source file: mills2.txt Date: Mon, 16 Oct 1995 08:42:42 -0700 From: "John H. Chalmers" From: mclaren Subject: Tuning and psychoacoustics - post 21 of 25 --- Throughout this series of posts, we have seen that the ear/brain system is not simple. As is now clear, the behavior of the auditory system when presented with complex tones is different than that predicted by Ohm/Helmholtz: "Helmholtz's working hypothesis has been put aside by later investigators, both those who worked in music and those who worked in psychoacoustics. Several reasons for this can be given. First, before the introduction of electroacoustic means of tone production and control in the 1920s, it was not possible to carry out the necessary psychoacuostical experiemnts, while Helmholtz's observations proved insufficient in many ways. Second, it turned out that musical theory has its own rules apart formm the perceptual relevance of the characteristics of the sounds it makes." [Rasch, R. A., and Plomp, R., "The Perception of Musical Tones," in "The Psychology of Music," ed. Diana Deutsch, 1982, pg. 21] But the behavior of the ear/brain system is also different from that predicted by Seebeck, Stumpf and Schouten: "It sounds obvious if we say that we hear a note a' if the fundamental of the frequency is 440 c.p.s. But what happens if we remove this fundamental by electrical means, leaving on ly the harmonics with frequencies of 880, 1320, 1760 cps, etc? Or if we take away the fundamental 440 cps and the second harmonic 880 cps? We learn from experiments that the perceived pitch level remains the same: a'. One may take away many of the lower harmonics without altering this. If this `mutilated' note is interrupted for only an interval of a second, the sensation is completely altered. Instead of the `residual tone' on a' we now hear another pitch which lies approximately in the region of the strongest remaining harmonics and is called `formant pitch.'" [Werner Meyer-Eppler, "Statistic and Psychologic Problems of Sound," Die Reihe, Vol. 1, No.1, 1955, pg. 55 (English edition, 1957)] This is the simplest and most compelling example of the contradictory behavior of the ear/brain system. The ability of the ear to extract an unambiguous fundamental frequency from a complex tone argues strongly in favor of the Ohm/Helmholtz hypothesis of the ear as a Fourier analyzer; but the ability of the ear to still hear a "residual" fundamental even after the fundamental and lower harmonics have been removed electronically cannot be explained by the ear-as-Fourier-analyzer model of hearing. Instead, "beats of mistuned consonances and periodicity pitch can be perceived even if each component tone is fed dichotically to a different ear. In such a case, of course, the complete vibration pattern never arises--what must arise is a superposition or interaction of the neural signals from both cochleas after they have been combined at the medullar or midbrain levels. (...) The fundamental neural message is given by the rate and the distribution in time with which individual impulses are fired along the axon." [Roederer, Juan, "The Physics and Psychophysics of Music," 1973, pg. 45] Thus the ability to hear such residual pitch argues against the Ohm/Helmholtz Fourier analyzer model of hearing, and for the Seebeck/Stumpf model of the ear as time-domain autocorrelator in which "the actual time distribution of [auditory nerve] impulses codes the information on repetition rate or periodicity pitch (see below)." [Roederer, Juan, "The Physics and Psychophysics of Music," 1973, pg. 45] However, the fact that after a brief interruption the ear hears exactly the same tone from which fundamental and lower harmonics have been electronically removed as having an entirely different pitch argues strongly against BOTH the Ohm/Seebeck frequency-domain and the Seebeck/Stumpf time-domain model of the ear, and instead supports the Fetis/Burns/Ward model of the ear as an adaptive system molded by learned responses. Thus in one simple experiment we have compelling evidence both for and against all 3 major theories of hearing. Other compelling evidence for the "contextual" behaviour of the ear/brain system abounds. "...it is to be emphasized that sound elements which are juxtaposed in time can have the effect that identical physical vibration procedures give rise to totally different sensations. The phenomenon has been particularly observed in the case of synthetic explosive sounds such as `p',`t', `k' which may be perceived in a totally different manner, depending on the vowels which are juxtaposed to them (3). To explain this one cannot attribute it to masking which has already been known for a long time becuase the influencing is effected by the following vowel (regressive dissimilation) as well as by the preceding." [Werner Meyer-Eppler, " Statistic and Psychologic Problems of Sound," Die Reihe, Vol. 1, No.1, 1955, pg. 55 (English edition, 1957)] Further evidence of complex phenomena possibly produced by interaction between the ear's frequency-domain and its time-domain processing functions was brought forward by Strong and Clark. "...in order to evaluate the relative significance of spectral and temporal envelopes, [they] resorted to an interesting process: they exchanged the spectral and temporal envelopes among the wind instruments and asked listeners to attempt to identify these hybrid tones. The results indicated that the spectral envelope was dominant if it existed in a unique way (as in the oboe, clarinet, bassoon, tube and trumpet); otherwise (as in the flute, trombone, and French