source file: m1413.txt Date: Mon, 11 May 1998 18:03:13 -0400 Subject: RE: 1% From: "Paul H. Erlich" >>According to Goldstein, the precision with which frequency information is >>transmitted to the brain's central pitch processor is between 0.6% and 1.2% >>within a certain optimal frequency range. Goldstein, J. L. 1973. "An >>optimum processor theory for the central formation of the pitch of complex >>tones." J. Acoust. Soc. Amer. Vol. 54 p. 1499. The roughness component is >>somewhat more permissive as to mistuning, but a 1% standard deviation would >>cause the >>accuracy to fall below 1/2 at around the point where benign beating begins >>to gives way to the roughness effects of the critical band. >..Just wondering what this means. It deals with mistuning plus or minus >0.3% - 0.6% of a dyad's logorithmic size? No? What then? And what does >it say can percieve about an interval so mistuned? No, 0.6% means a frequency change of 0.6%. Basically Goldstein dealt with the way complex tone pitch (virtual pitch) was inferred from sets of pure tones which departed by various amounts from an exact harmonic series. His experimental results were very well fit by a model with only one parameter, sigma, which could vary from listener to listener. Essentially the model states that the listener will interpret the set of pure tones as a certain virtual pitch if the pure tones have an rms deviation less than sigma from a harmonic series above that virtual pitch. As the deviations are increased further, the ear shifts to a different, lower virtual pitch, whose harmonics better fit the pure >tones (this is a very rough description).