226 Hz its85 dB SPL for most the manufacturer (GSI, Madsen). Additional information can be found in Silman and Silverman, Auditory diagnosis, ANSI std, ISO std. </div> <div>Ajith<br><br>&nbsp;</div> <div><span class="gmail_quote">On 8/31/06, <b class="gmail_sendername">AUDITORY automatic digest system</b> &lt;<a href="mailto:LISTSERV@xxxxxxxx">LISTSERV@xxxxxxxx</a>&gt; wrote:</span> <blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">There are 7 messages totalling 424 lines in this issue.<br><br>Topics of the day:<br><br>1. Cochlear liquid-particle trajectories. <br>2. tympanometry probe tone level? (4)<br>3. STFT vs Power Spectral in Musical recognition system ? (2)<br><br>----------------------------------------------------------------------<br><br>Date:&nbsp;&nbsp;&nbsp;&nbsp;Thu, 31 Aug 2006 10:37:03 +0000 <br>From:&nbsp;&nbsp;&nbsp;&nbsp;&quot;<a href="mailto:reinifrosch@xxxxxxxx">reinifrosch@xxxxxxxx</a>&quot; &lt;<a href="mailto:reinifrosch@xxxxxxxx">reinifrosch@xxxxxxxx</a>&gt;<br>Subject: Cochlear liquid-particle trajectories.<br> <br>Dear List,<br><br>The calculations on the &quot;new&quot; basilar-membrane stiffness<br>formula take longer than expected. If they give interesting<br>results, I shall communicate them in a week or two.<br><br>Yesterday I received a cochlear-mechanics report on a <br>transmission-line model, containing a graph of liquid<br>streamlines which disagree with what I have learned.<br><br>In the case of small-displacement surface waves on<br>the ocean it is helpful to change to a coordinate system <br>which moves along with the wave crests:<br><br>x' = x - c * t<br><br>(where x = longitudinal coordinate, t = time, c = phase<br>velocity). In the primed coordinate system, the liquid<br>particles make a stationary flow in the -x'-direction. <br>At the crests, the streamlines are far apart from each other,<br>so that the velocity v_x' of the particles is comparatively low.<br><br>Transformation back to the lab system yields closed,<br>elliptical liquid-particle trajectories (circular for short <br>waves).<br>In long waves (e.g., tsunamis far from the coast)&nbsp;&nbsp;the particle<br>trajectories are oblong ellipses; at the ocean floor, the length<br>of the short ellipse axis is zero.<br><br>In cochlear waves, the phase velocity c decreases if x <br>increases; the liquid-particle trajectories, however, are<br>similar (I believe) to those in ocean waves, i.e., closed<br>and approximately elliptical.<br><br>The elliptical particle trajectories do not cross the basilar <br>membrane (BM). At the BM, the liquid particles move<br>along the BM. That is why the mass of the organ-of-Corti<br>cells (which in the idealistic models do not move along<br>the BM) plays a role even if their density is equal to <br>that of the liquid.<br><br>A good introduction into cochlear waves was written<br>by the late G.K. Yates: Chapter 2 of the book &quot;Hearing&quot;,<br>B.C.J. Moore, ed., Academic Press, San Diego,<br>1995, Section II.B , pages 49-53.<br><br>With best wishes,<br><br>Reinhart Frosch.<br><br><br>Reinhart Frosch,<br>Dr. phil. nat.,<br>r. PSI and ETH Zurich,<br>Sommerhaldenstr. 5B,<br>CH-5200 Brugg.<br>Phone: 0041 56 441 77 72.<br>Mobile: 0041 79 754 30 32. <br>E-mail: <a href="mailto:reinifrosch@xxxxxxxx">reinifrosch@xxxxxxxx</a> .