back to list

integer detectors, Tenney complexity, mclaren, etc.

🔗Carl Lumma <carl@...>

8/4/2001 12:58:21 AM

~~~~~~~~~~~~~~~~
Integer Detector
~~~~~~~~~~~~~~~~

>>/.../ mythical "integer ratio detectors" inside
>>people's heads (which have been shown not to exist).
>
> "Not to my satisfaction, unless you've got something
>I haven't seen." -- Carl Lumma
>
> "The experiment does not support the hypothesis that the
>human ear is provided with some sort of frequency-ratio
>detector." [Plomp, R., W. A. Wagenaar and A. M. Mimpen,
>"Musical Interval Recognition with Simultaneous Tones,"
>Acustica, Vol. 29, 1973, pg. 101]

I'll have to look that one up. If you've previously
posted a larger excerpt from this paper, just let me
know the message number or approx. date.

> Carl Lumma also ignored the following result, which I
>also posted previously on this forum:
>
> "An experiment on the perception of melodic intervals by
>musically untrained observers showed no evidence for the
>existence of `natural' categories for musical intervals." [E. M.
>Burns and W. D. Ward, "Categorical Perception--Phenomenon or
>Epiphenomenon: Evidence from experiments in the perception of
>melodic musical intervals," J. Acoust. Soc. Am., Vol. 63, No.
>2, 1978, pp. 456-468]

That's _melodic_ intervals. This may be relevant to Boomsliter
and Creel's work, but not to my discussion of an integer detector,
which pertained to harmonic intervals only.

> "Our experimentation verified that `perfect' consonances
>are not a constituent of a Central African concept of the scale.
>These musicians do not judge a strict octave (1200 cents) to be
>better than a large major seventh (1150 cents) or a small minor
>ninth (1250 cents). On the contrary, the Banda Linda musicians
>prefer the small `octave' (1150 cents) in any register, probably
>because of the roughness it creates on the octaves that are
>always played simultaneously with double sticks in each hand."
>[Voisin, Frederic, "Musical Scales in Central Africa and Java:
>Modeling by Synthesis," Leonardo Music Journal, Vol. 4, 1994,
>pg. 89]

One could say (a) the effects of roughness interfere with those
of our hypothetical integer detector mechanism for these inharmonic
timbres and/or (b) this is a style choice; the existence of an
integer detector does not mean our music must 'satisfy' it.

> If humans had a built-in `integer detector' in their ear/
>brain system, all peoples everywhere would have perfect consonances
>as a constituent of their concept of the musical scales.

Doesn't follow. See (b) above.

> "MEG is now widely used to record the responses of the auditory
>cortex to sound stimuli. This technique is based on the fact that
>changes in neuronal potentials across cell membranes are accompanied
>by the flow of currents, and current flows induce magnetic fields.
>(..)
> "The authors failed to find a single neuron that responded
>to a harmonic series in a manner closely similar to the way it
>responded to the fundamental frequency of the harmonic series used.
>They concluded that pitch is not likely to be represented in the
>primary auditory cortex and other adjacent fields that they
>studied." [Weinberger, Norman M., "Music and the Auditory System,"
>in "The Psychology of Music," ed. Diana Deutsch, Academic Press:
>1999, pg. 67]

This incomplete quote could just as easily support an integer detector
hypothesis... only the fundamental of a complex tone is being passed
to the level that the authors were looking at (for example).

> The documented fact that pitch is mapped in the human
>tonotopically according to logarithmic pitch rather than frequency
>(op cit.) means that frequency ratios get lost when they are mapped
>into the human brain. As a result, there cannot exist frequency
>ratio detectors inside the human brain, since once converted to
>logarithmically arranged points of excitation within the brain,
>no ratio information between the notes of dyads remains...only
>information on the logarithmic pitch distance twixt the two tones.

Sounds promising, but I'm not 'choosing to ignore' anything by stating
that your excerpts are not satisfactory to prove what you claim is
proven... they are not even enough to convey complete ideas. The reader
of your posts is left with nothing but your frothing word, Brian.
If you're going to spend the bandwidth, might as well try and explain
the ideas you use to back your arguments.

>It stands to reason that Carl Lumma cannot provide us with
>such a citation, since he has not studied the psychoacoustic
>literature in sufficient detail to do so.

That's accurate, more or less. I don't have expertise in
psychoacoustics. I don't see evidence that you do, either,
Brian. In any case, if my claims are so patently false,
show me, or tell me what papers I should read. I'll get
the Weinberger and Plomp papers...

~~~~~~~~~~~~~~~~~
Tenney complexity
~~~~~~~~~~~~~~~~~

> First, Carl Lumma merely claims that Tenney's metric
>"has been found to agree fairly well with something people
>can hear." That is completely different from the claim
>that Tenney's metric involves "frequency ratio detectors"
>inside the human brain, which Carl was presumably trying
>to prove.

