|
I would like to add the observation that humans tend to either focus their attention to the frequency domain (analytic mode of perception, spectral pitch) or the time
domain (synthetic mode of perception, repetition/virtual pitch). I have tried to build a model that predicts the balance between the two, see chapter 5 of my PhD: https://alexandria.tue.nl/repository/books/304393.pdf A similar duality can be observed in HiFi audio where we tend to either focus on localization or immersion (non-localization), see my publication in the American scientist
issue of jan/feb2025: January-February - 2025 - Volume: 113 - Number: 1 | American Scientist Also readable on my web site: 2025americanScientistBeerendsEverdingenFinalPaper.pdf John Beerends From: AUDITORY - Research in Auditory Perception <AUDITORY@xxxxxxxxxxxxxxx>
On Behalf Of Richard F. Lyon Doug, thanks for your comments. I was thinking more of the physical correlates of what we call pitch (your "semiotic First", which isn't a framework I've ever paid attention to). My point is just that the percept of pitch is better explainable in terms of repetition
than in terms of sinusoid frequencies (Fourier components). I'm also "biased" here of course, as a long-time advocate of this viewpoint, which is not the default, but also not radical. There is a long history of back-and-forth between these viewpoints, going back to Ohm and Seebeck, which is recounted
in many places, including in my book Human and Machine Hearing. Fourier analysis, or characterization of signals and systems in terms of sinusoids, is very appropriate in the context of linear systems, but very limiting and sometimes misleading when applied inappropriately to nonlinear systems such
as the auditory system. The conflation of pitch (measured in Hz, typically) and fundamental frequency (also measured in Hz) is usually not a big deal, like when discussing signals that are actually periodic and have a fundamental component, but can lead to
confusions otherwise. I think the notion that we "reconstruct missing fundamentals" is part of that confusion. We agree, I expect, that the pitch of a periodic signal with a missing fundamental is the same as the repetition frequency (in typical cases, at least),
which would be the frequency of the fundamental if it were present. But we don't need to "reconstruct" that sinusoid, or measure its frequency, or acknowledge its presence or absence, to get at the pitch. In fact, most modern pitch estimation algorithms use an auditory-model-like analysis structure and arrive at the answer without any kind of fundamental reconstruction; they're just looking at how much time shift it takes for the output
of the nonlinear analysis to nearly match itself. No frequency domain needed. This method generalizes well to repeated noise bursts and other aperiodic signals that have a perceived pitch, whether they have discrete sinusoidal components or not. Anyway, I still need to read Adam's paper to understand where he's heading. Maybe we'll discuss more later. Dick On Sat, Apr 12, 2025 at 9:26 PM Douglas Scott <jdmusictuition@xxxxxxxxx> wrote: Hi Dick I would have thought the opposite: Pitch isn't usually perceived as a frequency in the sense of a rate of repetition, but rather as a semiotic First (i.e. a direct perception). This would be analogous to the way colour is perceived versus
the frequency of the light that produces the sensation, which is even more imperceptible directly. This is how one can "spoof" perceptual organs with false colour and reconstruct missing fundamentals. The Fourier transform, meanwhile, is explicitly about repetition of the circular motion of various types of circles. It is thus inherently a semiotic Second, while assigning a pitch name or class would be a Third. Of course, these concepts are fluid, so constructing an interpretation in your terms is also entirely possible, but I don't think it's the default way of seeing it. Perhaps I'm biased. Doug On Sat, 12 Apr 2025 at 01:31, Richard F. Lyon <0000030301ff4bce-dmarc-request@xxxxxxxxxxxxxxx> wrote: Adam, Thanks for bringing us back to topics more germane to the list. (While I agree there's a lot that needs to be discussed about the current state of craziness in the U.S., I also agree that this list is probably not so great a place to take
that on. I realize I'm speaking from a position of almost-retired and already signed up for Social Security, so not as much affected as many. When I was out on the "Hands Off" protest march, Auditory was not on my mind.) I took a glance at your paper, and it looks to me like you didn't do a clear enough job of distinguishing different notions of frequency (maybe I need to read more). Pitch, though a perceptual concept, is very closely related to the concept
of frequency as repetition rate. In physics and math, however, frequency is often thought of as a parameter of a Fourier transform, more related to sinusoids, or circular motion, than to repetition. Your paper barely mentions pitch, but I think it's a concept
that might help make some of your points better. I look forward to reading more of it. Dick On Fri, Apr 11, 2025 at 9:31 PM Adam Weisser <adam_weisser@xxxxxxxxxxx> wrote: Dear all, I would like to share with you my latest piece on the topic of frequency. While it is not strictly within auditory science, this work would have never materialized if it weren't for the deeply-ingrained
approach to acoustic signals that has time and frequency as more or less independent dimensions - two attributes of sound that are interrelated, but which we generally perceive as essentially different. Contrasting this understanding with the fundamental definition
of frequency in physics - the reciprocal of the period - leaves much to be elucidated and is duly riddled with paradoxes. In this manuscript, I review the different instantiations of the concept of frequency in physics, engineering, mathematics, and perception, partially tracing their historical evolution, in attempt
to answer the question of whether frequency can be counted as a separate dimension of reality, in addition to space and time. The resolution ties together time, frequency, and determinism in a highly counterintuitive manner: While I have no illusions regarding how this text (and its author) may come across given the extraordinary claims put forth, I have tried to go about it in the most rigorous way I could, while
methodically scrutinizing some long-held dogmas in the sciences, in hope of sparking further discussion, somewhere down the road. For what it's worth, I'm indebted to what I have learned (or maybe, mislearned) by virtue of being part of the auditory community. Yet, these days appear to be challenging to many members of
the Auditory List. I once had a good colleague (PKR) who lamented how we were never trained in "political acoustics" in engineering school, that would have endowed us with the necessary skill set to be able to juggle between the various demands of our whimsical
bosses. Even today, still nobody teaches this subject, and people are expected to figure out this vital skill as they go along, while trying to survive and stay true to what they were originally hired to do. It's perhaps ironic, because in every other respect,
hearing science must be one of the most unifying and all-encompassing of all human endeavors, being positioned in the nexus between physics, biology, psychology, neuroscience, perception, communication, language, music, engineering, medicine, environment,
architecture, computer science, mathematics, and many other disciplines. All this is in opposition to politics, which usually excels in sowing division, whether we are ready to get directly involved with it or not. I hope that we can transcend these unpleasant times both as individuals and as a community and eventually get back to what we do best, which is study hearing and sound. With wishes for better times to come, And thank you for your attention and time, Adam. |