Yahoo Tuning Groups Ultimate Backup tuning Perception of Consonance (was looking for MT traditions)

The thread on microtonal world traditions (mostly maqamat) seems interesting, admittedly I haven't had time to follow it all, have only given the most recent post a cursory glance of coffee this morning. But an issue has been raised which I would like to comment on, that being the perception of interval quality.

First and foremost, I think it is absolutely pointless to talk about aesthetic connotations of certain intervals and harmonies, tempered or not. This can never be but cultural programming and/or contextual, and is probably even considerably different depending on the individual. Thus to ponder whether a C major chord/arpeggio is inherently "happy" while a C minor is inherently "sad" is a waste of everybody's time. In this sense, I am completely in agreement with the atonalists (the only point!) in that emotive content is contextual, and the composer can create the functional role he wants. Granted, he must work against the predominant cultural programming if he does not want to more or less follow its dictates.

Beyond that, we are left with the question of the perception of JI. Basically, the field of enquiry is also impossible to explore in any truly objective sense, since all subjects have been preprogrammed more or less. As pointed out, a fetus in the womb is exposed to sounds. Not that I believe the stuff about the sound of the human voice predisposing the subject to a harmonic series. It's more so that all cultures produce some sort of music, and the fetus also hears music, either the mother singing, or the sounds which arrive through the stomach walls. So all samples are tainted.

Having admitted that, we can nonetheless make some attempts at examining our perception. Basically, it boils down to hardware versus software; which types of perception are innate due to the basic perceptual equipment we come into the world with, and which types of perception must be learned, either through cultural bias or through repeated exposure to patterns (there is a slight difference between the two).

I have a lot of trouble with many studies published in the perceptual sciences literature. You usually don't have to read very far to find some elephant in the room, some glaring basic conceptual error which dooms the conclusions to partial or complete invalidity. They include things like using ET intervals instead of JI to explore why people "like" traditional harmonies, or using pure sine waves instead of complex harmonic sounds. At a more subtle level, it is often easy to imagine a hypothetical situation which will undermine the basic thesis. For example, the theories of consonance that postulate neural mode locking fail to explain why we accept slightly detuned fifths or thirds as being consonant, especially when the stimulus is two pure sine waves, when the period of repetition may be thousands of cycles. Ditto the new fad in neural science, brain scans. So what if certain centers light up with certain stimuli? Are we seeing something inherent, or merely the recognition of learned patterns? Who can tell? In any event, if anybody out there has a favorite article proposing a neural-based explanation, I'd love to give it a read. Bet'cha I can find a fatal flaw somewhere, and you usually don't have to be a rocket scientist to do so.

This is but a natural result of the "publish or perish" doctrine of modern academia. If everybody spent as much time and effort trying to imagine experiments that would DISprove their theories as they do concocting arguments for why the theory is valid, as any real scientist MUST do, probably only about 5% of the current output would every reach publication. Although this would greatly increase our knowledge, it would, however, destroy the basis for an entire universe of academic pecking order. Furthermore, it would dismantle the entire system of paid-vacation perks euphemistically referred to as "academic conferences", to say nothing of the leach-like industry of journal publication. BTW, I just learned recently that in many cases of scientific journals, the authors PAY to have their stuff published. In the recording bizz, this used to be called "Vanity Labels".

Back to perception. Personally, I'm a Helmholtz/Sethares fan, based upon my personal experience in composing (which I haven't done for 30 years), tuning keyboards professionally, making instruments and judging their sound, and performing listening experiments on music students in the classes I teach. I'll share with you all a few of these.

First off, sine wave studies are pointless. There is no natural source of sine waves, thus the ability to use them exclusively to make music is a relatively recent capability and therefore has nothing to do with what humans have been up to for thousands of years. I have also repeatedly proved to myself that harmonic perception goes out the window when sine waves are used. I play tricks on classrooms of students, things like playing them a very narrow fifth a telling them it is perfect, and then asking them it sounds perfect to them or not. Depending on the size of the fifth, you get differing numbers of students who agree, naturally the narrower the fifth, the more disagreement. However, in all cases, if you let them listen to it for about a minute, and then give them a pure fifth, they all inevitably say it is now too wide, the narrower the first fifth, the more they think the second is too wide. You can do it the other way around, starting with a wide fifth, and get the same results; they think a pure fifth is too narrow. I usually give them a dose of aural "gari" in order to clean their sonic palette between servings, i.e. 10 seconds of white noise between the fifths. Major thirds also work, though less compelling, I think primarily because you are fighting against the cultural programming of the ET third, something which different people seem more or less capable of overcoming quickly when presented with pure intervals... but that's another story.