<br><br>------------------------------<br><br>Date:&nbsp;&nbsp;&nbsp;&nbsp;Thu, 31 Aug 2006 10:08:03 -0500<br>From:&nbsp;&nbsp;&nbsp;&nbsp;&quot;Ferguson, Sarah Hargus&quot; &lt;<a href="mailto:safergus@xxxxxxxx"> safergus@xxxxxxxx</a>&gt;<br>Subject: tympanometry probe tone level?<br><br>Hello list! Students ask the darnedest questions. Today in Intro to<br>Audiology, one of the students asked what the presentation level is for<br>the probe tone used in tympanometry. I didn't find it in a quick Google <br>search (not even a manufacturers' specs I looked at) or during a quick<br>peek in a classic audiology textbook - perhaps someone out there knows<br>off-hand what it is?=20<br><br>~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~=20 <br>Sarah Hargus Ferguson, Ph.D., CCC-A<br>Assistant Professor<br>Department of Speech-Language-Hearing: Sciences and Disorders=20<br>University of Kansas=20<br>Dole Center=20<br>1000 Sunnyside Ave., Room 3001=20<br>Lawrence, KS&nbsp;&nbsp;66045 <br>office: (785)864-1116<br>Speech Acoustics and Perception Lab: (785)864-0610=20<br><a href="http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html">http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html</a><br><br>------------------------------ <br><br>Date:&nbsp;&nbsp;&nbsp;&nbsp;Thu, 31 Aug 2006 11:45:04 EDT<br>From:&nbsp;&nbsp;&nbsp;&nbsp;Harriet B Jacobster &lt;<a href="mailto:Hjacobster@xxxxxxxx">Hjacobster@xxxxxxxx</a>&gt;<br>Subject: Re: tympanometry probe tone level?<br><br>-------------------------------1157039103 <br>Content-Type: text/plain; charset=&quot;US-ASCII&quot;<br>Content-Transfer-Encoding: 7bit<br><br>It's around 85 dBSPL<br><br>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~<br>Harriet Jacobster, Au.D., CCC-A, FAAA<br>Board Certified in Audiology <br><br>-------------------------------1157039103<br>Content-Type: text/html; charset=&quot;US-ASCII&quot;<br>Content-Transfer-Encoding: quoted-printable<br><br>&lt;!DOCTYPE HTML PUBLIC &quot;-//W3C//DTD HTML 4.0 Transitional//EN&quot;&gt; <br>&lt;HTML&gt;&lt;HEAD&gt;<br>&lt;META http-equiv=3DContent-Type content=3D&quot;text/html; charset=3DUS-ASCII&quot;&gt;<br>&lt;META content=3D&quot;MSHTML 6.00.2900.2963&quot; name=3DGENERATOR&gt;&lt;/HEAD&gt;<br>&lt;BODY id=3Drole_body style=3D&quot;FONT-SIZE: 10pt; COLOR: #000000; FONT-FAMILY:=20= <br>Arial&quot;=20<br>bottomMargin=3D7 leftMargin=3D7 topMargin=3D7 rightMargin=3D7&gt;&lt;FONT id=3Drol=<br>e_document=20<br>face=3DArial color=3D#000000 size=3D2&gt;<br>&lt;DIV&gt;It's around 85 dBSPL&lt;/DIV&gt;<br>&lt;DIV&gt;&amp;nbsp;&lt;/DIV&gt; <br>&lt;DIV&gt;~~~~~~~~~~~~~~~~~~~~~~~~~~~~~&lt;/DIV&gt;<br>&lt;DIV&gt;Harriet Jacobster, Au.D., CCC-A, FAAA&lt;/DIV&gt;<br>&lt;DIV&gt;Board Certified in Audiology&lt;/DIV&gt;&lt;/FONT&gt;&lt;/BODY&gt;&lt;/HTML&gt;<br><br> -------------------------------1157039103--<br><br>------------------------------<br><br>Date:&nbsp;&nbsp;&nbsp;&nbsp;Thu, 31 Aug 2006 08:47:42 -0700<br>From:&nbsp;&nbsp;&nbsp;&nbsp;Navid Shahnaz &lt;<a href="mailto:nshahnaz@xxxxxxxx">nshahnaz@xxxxxxxx </a>&gt;<br>Subject: Re: tympanometry probe tone level?<br><br>Hi<br>It depends on the probe tone frequency and it is usually set at a level bel=<br>ow the level the elicit the aural reflex. It varies a bit between manufactu= <br>rer; however, it is close the the following values.