Nope. I never said anything about Tenney's metric having
to do with the integer detector hypothesis. It's a
completely different topic, as you seem to agree.

>But the most important point is debunking James
>Tenney's musically meaningless mathematical calisthenics
>(the so-called "Tenney dissonance" metric) is the fact
>that neither James Tenney nor Carl Lumma nor Paul Erlich
>seems to recognize the vast chasm which separate
>acoustic roughness (which Tenney's metric purports to
>measure, but actually doesn't) and M*U*S*I*C*A*L
>dissonance.

Bzzz. Tenney complexity does not purport to measure
roughness. We've speculated it has to do with the
periodicity mechanism (it looks a lot like harmonic
entropy over the rationals, as I mentioned), but all
I said was that it corresponds to something people
(like Tenney, Genovese, Myself, Breed, Erlich, Keenan,
Pehrson) seem to be able to hear.

> There exists an example using major sevenths which
>sound smooth when played with the correct Csound timbres,
>while a perfect 2:1 octave sounds unbearably rough,
>on track 14 of my CD "Introduction To Microtonality."
> The same track contains Csound examples showing
>a diminished fifth dyad which sounds smooth, while
>the 3:2 interval played with the same timbres sounds
>unbearably rough, and an example showing a 12-equal
>minor seventh interval played as a dyad which sounds
>smooth, while the 2:1 octave sounds unbearably rough.

Right- the roughness obscures the effect. I've got the
CD, you know.

> Naturally Carl Lumma was too arrogant to bother to
>listen to my CD and consequently too ignorant of
>such simple and obvious counterexamples to realize that
>his claim about the debunked and pervasively false
>and musically useless metric known as "Tenney
>dissonance" was systematically false.

All wrong.

~~~~~~~~~~~~~~~~~~~~
Boomsliter and Creel
~~~~~~~~~~~~~~~~~~~~

>>"Where in that quote, Brian, are you getting difference
>>tones?"
>
> If Carl Lumma were knowledgable of Boomsliter &
>Creel's theory of the auditory system, he would
>realize that Boomsliter & Creel model the human
>auditory system's frequency detection mechanism
>on difference tones -- a methodology now known to
>be incorrect.

I've read every extant paper by Boomsliter and Creel,
creating, at Erv's request, what may be the most
complete library of Boomsliter and Creel materials on
the planet, distributing parts back to Erv and Chalmers...

You'll have to show me, Brian, where difference tones
have anything to do with extended reference. As I
said, Boomsliter and Creel made many different guesses
as to how the ear/brain was getting the preference
of small just ratios... none of which were compelling,
considered anything more than speculation by their authors,
and none of which have anything to do with their main
thesis.

Instead, the one paragraph you've actually produced
about difference tones and extended reference was
total gibberish... like your stuff about "5-limit JI"
and adaptive tuning, it smacked of bluffery.

>>>There is not one shred of evidence for the false claim
>>>that singers or string players "naturally play in just
>>>intonation,"
>
>"You never said what type of JI you meant -- melodic or
>harmonic -- and you cussed me out when I asked. At the
>time, we were discussing adaptive tuning, which deals with
>harmonic intonation. Now we're discussing B&C, who deal
>with melodic intonation. Which, or both, is it?"
>
> The classic logical fallacy of the irrelevant
>objection.

My comment was a question, not an objection.

> JI is JI. No matter whether you retune chords and
>consequently warp and twist melodic pitches out of melodic
>tune, or you retune melodic pitches and consequently
>warp and twist vertical pitches out of tune...one way
>or the other, ya gotta warp and twist something out of
>tune somewhere, either melodically or harmonically.

"Warp" melodically "out of tune", compared to 12-tET, you
mean. Unlike you, I tend to like comma shifts in melodies.
Like David Doty, I'd call them feature, not a bug.

But let's consider what you're calling an irrelevant
objection... would you call meantone "out of tune"?
Would you call it JI? It has just major thirds... it has
some degree of vertical JI. What about JdL's adaptive
temperament? It can hide melodic "warping" very well
in most cases -- in fact, beyond the ability of even the
best ears to notice it, on balance.

> Listening experiments have proven conclusively
>that warping and twisting melodic pitches out of tune
>sounds more out of tune than warping and twisting harmonic
>pitches in chords out of tune...

Experiment_s_? Only the Rasch seems pertinent here, and
it has never been duplicated. Also, "more out of tune"
is quite a jump from "more objectionable", which is what
the experiment actually tries to measure. Finally,
"objectionable" doesn't mean very much in this experiment...
it isn't a scientific quantity as Rasch uses it. The
testimonial evidence I've seen to the contrary is enough
to convince me that more is going on here than Rasch
reports.

-Carl