Needless to say, when harmonics are present, none of this trickery is successful. All but the most intonationally deficient musicians (usually button-pushers, i.e. keyboard players) instantly hear that a detuned fifth is detuned, and they will not accept the suggestion that it is not so. Furthermore, they instantly recognize a pure fifth as pure, and do not say it is too wide or too narrow. So for me, this means that the presence of harmonics is essential in some way. This is the elephant in the room for all studies done with sine waves. Yeah, they find neural activity with the sound of a fifth produced by two sine waves. But do they think of testing greater and lesser degrees of detuned fifths for the same neural response? Not that I know of. Thus we have no way of knowing whether this has something to do with two frequencies with a precise ratio of 3:2 precisely because they have a simple ratio or if it is simply the recognition of a learned interval.

The other experiments I have done are most instructive. I have deconstructed the consonant and dissonant components of simple harmonies. Take, for example, two complex harmonic tones sounding a pure major third on middle c (c' - sorry, having lived in the organology world so long I have an aversion to the "scientific" note terminology and refuse to learn/use it). I've done it with various spectral contents, including a wave form extracted from a sample of me singing the vowel "Ah" at Bb below tenor c, violin and cello wave forms, etc. I now use a basic "average" sound, which has more or less equal level for the first 4 harmonics, with a slight reduction of the fundamental (as is often observable in real sounds), and from there a more or less linear reduction to zero at about the 25th harmonic. The results are similar in all cases.

The first thing to do is play the interval and the detune it, both up and down, exploring the qualitative difference in timbre. Forget about beats, just listen to the quality of the sound as though you were a composer experimenting with orchestration. Now, remove the consonant harmonic components. This can be done is several ways. One possibility is to take away all multiples of 5 from the lower tone. Another is to take away all multiples of 4 from the upper tone. You can even mix and match, removing 5 from the lower, 8 from the upper, 15 from the lower, 16 from the upper, etc. The ultimate "deconsonofication" is to remove the congruent harmonics from both notes. You will find that the compelling nature of the pure interval ceases to become compelling, and that it is far more difficult to hear the difference between the pure interval and even severely detuned versions. A difference is still there, but it is an extremely subtle variation in the quality Helmholtz called "roughness", which I think of as the "mosquito factor", which is an disturbing buzziness which irritates in a manner similar that the faint sound of mosquito in the darkness of your bedroom on a summer night irritates. At the risk of crossing the border into aesthetics, I would have to say that this quality, the more it is present, the more unpleasant the sound in some basic way, like a wool sweater with no undershirt.

The other thing to do, naturally is take away the dissonance. In the case of a pure third with none of the (normally) consonant components, you can reduce the remaining annoyance by removing one or the other of the harmonics which is pairing at a dissonance maxima according to the critical bandwidth. Instantly the sound is less bothersome. It is no more compelling or "consonant", but at least it is not so bothersome. The lower down this pair, the better the improvement.

Finally, though it hardly needs demonstrating, it is nonetheless instructive to do the opposite: reduce the interval to only its consonant components. Simply remove all harmonics except multiples of 5 from the lower tone and multiples of 4 for the upper with a pure tuning of the interval. It is quite easy to hear that the sounds are completely identical except for the fundamentals. It's fun to turn the fundamentals on and off, and hear how the identity is absolute when they are absent. When this consonant-only interval is detuned, you find the same sort of gradual transition from identical sound to two separate sounds through a region of maximum annoyance that happens with a single tone as you move through the critical bandwidth.

So I see the perception of consonance/dissonance in its most universal and fundamental sense as mix between two conflicting stimuli: unity and noise. The more consonant, the more components of unity and the less components of noise, the more dissonant, the less unity and the more noise. It's somewhat akin to signal to noise ratio in audio. A dissonant interval is like bad reception. Every inteval other than the octave has some noise. Detuning the interval introduces even more noise by converting the "signal" component itself (the normally congruent components) more or less into a noise component.

I don't think any of this has to be learned in any way, not even from exposure to harmonic vocal sounds in the womb. It is simply the combination of two "hard-wired" capabilities: the ability to resolve frequency, and the ability to recognize similarities in stimuli. In essence, it all boils down to a single question: why does separation by 25% of the critical bandwidth annoy us? Probably simply because the brain can't decide what to make of it, it's neither fish nor fowl, and it takes a lot of mental effort to no avail, since the problem is irresolvable until the interval either increases or decreases.