<br>226 Hz @xxxxxxxx 85dB SPL =B1 1.5dB<br>1000 Hz (a probe tone frequency of choice for infant assessment) @xxxxxxxx 75dB SP=<br>L =B1 1.5dB<br>Hope this helps.<br>Best<br>Navid<br> -------------------------------<br>Navid Shahnaz, Ph.D<br>Assistant Professor<br>School of Audiology &amp; Speech Sciences<br>Faculty of Medicine<br>5804 Fairview Ave.<br>Vancouver, B.C. V6t 1Z3<br>Canada<br>Tel. 604-822-5953 <br>Fax. 604-822-6569<br>E-mail: <a href="mailto:nshahnaz@xxxxxxxx">nshahnaz@xxxxxxxx</a><br>Website: <a href="http://www.audiospeech.ubc.ca/school/faculty/navid/">http://www.audiospeech.ubc.ca/school/faculty/navid/ </a><br><br>---------- Original Message ----------------------------------<br>From: &quot;Ferguson, Sarah Hargus&quot; &lt;<a href="mailto:safergus@xxxxxxxx">safergus@xxxxxxxx</a>&gt;<br>Reply-To: &quot;Ferguson, Sarah Hargus&quot; &lt; <a href="mailto:safergus@xxxxxxxx">safergus@xxxxxxxx</a>&gt;<br>Date:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Thu, 31 Aug 2006 10:08:03 -0500<br><br>&gt;Hello list! Students ask the darnedest questions. Today in Intro to<br>&gt;Audiology, one of the students asked what the presentation level is for <br>&gt;the probe tone used in tympanometry. I didn't find it in a quick Google<br>&gt;search (not even a manufacturers' specs I looked at) or during a quick<br>&gt;peek in a classic audiology textbook - perhaps someone out there knows <br>&gt;off-hand what it is?<br>&gt;<br>&gt;~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~<br>&gt;Sarah Hargus Ferguson, Ph.D., CCC-A<br>&gt;Assistant Professor<br>&gt;Department of Speech-Language-Hearing: Sciences and Disorders <br>&gt;University of Kansas<br>&gt;Dole Center<br>&gt;1000 Sunnyside Ave., Room 3001<br>&gt;Lawrence, KS&nbsp;&nbsp;66045<br>&gt;office: (785)864-1116<br>&gt;Speech Acoustics and Perception Lab: (785)864-0610<br>&gt;<a href="http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html"> http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html</a><br>&gt;<br><br><br><br><br><br>________________________________________________________________<br>Sent via the WebMail system at <a href="http://audiospeech.ubc.ca"> audiospeech.ubc.ca</a><br><br><br><br><br><br>------------------------------<br><br>Date:&nbsp;&nbsp;&nbsp;&nbsp;Thu, 31 Aug 2006 11:20:43 -0500<br>From:&nbsp;&nbsp;&nbsp;&nbsp;Jeremy Federman &lt;<a href="mailto:jeremy.federman@xxxxxxxx">jeremy.federman@xxxxxxxx </a>&gt;<br>Subject: Re: tympanometry probe tone level?<br><br>Hi -<br><br>I suspect it depends on the frequency being used (and maybe the equipment,<br>too), but I believe the probe tone level at the standard 226 Hz should be 85 <br>dB SPL plus or minus 1.5 dB.&nbsp;&nbsp;For additional calibration info, you could<br>check the ANSI standard...<br><br>With kind regards,<br><br>Jeremy<br><br>On 8/31/06 10:08 AM, &quot;Ferguson, Sarah Hargus&quot; &lt;<a href="mailto:safergus@xxxxxxxx"> safergus@xxxxxxxx</a>&gt; wrote:<br><br>&gt; Hello list! Students ask the darnedest questions. Today in Intro to<br>&gt; Audiology, one of the students asked what the presentation level is for<br>&gt; the probe tone used in tympanometry. I didn't find it in a quick Google <br>&gt; search (not even a manufacturers' specs I looked at) or during a quick<br>&gt; peek in a classic audiology textbook - perhaps someone out there knows<br>&gt; off-hand what it is?