horn), the temporal envelope was at least as important." [Risset, Jean-Claude, "Exploration of Timbre by Analysis and Synthesis," in "The Psychology of Music," ed. Diana Deutsch, 1982, pg. 36] Some results are simply inexlicable by any of the above models: "Doughty and Garner (1948)...concluded that pitch was unchanging for tones of 25 msec and longer, but that 12-msec and 6-msec tones have a lower pitch. However, Boomsliter et al., (1964) emphasize that the transition from "click" to "tone" depends on the intensity. Swigart (1964) has reported a perhaps related phenomneon for repeated short bursts of tone. If one presents successive 8-msec bursts of 1000-Hz tone with 1-msec pauses between (i.e., if one cuts out every ninth cycle), the pitch is significantly lower than that of a continuous 1000-Hz tone. Just why, however, is still unclear." [Ward, W.D., "Musical Perception," in "Foundations of Modern Auditory Theory," ed. J.V. Tobias, Vol. 1, pg. 429.] "An aspect of pitch perception that is still regarded as somewhat mysterious despite a goodly amount of experimentation is "aboslute" (or "perfect") pitch." [Ward, W.D., "Musical Perception," in "Foundations of Modern Auditory Theory," ed. J.V. Tobias, Vol. 1, pg. 429.] "In 1973 David M. Green published some interesting results on temporal acuity. He measured the ear's ability to discriminate between two signals that have different waveforms but the same energy spectrum. An example of such signals is any short waveform and the same waveform reversed in time, such as those in figure 10-5. Here Part A is a sound of decreasing frequency, Part B is a sound of increasing frequency. Green found that the ear can tell the difference between two such waveforms if their duration is greater than 2 milliseconds." [Pierce, J.R., "The Science of Musical Sound," 2nd ed., 1992, pg. 149] "On the hypothesis that critical band filters can be identified with auditory nerve filters, the model of Zwicker and Scharf was tested by Pickles (1983). (...) The results agreed with the psychophysical data, in that the summed activity increased with stimulus bandwidth, for wider timulus bandwidths. (Fig. 9.13 B). However, there was no clear sign of a flat portion in the function at narrow bandwidths. The reason for this is not known..." [Pickles, James O., "An Introduction to the Physiology of Hearing," Academic Press, 2nd ed., 1988, pg. 283] No model of the ear/brain system convincingly explains any of the above results, or for that matter one of the best-known quirks in musical perception: perfect pitch. Musical aesthics is yet another abyss which has swallowed many a psychoacoustic researcher. "Why some vibration patterns appear more "beautiful" than others is not known. A great deal of research (good and bad) has been attempted, for instance, to find out what physical characteristics make a Stradivarius violin a great instrument. Many of these characteristics are dynamic in character, and most of them seem more related to the major or minor facility with which the player can control the wanted tone "color" (timbre), than to a "passive" effect on a listener. To a large extent the impression on the listener is based on learned experience..." [Roderer, Juan, "The Physics and Psychophysics of Music," 1973, pg. 134] But by far the most striking gaps in current knowledge of the ear/brain system involve the question of how the human auditory performs 3-dimensional sound localization. A Head-Related Transfer Function can be measured empircially for each individual by means of microphones placed in the ear canals and an extremely high-speed computing engine, but to date no one has offered a general mathemtical model which predicts the HRTF, or the location of a specific sound using that mathematically- modelled HRTF. The best example of this complete gap in our psychoacoustic knowledge is the inadequacy of current stereo speakers. Even a 6-year-old child easily detects the difference between a recording reproduced from digital media on high-quality stereo speakers, and a live performance: but to date no one has succeeded in formulating a mathematical model of hearing which details the difference in hard numbers. [See Moeller, H. K., Soerenson, M. F., Hammershoei, D., and Jensen, C. B, "Head Related Transfer Functions of Human Subjects," J. Audio Eng. Soc., 43(5), 1995 May, pp. 300-321; also see Appleton, J., "Machine Songs III: Music In the Service of Science-- Science in the Service of Music," Computer Music Journal, 16(3), 1992,pp. 17-21] Clearly, there remain many unexplained aspects in the behavior of the ear/brain system. The next post examines the mass of evidence gleaned from the many psychoacoustics results confirmed and supported by a wealth of modern. Although the modern psychoacoustic data is complex, it does support some conclusions. These will be given in the next post, after which the various biases and prejudices of the psychoacoustic researchers themselves will be examined...with an eye to determining how much or how little their prejudices affected the conclusions each major researcher drew from hi/r research. --mclaren Received: from eartha.mills.edu [144.91.3.20] by vbv40.ezh.nl with SMTP-OpenVMS via TCP/IP; Tue, 17 Oct 1995 01:35 +0100 Received: from by eartha.mills.edu via SMTP (940816.SGI.8.6.9/930416.SGI) for id QAA13399; Mon, 16 Oct 1995 16:34:38 -0700 Date: Mon, 16 Oct 1995 16:34:38 -0700 Message-Id: <199510162332.AA046446328@athena.ptp.hp.com> Errors-To: madole@ella.mills.edu Reply-To: tuning@eartha.mills.edu Originator: tuning@eartha.mills.edu Sender: tuning@eartha.mills.edu