Subject: Re: Does anybody know a similar study? From: Owen Brimijoin <owen@xxxxxxxx> Date: Wed, 18 Jun 2014 09:30:15 +0100 List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>This is a multi-part message in MIME format. --------------040008010101030209060502 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: quoted-printable Hi Massimo, Repetition can help immensely; I'd suggest that you take a peek at some=20 nice work by Josh McDermott et al on the effects of repetition in noisy=20 signals: McDermott, J.H., Wrobleski, D., Oxenham, A.J. (2011) Recovering sound=20 sources from embedded repetition. Proceedings of the National Academy of=20 Sciences, 108, 1188-1193. Also (and at the risk of self-promotion) I will say that we published a=20 study last year on a somewhat related phenomenon that you are alluding=20 to here, that of hearing things that are not in a signal. We presented=20 listeners with broadband noise that was -- in the long term white -- but=20 changed randomly in its spectrum from moment to moment. We then asked=20 them to press a button when they heard one of two vowel sounds. In some=20 trials they were listening for [a] and in others [i:]. Then using=20 response triggered averaging (reverse correlation), we averaged the=20 spectrum of the signals leading up to each button press and voil=E0: what= =20 emerged was the spectrum of the vowel they were listening for. So despite the fact that we did not embed any actual vowel sounds in the=20 random noise, listeners were responding to subtle variations in the=20 noise that were somehow similar enough to their internal representations=20 of those sounds. Turns out to be a nice (and relatively unbiased) way to=20 estimate what someone /imagines /a particular sound to be. See the JASA=20 article here: http://scitation.aip.org/content/asa/journal/jasa/133/2/10.1121/1.4778264 all the best, -Owen. ------------------------------------- W. Owen Brimijoin, PhD Investigator Scientist MRC/CSO Institute of Hearing Research Glasgow, United Kingdom +44 (0) 141 201 8766 owen@xxxxxxxx ------------------------------------- On 17/06/2014 15:46, Massimo Grassi wrote: > Dear list members, > > yesterday I colleague played me a sample (a sentence) of highly=20 > degraded speech. It was a recording made in a highly noisy=20 > environment. It included speech (a conversation) that was hardly=20 > intelligible except for a few occasional words. > > The colleague asked me to listen to the sample and pay attention=20 > whether I was able to spot a few target words. These words were not=20 > intelligible to me. > > The colleague then selected a portion of the recording and played it=20 > in loop. That portion included (according to him) one target word.=20 > After a few loops I was able to "perceive" the word. > > This is exactly the problem. I'm wandering whether it was just a=20 > suggestion due to the repeated listening of an ambiguous auditory=20 > signal. A kid of auditory Rorschach test: there seem to be nothing at=20 > the beginning but if you keep listening you can hear whatever you like. > > Is there anybody out there that is aware of studies that investigated=20 > whether listening in loop to an ambiguous signal can lead to hear=20 > things that are not in the signal? > > I didn't find anything yet. > > Thank you all in advance, > m > --------------040008010101030209060502 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit <html> <head> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body style="background-color: rgb(255, 255, 255); color: rgb(0, 0, 0); font-family: Helvetica,Arial,sans-serif; font-size: 13px;" text="#000000" bgcolor="#FFFFFF"> <div class="moz-cite-prefix">Hi Massimo,<br> <br> Repetition can help immensely; I'd suggest that you take a peek at some nice work by Josh McDermott et al on the effects of repetition in noisy signals:<br> <br> McDermott, J.H., Wrobleski, D., Oxenham, A.J. (2011) Recovering sound sources from embedded repetition. Proceedings of the National Academy of Sciences, 108, 1188-1193.<br> <br> Also (and at the risk of self-promotion) I will say that we published a study last year on a somewhat related phenomenon that you are alluding to here, that of hearing things that are not in a signal. We presented listeners with broadband noise that was -- in the long term white -- but changed randomly in its spectrum from moment to moment. We then asked them to press a button when they heard one of two vowel sounds. In some trials they were listening for [a] and in others [i:]. Then using response triggered averaging (reverse correlation), we averaged the spectrum of the signals leading up to each button press and voilà: what emerged was the spectrum of the vowel they were listening for. <br> <br> So despite the fact that we did not embed any actual vowel sounds in the random noise, listeners were responding to subtle variations in the noise that were somehow similar enough to their internal representations of those sounds. Turns out to be a nice (and relatively unbiased) way to estimate what someone <i>imagines </i>a particular sound to be. See the JASA article here:<br> <br> <a class="moz-txt-link-freetext" href="http://scitation.aip.org/content/asa/journal/jasa/133/2/10.1121/1.4778264">http://scitation.aip.org/content/asa/journal/jasa/133/2/10.1121/1.4778264</a><br> <br> all the best, <br> -Owen.<br> <br> <pre class="moz-signature" cols="72">------------------------------------- W. Owen Brimijoin, PhD Investigator Scientist MRC/CSO Institute of Hearing Research Glasgow, United Kingdom +44 (0) 141 201 8766 <a class="moz-txt-link-abbreviated" href="mailto:owen@xxxxxxxx">owen@xxxxxxxx</a> -------------------------------------</pre> On 17/06/2014 15:46, Massimo Grassi wrote:<br> </div> <blockquote style="font-size: medium;" cite="mid:13281_1403064679_53A11167_13281_262_20_53A054E0.70104@xxxxxxxx" type="cite">Dear list members, <br> <br> yesterday I colleague played me a sample (a sentence) of highly degraded speech. It was a recording made in a highly noisy environment. It included speech (a conversation) that was hardly intelligible except for a few occasional words. <br> <br> The colleague asked me to listen to the sample and pay attention whether I was able to spot a few target words. These words were not intelligible to me. <br> <br> The colleague then selected a portion of the recording and played it in loop. That portion included (according to him) one target word. After a few loops I was able to "perceive" the word. <br> <br> This is exactly the problem. I'm wandering whether it was just a suggestion due to the repeated listening of an ambiguous auditory signal. A kid of auditory Rorschach test: there seem to be nothing at the beginning but if you keep listening you can hear whatever you like. <br> <br> Is there anybody out there that is aware of studies that investigated whether listening in loop to an ambiguous signal can lead to hear things that are not in the signal? <br> <br> I didn't find anything yet. <br> <br> Thank you all in advance, <br> m <br> <br> </blockquote> <br> </body> </html> --------------040008010101030209060502--