<br>&gt;<br>&gt; ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ <br>&gt; Sarah Hargus Ferguson, Ph.D., CCC-A<br>&gt; Assistant Professor<br>&gt; Department of Speech-Language-Hearing: Sciences and Disorders<br>&gt; University of Kansas<br>&gt; Dole Center<br>&gt; 1000 Sunnyside Ave., Room 3001 <br>&gt; Lawrence, KS&nbsp;&nbsp;66045<br>&gt; office: (785)864-1116<br>&gt; Speech Acoustics and Perception Lab: (785)864-0610<br>&gt; <a href="http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html">http://www.ku.edu/~splh/ipcd/Faculty/FergusonBio.html </a><br>&gt;<br><br>--<br>Jeremy Federman, MS, CCC-A<br>Vanderbilt University Medical Center<br>Dan Maddox Hearing Aid Research Lab<br>Department of Hearing and Speech Sciences<br>Medical Center East, South Tower<br>1215 21st Ave. South, Room 8310 <br>Nashville, TN&nbsp;&nbsp;37232-8242<br><a href="mailto:jeremy.federman@xxxxxxxx">jeremy.federman@xxxxxxxx</a><br><br>------------------------------<br><br>Date:&nbsp;&nbsp;&nbsp;&nbsp;Thu, 31 Aug 2006 18:32:56 -0400<br>From:&nbsp;&nbsp;&nbsp;&nbsp;Arturo Camacho &lt; <a href="mailto:acamacho@xxxxxxxx">acamacho@xxxxxxxx</a>&gt;<br>Subject: Re: STFT vs Power Spectral in Musical recognition system ?<br><br>One problem of the square-root compression is that its slope<br>approaches infinity as the magnitude M approaches zero. A more <br>appropriate approach may be to use log(1+KM), where K is a constant to<br>be determined. The response of this function is almost logarithmic for<br>high magnitudes and almost linear for low magnitudes. Of course, the<br> determination of the optimal value for K given an input is not<br>trivial.<br><br>Arturo<br>--<br>__________________________________________________<br><br>Arturo Camacho<br>PhD Candidate<br>Computer and Information Science and Engineering <br>University of Florida<br><br>E-mail: <a href="mailto:acamacho@xxxxxxxx">acamacho@xxxxxxxx</a><br>Web page: <a href="http://www.cise.ufl.edu/~acamacho">www.cise.ufl.edu/~acamacho</a><br>__________________________________________________ <br><br>On Fri, 25 Aug 2006, Richard F. Lyon wrote:<br><br>&gt; Edwin,<br>&gt;<br>&gt; A power spectral density is only defined for stationary signals, not<br>&gt; music.&nbsp;&nbsp;The STFT generalizes it to short segments, if you use the <br>&gt; squared magnitude.<br>&gt;<br>&gt; The difference between the absolute value, square, log, etc. are just<br>&gt; point nonlinearities that do not change the information content, but<br>&gt; do change the metric structure of the space a bit.&nbsp;&nbsp;Log is too <br>&gt; compressed, leading to too much emphasis on near-silent segments,<br>&gt; while the square (the power you ask about) is too expanded, leading<br>&gt; to too much emphasis on the louder parts.&nbsp;&nbsp;A good compromise is <br>&gt; around a square root or cube root of magnitude (roughly matching<br>&gt; perceptual magnitude via Stevens's law), but the magnitude itself is<br>&gt; also sometimes acceptable, depending on what you're doing.