Re: perceptual segregation of sound (Kevin Austin )


Subject: Re: perceptual segregation of sound
From:    Kevin Austin  <kevin.austin@xxxxxxxx>
Date:    Thu, 27 Apr 2006 08:43:27 -0400

Hi Mark I am not a psychologist but have worked on auditory analysis in an electroacoustic (music) context quite extensively. This is one of the 'non-questions' I bring to my classes. My long-term perspective is that the question may make assumptions about the linearity of (the perception of) time, and the nature of sonic identity. To get to this place my proposition is that sound is the displacement of air. Assume a mono source, eg a recording of Furtwangler conducting the Beethoven Seventh in 1943. Using a 78 rpm disc (an analog for the ear-drum), the needle is in only one location at any point in 'time'. We look at the groove and have change of amplitude. Quantize the time, eg at 44.1kHz. How many samples are required before the first beat is heard. This information, as I see it, is not intrinsic to the source, it is about how the perceptual system interprets the information. To ask how many sounds we can hear "at the same time" (= concurrently), may not be a valid question. Let me try it this way ... "how many things can you see concurrently?" Does the question make 'sense'? It is linguistically well-formed, but has no clear semantic. The first chord of the Beethoven arrives at my ear. I need to wait until I have 'windowed' enough of the sound to start to segregate / integrate the sound. From my experience, my perception works on my memory of the sound. And there are multiple levels of windowing, as I can also 'scan' the entire symphony in a few seconds. (Think Google Earth.) An experiment ... for a moment, as you read this, right now, clap your hands. Clap again. Listen to the 'sounds'. If this is followed, the 'sounds' were gone before you were asked to "listen", but most people are able to 'listen back' (recall) what they heard. Asking the psychometric / linguistic question of "How many sounds did you hear", may result in answers from "One. I heard two claps.", to "Two. I heard one clap and then another clap.", or ... more. Since I was writing this, I also heard the click of the keys, the truck that went by, the floor creak and the fan from the computer ... and my tinnitus. I rub my hands and hear my clothes rub .... and I listen to the voice in my head dictating or re-reading these words. The chair creaks. Some of these are temporally linear, and I segregate them into 'unique events'. The claps are not streamed into the truck, but the two claps may be streamed together. In my experience, this will depend upon how I choose to group and segregate the sounds in my head; the condition is not inherent in the physical displacement of the air. And the Beethoven Seventh. Memory and learning. When I was younger I heard that first chord as a three dimensional spectral explosion, and could 'identify' the oboe and violin. Maybe I could segregate the oboe because he doesn't cut off. (Beethoven had a problem with oboes cutting off, see the recapitulation of the first movement of the Fifth.) I have a simultaneous integration and segregation ... which is (as we know), impossible. (It is not possible, I am told, to see both versions of the Necker cube simultaneously. When I look on the page I can't, but when I recall the image in my minds' eye, I can.) I have somewhat similar experiences with sound, and I notice that I seem to spend less and less time 'hearing' the sounds arriving, and more and more time 'listening' to my internal representations. Even as the oboe holds the the high 'A', I (pre-)hear the following "E", and seek out the viola in the previous chord. I "hear" the present, past and future "at the same time". It has been my considered opinion for a long time that if one wants to understand the potential and limits of human perception, one doesn't work with 'ordinary' subjects. I would be a dreadful person to test for physical endurance as the results from testing me would be mundane. However, to start to answer your question, I would propose that you only work with people who have 'really good' hearing. Working from this position, start to work out how they perceive and interpret the stimulus. In my classes, I have had students who can "go back" to a sound (or sounds) they heard and extract components that they did not 'hear' when the sound was presented. In one case a student re-listened to a piece he had heard a couple of weeks previously.) It is, in my experience, possible to raise the level of awareness regarding integration, segregation, streaming, grouping etc, through detailed and focused listening exercises. This response may not be much use to you, but it's about as good as I can get it on paper at this time. Best wishes Kevin Oh ... and how did this evolve? The listener who heard both lions had children, the one who heard only one was called "supper". >Date: Wed, 26 Apr 2006 19:08:39 +0100 >From: Mark Every <m.every@xxxxxxxx> >Subject: > >Dear all, > >I have a question of general interest about auditory scene analysis, and >would be grateful for any views on the subject. > >Humans have the remarkable ability to interpret multiple events and >perceive distinct sources within a complex sound environment. The >concept of perceptual streams of information is explained in (Bregman >A.S., Auditory Scene Analysis: The Perceptual Organization of Sound, MIT >Press, 1990). My question is: are we really capable of perceptually >segregating multiple sources concurrently, or are we just focusing our >attention on one source, and then shifting it very quickly to another >source? Evolutionarily speaking, I can see it would be advantageous if >we were being chased by two lions rather than one, to be able to >concentrate on both simultaneously. However, I also have the impression >that if lion 1 roars, we will forget about lion 2 for a moment, and vice >versa. If both roar at the same time, are we now listening to two >separate lions or just a general sound of lions roaring? >In any case, run for it... > >Best, >Mark > >-- >Mark Every <m.every@xxxxxxxx> >CVSSP, SEPS, University of Surrey


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