gints@prophet.sgi.com (Gints Klimanis) (03/20/91)
In article <3719@ssc-bee.ssc-vax.UUCP>, carroll@ssc-vax (Jeff Carroll) writes: >In article <1991Mar18.195507.25639@odin.corp.sgi.com> gints@prophet.esd.sgi.com (Gints >Klimanis) writes: >> >>In any case, it is not a perfect ability. You can handle pitches with >>about the same accuracy as distinguishing flavors of vanilla ice cream. >>Sure, they're vanilla but HOW vanilla. I doubt anyone was able to >>recognize pitches in units of Hertz. There is too much other evidence >>that would discredit this type of account. > With the aid of my pocket calculator (for taking logarithms to >the base 2^(1/12)), I can get within about 3-5 Hz. I'm sure that I could >get down to 1 Hz with a little practice. My argument against measuring in Hertz, an absolute measure, is that a unit of cycles per second is contrary our anatomy and both pitch and amplitude perception. Both pitch and amplitude perception are calibrated on a non-linear scale. We are less sensitive to small variations in Hertz at high pitches. Same story for loudness. If you can perceive 5 Hz discrepencies at pitches near 10 kHz, you should be able to perceive discrepencies near 100 Hz with a sensitivity that is two orders of magnitude greater. -- Jeff Carroll carroll@ssc-vax.boeing.com
kalt@nmsu.edu (Kerry Alt) (03/20/91)
OK, enough hearsay! Has anyone seen any DEFINITIVE results where subjects have been run before and after using the Burge tests? Surely there must be some REAL data out there to settle this. I admit I'm skeptical, but surely the Music Perception types should have some useful input here... -- ************************************************* ************************************************* ** Kerry Alt at The Computer Operations Group ** ** kalt@nmsu.edu ** ** ph. 505/646-5318 (at COG) ** ** 646-3645 (office) ** ** 524-7247 (home) ** ************************************************* *************************************************
carroll@ssc-vax (Jeff Carroll) (03/20/91)
In article <1991Mar19.195901.6769@odin.corp.sgi.com> gints@prophet.sgi.com (Gints Klimanis) writes: >If you can perceive 5 Hz discrepencies at pitches near 10 kHz, you >should be able to perceive discrepencies near 100 Hz with a sensitivity >that is two orders of magnitude greater. Good point. My problem is that I've been an engineer for too long, and have gotten used to thinking in Hz. My gut feeling, though, is that it's not that simple; that is, that I don't perceive pitch as a simple log-frequency scale. -- Jeff Carroll carroll@ssc-vax.boeing.com
talley@hpuxa.acs.ohio-state.edu (James T. Talley) (03/21/91)
I know that Mark Rush did his dissertation research here at OSU on whether perfect pitch could be learned. I know that he examined Burge's method. I'll check with him for more info and summarize. James T. Talley
edhall@rand.org (Ed Hall) (03/21/91)
In article <3728@ssc-bee.ssc-vax.UUCP> carroll@ssc-vax.UUCP (Jeff Carroll) writes: >In article <1991Mar19.195901.6769@odin.corp.sgi.com> gints@prophet.sgi.com (Gints Klimanis) writes: >>If you can perceive 5 Hz discrepencies at pitches near 10 kHz, you >>should be able to perceive discrepencies near 100 Hz with a sensitivity >>that is two orders of magnitude greater. > > Good point. My problem is that I've been an engineer for too long, and >have gotten used to thinking in Hz. > > My gut feeling, though, is that it's not that simple; that is, that >I don't perceive pitch as a simple log-frequency scale. No, it isn't. The human ear seems to be most sensitive to pitch changes (where "pitch" is a log scale) in the midrange of, say, 200-4000Hz. It is also most sensitive to absolute volume and to volume variations in this range. Probably not coincidentally, this is also the range necessary for the intelligibility of speech. There are several books on the subject of acoustics and music that go into this sort of thing in depth (although a few mysteries still remain). I'd post a reference, but all my books are sealed in boxes ready to move to my new home... -Ed Hall edhall@rand.org