<br> &gt;<br>&gt; Dick<br>&gt;<br>&gt; At 7:12 AM -0700 8/25/06, Edwin Sianturi wrote:<br>&gt; &gt;Content-Type: text/html<br>&gt; &gt;X-MIME-Autoconverted: from 8bit to quoted-printable by<br>&gt; &gt;<a href="http://torrent.cc.mcgill.ca"> torrent.cc.mcgill.ca</a> id k7PED6jh031610<br>&gt; &gt;<br>&gt; &gt;Hello,<br>&gt; &gt;<br>&gt; &gt;I am just a master student, doing my internship. Right now, I am<br>&gt; &gt;building a musical instrument recognition system. I have read <br>&gt; &gt;several papers on it, and I am just curious:<br>&gt; &gt;<br>&gt; &gt;All the papers/journals that I have read use the STFT, a.k.a the<br>&gt; &gt;|X(t,f)| of a signal x(t), in order to extract several (spectral) <br>&gt; &gt;features to be used as the input to the recognition system.<br>&gt; &gt;<br>&gt; &gt;What are the reasons behind using the |X(t,f)| instead of using the<br>&gt; &gt;&quot;power spectral&quot; |X(t,f)|^2 ?<br> &gt; &gt;(technically speaking, a power spectral density is the expectation<br>&gt; &gt;of |X(f)|^2, i.e. E(|X(f)|^2) )<br>&gt; &gt;<br>&gt; &gt;Thanks in advance,<br>&gt; &gt;<br>&gt; &gt;Edwin SIANTURI<br>&gt; &gt;<br>&gt; <br>&gt;<br><br>------------------------------<br><br>Date:&nbsp;&nbsp;&nbsp;&nbsp;Thu, 31 Aug 2006 16:00:31 -0700<br>From:&nbsp;&nbsp;&nbsp;&nbsp;&quot;Richard F. Lyon&quot; &lt;<a href="mailto:DickLyon@xxxxxxxx">DickLyon@xxxxxxxx</a>&gt;<br>Subject: Re: STFT vs Power Spectral in Musical recognition system ? <br><br>Arturo, I totally agree with the idea of using log(1 + KM), or log(M<br>+ epsilon) as I usually do it.&nbsp;&nbsp;This kind of nonlinearity is<br>especially important in systems with an imprecisely known zero level<br>or a variable noise floor.&nbsp;&nbsp;On the other hand, a power law, though it <br>has an infinite slope at 0, is not half as bad as a plain log, and<br>lots of people use that anyway.&nbsp;&nbsp;A stabilized power law, like (M +<br>epsilon)^(1/3) is another good choice, probably more in line with<br>perception that letting it go log-like at high magnitudes.&nbsp;&nbsp;With any <br>of these, adjusting your parameters to accommodate a realistic range<br>of input signal levels becomes important; you can no longer ignore<br>scale factors and hope for algorithms to work fine on inputs varying<br>over many orders of magnitude in scale. <br><br>Dick<br><br>At 6:32 PM -0400 8/31/06, Arturo Camacho wrote:<br>&gt;One problem of the square-root compression is that its slope<br>&gt;approaches infinity as the magnitude M approaches zero. A more<br>&gt;appropriate approach may be to use log(1+KM), where K is a constant to <br>&gt;be determined. The response of this function is almost logarithmic for<br>&gt;high magnitudes and almost linear for low magnitudes. Of course, the<br>&gt;determination of the optimal value for K given an input is not <br>&gt;trivial.<br>&gt;<br>&gt;Arturo<br>&gt;--<br>&gt;__________________________________________________<br>&gt;<br>&gt;&nbsp;&nbsp;Arturo Camacho<br>&gt;&nbsp;&nbsp;PhD Candidate<br>&gt;&nbsp;&nbsp;Computer and Information Science and Engineering<br> &gt;&nbsp;&nbsp;University of Florida<br>&gt;<br>&gt;&nbsp;&nbsp;E-mail: <a href="mailto:acamacho@xxxxxxxx">acamacho@xxxxxxxx</a><br>&gt;&nbsp;&nbsp;Web page: <a href="http://www.cise.ufl.edu/~acamacho">www.cise.ufl.edu/~acamacho</a><br>&gt;__________________________________________________ <br>&gt;<br>&gt;On Fri, 25 Aug 2006, Richard F. Lyon wrote:<br>&gt;<br>&gt;&gt;&nbsp;&nbsp;Edwin,<br>&gt;&gt;<br>&gt;&gt;&nbsp;&nbsp;A power spectral density is only defined for stationary signals, not<br>&gt;&gt;&nbsp;&nbsp;music.