Since the same sort of critical bandwidth mechanism has been observed in cats, I think this is something inherent in the structure of our perceptual apparatus. Thus I believe that there is a strong argument to be made for a universal, natural basis for consonance in a purely psychoacoustic sense. What you do with that in a musical context is pretty much up to you, and the natural order of things can be rather easily "unlearned" by indoctrination, as we all know is the case with ET major and minor thirds and sixths. Nonetheless, once one becomes aware of the natural order of things, it is difficult to go back to the state of ignorant bliss, much the same way that once you start hearing the dripping faucet, it drives you crazy until you fix it, whereas before, you didn't even notice it. I now have a lot of trouble listening to any organ tuned in ET, for example, whereas for many years I happily played all manner of literature, from Frescobaldi to Messiaen, on a regular old garden variety organ. Now I instantly recognize the ubiquitous dirty acoustic haze which makes all intervals (except octaves, fourths, and fifths) sound more or less the same. Shades of gray. Ditto jazz piano, although I like jazz piano much more than ET organ, since in jazz piano the name of the game is precisely about playing with shades of gray, not pretending they don't exist. Once again, this rediscovery is neither learning nor programming, it is in fact unlearning and deprogramming, just becoming aware of a natural state which you had either forgotten or never noticed, just as you can suddenly become aware of the sound of your own breathing, or the blood rushing through your body with each heartbeat.

Me `umble opin-ee.

Ciao,

🔗Michael <djtrancendance@...>

Paul>"Thus to ponder whether a C major chord/arpeggio is inherently "happy" while a C minor is inherently "sad" is a waste of everybody's time. In this sense, I am completely in agreement with the atonalists (the only point!) in that emotive content is contextual, and the composer can
create the functional role he wants. "

Agreed! Throw a C-major in a chord progression full of diminished chords, for example, and I've found the C major can sound tense or, rather, "out of place" in comparison. The one thing I will say, across cultures, is that playing multiple chords that are either very high limit or have high critical band dissonance will almost undeniably create a sense of tension and "vagueness/instability" for just about anyone even accustomed to any culture...but that's in fairly extreme cases. And still, that tension can be "happy" or "sad" tension depending on the person.

>"(errors in perceptual theories include) using pure sine waves instead of complex harmonic sounds"

The way around this seems to be breaking complex harmonic sounds into series of sine-waves and then analyzing the dyads between both overtones and root tones of each instrument/note played.
Which is exactly what Sethares does...

>"Ditto the new fad in neural science, brain scans. So what if certain
centers light up with certain stimuli? Are we seeing something inherent,
or merely the recognition of learned patterns?"

Exactly...time and again I've been told "no one will believe even any listening-tested theory as scientific unless backed by brain-scans" and we have no way to tell what of the brain scans are nature vs. nurture. Someone who likes a band may light up very differently when listening to their music from someone who does not like it...in that case, nurture seems to dominate.

>"Back to perception. Personally, I'm a Helmholtz/Sethares fan, based upon
my personal experience in composing (which I haven't done for 30
years), "

As am I...it's funny just how much people cherish JI over Sethares' work. In my opinion, both are very valid and important.

>"You can do it the other way around, starting with a wide fifth, and get the same results; they think a pure fifth is too narrow"

It seems the pattern is...people generally prefer "overly sharp" versions of an interval to similarly flat ones. I've noticed this myself when I give myself blind listening tests.

>"In essence, it all boils down to a single question: why does separation
by 25% of the critical bandwidth annoy us? Probably simply because the
brain can't decide what to make of it, it's neither fish nor fowl, and
it takes a lot of mental effort to no avail, since the problem is
irresolvable until the interval either increases or decreases."

My best guess is that the mind puts in the same amount of effort it takes to analyze two tones and, to an extent, only gets one tone. There is a theory of consonance somewhere that involves consonance=amount-of-information-received/effort more or less.

Another weird thing is the debate on what maximum critical band dissonance occurs at.
Sethares appears to say maximal critical band dissonance is right around the 100 cent 12TET semitone and becomes less as two intervals become closer http://sethares.engr.wisc.edu/images/image3.gif, however people on here seem to think it is much more like 50 cents. What do you think between these two and why?

>"Nonetheless, once one becomes aware of the natural order of things, it is difficult to go back to the state of ignorant bliss, much the same way that once you start hearing the dripping faucet, it drives you crazy until you fix it, whereas before, you didn't even notice it."

