Re: Auditory wheel (Neil Waterman )


Subject: Re: Auditory wheel
From:    Neil Waterman  <neil.waterman@xxxxxxxx>
Date:    Fri, 19 Mar 2010 10:51:14 -0400
List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>

Hi List, I don't want to be a "Debbie Downer" (SNL oblique reference : http:// en.wikipedia.org/wiki/Debbie_Downer for those not in the USA), but my practical experience is that the auditory equivalent of the color wheel does not exist, beyond simple periodic tones (which do not wrap around). Here's why - the audio signature of say a piano is equivalent visually to looking at a familiar object, let's pick the Capitol Building in Washington DC, we recognize it from the composite elements. Different pictures of the same building (taken in Winter/ Summer/Day/Night) are somewhat equivalent to taking the same piano and putting it in different acoustic spaces - we still recognize the piano, but it's different every time. The audio signature of the piano is the sum of many parts, just as the Capitol Building is. When it comes to perceiving the color of the Capitol Building using a color wheel as a guide, I'm guessing most folks would characterize it as off-white in a creamy direction (as opposed to off-white in a blue sense for example). Our brain is able to simplify the perception. We come to a very similar color conclusion irrespective of how it is presented (Winter/Summer/etc). I do not believe we can do that with "complex sounds" - instruments of nearly every kind, have complex harmonics. With sound we are very tuned to at least two dimensional aspects - pitch and loudness, and these interact. The same "sound" played loud will possibly sound different if it is then played quietly. With a color wheel we are segmenting the frequency range of the visual spectrum into some discrete blocks. We understand Red, Blue, etc, in a different way from that of sound, and do not find the visual perception of the color wheel wrapping around as a discontinuity even though in frequency terms it is. This is because we are used to perceiving colors as mixing across the spectrum. We do not handle the pitch of sound this way. If we sweep a tone it does not wrap, and our brain would not really understand that. The Shepard Tones are an amusing phenomena, but really do not do what the color wheel does, because at any point in the tone cycle, you cannot really identify the "sound color". I'm writing this not as a heavy-weight academic, so apologies for any incorrect terminology, etc, but I do work with the perception of sound day in day out - I am a project engineer for a company that builds flight simulators, specifically we produce a system that creates the sound environment of the aircraft, and hence understand pretty well how humans perceive sound from a practical sense. We synthesize the sound of complex sources (engines, propellors, air- flows, etc) and as such have to deal with human subjective perception of the sounds we create versus the real sounds in the aircraft. Anyway that's just my 2 cents. Regards, Neil On Mar 18, 2010, at 6:22 PM, Michael H. Coen wrote: > Hello list, > > We are all familiar with the notion of a visual color wheel, e.g., a > continuous, circular representation of colors in some color space such > as HSV. (Here's a wikipedia page with some visual examples: > http://en.wikipedia.org/wiki/Color_wheel) > > Of course, there are many different color wheels, given that there are > many different color spaces in which to model them. > > I'm looking for the auditory equivalent of a color wheel. Namely, a > parametrized, continuous method for generating a series of sounds that > form a "perceptual loop" that has no perceived gaps. > > I've coded a several of these, e.g., a violin morphing into a piano > morphing into a clarinet, which then morphs back into a violin. They > are all playing middle C and the ASDR envelopes for the generated, > equal-length samples are identical. > > I'm doing this to measure discriminative acuity in distinguishing > sounds > along the auditory wheel, as part of a larger multimodal perceptual > experiment. > > However, what I've found is that people with musical backgrounds have > far greater discriminatory power in separating nearby sounds than > those > who have little training. There are also "unintended" clues, such as > harmonic complexity, which people appear differentially sensitive to. > > Thus, I wonder if there is any work in creating such an auditory wheel > that might be expected to reduce bias due to background and/or > culture. > E.g., using a heptatonic music scale may not be a good idea and > sounds > derived from familiar instruments are probably best avoided as > well. I > would like participants to be on an equal auditory footing, so to > speak. > I'll repeat that it is essential that there be a "loop" in the sound > presentation that is not due to a simple repetition of sounds, e.g., a > sine wave rising and then falling in frequency would not be useful. > > Any pointers, suggestions, code, etc., would be most welcome. > > Best, > Michael Coen >


This message came from the mail archive
/home/empire6/dpwe/public_html/postings/2010/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University