&nbsp;&nbsp;The STFT generalizes it to short segments, if you use the <br>&gt;&gt;&nbsp;&nbsp;squared magnitude.<br>&gt;&gt;<br>&gt;&gt;&nbsp;&nbsp;The difference between the absolute value, square, log, etc. are just<br>&gt;&gt;&nbsp;&nbsp;point nonlinearities that do not change the information content, but<br>&gt;&gt;&nbsp;&nbsp;do change the metric structure of the space a bit.&nbsp;&nbsp;Log is too <br>&gt;&gt;&nbsp;&nbsp;compressed, leading to too much emphasis on near-silent segments,<br>&gt;&gt;&nbsp;&nbsp;while the square (the power you ask about) is too expanded, leading<br>&gt;&gt;&nbsp;&nbsp;to too much emphasis on the louder parts.&nbsp;&nbsp;A good compromise is <br>&gt;&gt;&nbsp;&nbsp;around a square root or cube root of magnitude (roughly matching<br>&gt;&gt;&nbsp;&nbsp;perceptual magnitude via Stevens's law), but the magnitude itself is<br>&gt;&gt;&nbsp;&nbsp;also sometimes acceptable, depending on what you're doing. <br>&gt;&gt;<br>&gt;&gt;&nbsp;&nbsp;Dick<br>&gt;&gt;<br>&gt;&gt;&nbsp;&nbsp;At 7:12 AM -0700 8/25/06, Edwin Sianturi wrote:<br>&gt;&gt;&nbsp;&nbsp;&gt;Content-Type: text/html<br>&gt;&gt;&nbsp;&nbsp;&gt;X-MIME-Autoconverted: from 8bit to quoted-printable by<br>&gt;&gt;&nbsp;&nbsp;&gt; <a href="http://torrent.cc.mcgill.ca">torrent.cc.mcgill.ca</a> id k7PED6jh031610<br>&gt;&gt;&nbsp;&nbsp;&gt;<br>&gt;&gt;&nbsp;&nbsp;&gt;Hello,<br>&gt;&gt;&nbsp;&nbsp;&gt;<br>&gt;&gt;&nbsp;&nbsp;&gt;I am just a master student, doing my internship. Right now, I am <br>&gt;&gt;&nbsp;&nbsp;&gt;building a musical instrument recognition system. I have read<br>&gt;&gt;&nbsp;&nbsp;&gt;several papers on it, and I am just curious:<br>&gt;&gt;&nbsp;&nbsp;&gt;<br>&gt;&gt;&nbsp;&nbsp;&gt;All the papers/journals that I have read use the STFT, a.k.a the<br>&gt;&gt;&nbsp;&nbsp;&gt;|X(t,f)| of a signal x(t), in order to extract several (spectral)<br>&gt;&gt;&nbsp;&nbsp;&gt;features to be used as the input to the recognition system.<br>&gt;&gt;&nbsp;&nbsp;&gt;<br>&gt;&gt;&nbsp;&nbsp;&gt;What are the reasons behind using the |X(t,f)| instead of using the <br>&gt;&gt;&nbsp;&nbsp;&gt;&quot;power spectral&quot; |X(t,f)|^2 ?<br>&gt;&gt;&nbsp;&nbsp;&gt;(technically speaking, a power spectral density is the expectation<br>&gt;&gt;&nbsp;&nbsp;&gt;of |X(f)|^2, i.e. E(|X(f)|^2) )<br>&gt;&gt;&nbsp;&nbsp;&gt;<br>&gt;&gt;&nbsp;&nbsp;&gt;Thanks in advance, <br>&gt;&gt;&nbsp;&nbsp;&gt;<br>&gt;&gt;&nbsp;&nbsp;&gt;Edwin SIANTURI<br>&gt;&gt;&nbsp;&nbsp;&gt;<br>&gt;&gt;<br>&gt;&gt;<br><br>------------------------------<br><br>End of AUDITORY Digest - 30 Aug 2006 to 31 Aug 2006 (#2006-197)<br>*************************************************************** <br><br></blockquote></div><br><br clear="all"><br>-- <br>-----------------------------------------------------------<br>Stand on the shoulders of giants<br><br>Ajith Kumar Uppunda<br>Post doctoral fellow,<br>Department of communication sciences and disorders, <br>Speech research laboratory,<br>Northwestern University, Evanston, IL, 60208.<br>Ph: 847 491 2430<br>Email&nbsp;&nbsp;<a href="mailto:a-uppunda@xxxxxxxx">a-uppunda@xxxxxxxx</a> ------=_Part_72126_28863797.1157120093958--


This message came from the mail archive
http://www.auditory.org/postings/2006/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University