Exactly...such has also been proven in cases of permanent Tinnitus (ringing in the ears): just about the best treatment for it, as of now, is to teach people to ignore it.

🔗Paul <paul@...>

--- In tuning@yahoogroups.com, Michael <djtrancendance@...> wrote:
>

>
> >"(errors in perceptual theories include) using pure sine waves instead of complex harmonic sounds"
>
> The way around this seems to be breaking complex harmonic sounds into series of sine-waves and then analyzing the dyads between both overtones and root tones of each instrument/note played.
> Which is exactly what Sethares does...

Which is what I do as well.
>
>
> >"Ditto the new fad in neural science, brain scans. So what if certain
> centers light up with certain stimuli? Are we seeing something inherent,
> or merely the recognition of learned patterns?"
>
> Exactly...time and again I've been told "no one will believe even any listening-tested theory as scientific unless backed by brain-scans" and we have no way to tell what of the brain scans are nature vs. nurture.

Yeah, brain scans are the new flavor of the month in perceptual awareness. Note that it is also very convenient for the researchers, since if you ain't got the fancy-pants equipment, you can stress-test the theory. But just ignore these types, as they will ignore you. They are not interested in anything except getting invited to the next conference at the next exotic location.

;-)

>
> >"Back to perception. Personally, I'm a Helmholtz/Sethares fan, based upon
> my personal experience in composing (which I haven't done for 30
> years), "
>
> As am I...it's funny just how much people cherish JI over Sethares' work. In my opinion, both are very valid and important.

Well, Sethares IS JI, except he demonstrates the basis of JI (which of course is what Helmholtz did 150 years ago). Trad JI is only valid with harmonic spectra. His greatest contribution is the demonstration that warping the gamut to fit a non-harmonic spectra (I hate calling a gamut a scale) makes a new kind of xenharmony. That is the logical outcome of Helmholtz.

That said, Sethares goes wrong on two counts:

(1) His demo pieces of trad harmony should have been done in JI, since 12tet is already a departure from his own theory.

(2) His articles/chapter on temperament are misguided, since traditionally a temperament is a compromised gamut, and band-aide solution, a less than ideal, and his whole shtick is that the spectrum dictates the gamut. I think you can no more reverse-engineer the structure of a trad temperament by analyzing music supposedly written in that temperament (how do we know?) than you can write good music by feeding a temperament into a computer.

>
> >"You can do it the other way around, starting with a wide fifth, and get the same results; they think a pure fifth is too narrow"
>
> It seems the pattern is...people generally prefer "overly sharp" versions of an interval to similarly flat ones. I've noticed this myself when I give myself blind listening tests.

Is that with sine waves? Personally, I just find my judgement is rudderless with sines, neither wide nor narrow in general, just lost...
>
> My best guess is that the mind puts in the same amount of effort it takes to analyze two tones and, to an extent, only gets one tone. There is a theory of consonance somewhere that involves consonance=amount-of-information-received/effort more or less.
>
> Another weird thing is the debate on what maximum critical band dissonance occurs at.
> Sethares appears to say maximal critical band dissonance is right around the 100 cent 12TET semitone and becomes less as two intervals become closer http://sethares.engr.wisc.edu/images/image3.gif, however people on here seem to think it is much more like 50 cents. What do you think between these two and why?

I agree with the 50 cents, but it is different in different registers, a low frequencies even a fifth can be a dissonance. Maybe Sethares is illustrating the situation at around 200 Hz, which is pointless for harmonic/overtone analysis. Here's my graph I use in my classes, made by running a bunch of frequencies through the Plomp/Levelt formula (in Catalan, sorry, but you get the idea):

http://www.polettipiano.com/MaxDis,jpg

>
> >"Nonetheless, once one becomes aware of the natural order of things, it is difficult to go back to the state of ignorant bliss, much the same way that once you start hearing the dripping faucet, it drives you crazy until you fix it, whereas before, you didn't even notice it."
>
> Exactly...such has also been proven in cases of permanent Tinnitus (ringing in the ears): just about the best treatment for it, as of now, is to teach people to ignore it.
>
Tell me about it. As a kid I hunted and played in a rock band. Now I'm paying for it.

Ciao,

Perception of Consonance (was looking for MT traditions)

🔗Paul <paul@...>

🔗Michael <djtrancendance@...>

🔗Paul <paul@...>

🔗Carl Lumma <carl@...>

🔗Paul <paul@...>

🔗Paul <paul@...>