Subject: Re: Computational ASA -- how many sources can humans perceive? From: Peter Lennox <peter(at)LENNOX01.FREESERVE.CO.UK> Date: Mon, 3 May 2004 13:43:14 +0100On the evidence below: Some of this might be that perception in general (as against specifically auditory prc) isn't good at simulteneity. Look at the numbers of dots on a die - we can all recognise "five" because of the pattern that is the conventional way of arranging 5 dots on a die.alter the pattern, and see what happens. Similarly, I have 6 hens wandering round my garden; there are two black ones, two marans (speckeldy grey), one brown hen and one smaller yokohama. If they happen to be lined up just so, I can instantly tell that they are all present. If they are well separated - even if in my field of view simultaneously, I can't tell as easily; I have to count them. Counting isn't simultaneous perception, it's sequential. I had once been told that hens were stupid because they could only count: one, two, and 'many'; it turns out that I'm not all that much brighter! But is the ability to correctly report the number of simultaneous items synonymous with 'perception'? - after all, when I drive at (relatively) high speed down the road, I'm clearly not able to accurately report the number of items; however, I generally manage to avoid them. Am I able, therefore, to perceptually use 'non-segregated', holistic information? This is really an argument for a kind of 'perceptual background/context' that contains items that have the potential for segregation into foreground objects of attention, but surely not simultaneously. Cocktail party effects just wouldn't work if we couldn't understand anything if there were more than three talkers. But I don't need to consciously judge the number of talkers to hear if my name is spoken; surely, I don't need to unconciously judge the number either? On the other hand, if I was asked to find a particular talker in a crowded large room by hearing alone, could I do it? - given enough time, possibly. Likewise, in the rain it is rather difficult to focus on a particular raindrop - however, if there is a regular drip from a particular feature, one can pick that out, and after a while, one can pick out scores - but it's not simultaneous, or effortless. There seems to be a possibility of conflating perception with attention, perhaps? - or rather, it's very difficult to tease them apart, but this is what would be necessary if one is to study the equivalent of scene analysis modules that apply particular algorithms to an auditory compound. In any event, is it a reasonable assumption that humans actually do implement ASA as some kind of stand alone application? - surely, this would be some kind of special case, whereas the generality of scene analysis problems would be approached with all senses, and these might integrate at even quite peripheral levels, so that 'scene analysis' is more often an auditory/visual/olfactory/haptic kind of problem? regards ppl >"For instance, > Jennifer Tufts and Tom Frank J. Acoust. Soc. Am. 101 , 3107 (1997) found > that the accuracy of judging the number of talkers in a multitalker mixture > drops considerably when there are more than 3 talkers. "There is also a > report by David Huron (Music Perception, Vol. 19, No. 1 (2001) pp. 1-64., > or on-line > http://www.music-cog.ohio-state.edu/Huron/Publications/huron.voice.leading.h tml > ) that estimating the number of musical lines in > polyphonic music worsens considerably after 3. Some anecdotal evidence > for this limit also comes from movie sound effect designers. This is a > citation from Walter Murch, a renown sound effect artist: "There is a rule > of thumb I use which is never to give the audience more than two-and-a-half > things to think about aurally at any one moment. Now, those moments can > shift very quickly, but if you take a five-second section of sound and feed > the audience more than two-and-a-half conceptual lines at the same time, > they can't really separate them out. There's just no way to do it, and > everything becomes self-canceling." (cited from > http://www.filmsound.org/murch/waltermurch.htm)" > > Any thoughts, comments, and references relevant to this issue are > appreciated. > > > > ------------------------------------------------------------- > Valeriy Shafiro > Communication Disorders and Sciences > Rush University Medical Center > Chicago, IL > > office (312) 942 - 3298 > lab (312) 942 - 3316 > email: valeriy_shafiro(at)rush.edu