A coming change

Comments and questions about AP Avenue.
aruffo
Site Admin
Posts: 1692
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Post by aruffo » Mon Nov 21, 2005 7:01 am

A quick look-- still only a quick look, since I'm rushing off now to rehearsal-- suggests that I do disagree (in some ways) with your application of the terms "relative pitch" and "absolute pitch". I'll have to explain later; I'm just writing this little bit as a marker to remind myself about it. In the meantime, I wonder if you've read the first part of the Secret of Lesson One...

Guest

Post by Guest » Mon Nov 21, 2005 9:15 am

When I go there, I'm missing some pictures on that page. Can you please take a look and fix?

Thanks,
Lou

aruffo
Site Admin
Posts: 1692
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Post by aruffo » Mon Nov 21, 2005 10:24 am

Eep! I wonder where those pictures went? I'm not at a place where I can fix em right now... but I can tell you that the pictures that are missing are just my speculation about why octave equivalence exists. There is the visual coincidence of 2-to-1 frequencies "ending at the same point", but that doesn't prove anything.. I created this page when I was still at Phase One and (almost) purely in the speculative phase.

Anyway, I have a few minutes, so let me turn the conversation away from my unthinking knee-jerk response and back toward pitch stuff.

The study referenced on that page (the "Secret of Lesson One") indicates that children perceived pitches to be "brighter" or "darker". You wrote

Once the child learns "C" the instructor can say, "ok now this is also C, but up an octave." What is that? It's RELATIVE PITCH!


but this rhetorical question is not necessarily rhetorical. In the We Hear and Play system, a new octave is introduced by telling the student "red, but a brighter red." That isn't "up an octave"; it's not even "up" or any direction at all. It is a different quality of the same pitch. I myself have been successful at getting a non-musical adult to sing an octave interval by asking him to sing the same pitch but to make it "thinner". To drive it back to the central point-- it seems more likely to me that a child learning pitch sounds would naturally perceive each pitch to be equivalent regardless of the octave, due to its shades of "brightness"; instead of having to gradually learn that these pitches were the same, they would instead have to be trained to understand that the pitches were different.

Absolute listeners do not perceive "height".

However, I have to agree with you one hundred percent that normally, your rhetorical question is most definitely rhetorical. Children are definitely trained to think that "this is also C, but 'up' an octave" and therefore are trained away from the absolute perception which reveals that a C is a C in any octave. I was explicitly told that when I was learning Suzuki piano as a littleun (with a "wandering note" toy).

Researchers recently have manipulated spectral envelopes so that the same chroma can be only one semitone "higher" and still be the same pitch. This suggests to me that octave equivalence is itself an unnecessary limitation; that being free to perceive the chroma separate from the octave designation allows the listener to perceive not merely eight or nine octaves but an infinite spectrum of "height".

Bass Geek
Posts: 10
Joined: Fri Nov 18, 2005 11:07 pm
Location: NYC

Post by Bass Geek » Mon Nov 21, 2005 10:32 am

LouMan5 wrote: Many people with perfect pitch still make octave mistakes ... Once the child learns "C" the instructor can say, "ok now this is also C, but up an octave." What is that? It's RELATIVE PITCH!


Interesting - I've never thought of octave recognition as relative pitch ... because there is so much sameness between the notes, at least to my ears. But this clears something up for me: When I have played around with pitch recognition with my young piano students, I have often played C's in many octaves, and asked them "Do you hear how these are all similar, how they are all C?" And the usual response is a blank look of confusion, or a definite "no" from the braver students. This confused me before, but now it makes more sense.

I also had a theory student once who was a very talented organist with "naturally" acquired AP. (She said that most of her life she just assumed everyone could hear distinct pitches). Anyway, not only could she not determine octaves, but she couldn't even tell which note was higher or lower than another! This presented great difficulties in RP testing - she could identify what chord type she heard, but not the inversion ... and of course, even intervals were inverted a good deal of the time. This kind of AP actually seems to me to be counterproductive to interpreting music. If you can hear the pitches, but have no idea of spatial relationships, aren't you only getting a very fuzzy picture of the whole?

etaxier
Posts: 136
Joined: Tue Oct 04, 2005 1:57 am
Location: NYC

Post by etaxier » Mon Nov 21, 2005 4:18 pm

Aruffo said:
Researchers recently have manipulated spectral envelopes so that the same chroma can be only one semitone "higher" and still be the same pitch.


I read about this elsewhere on the site, and I still find it rather fascinating... do you have any mp3s or links to recordings of this phenomenon?

aruffo
Site Admin
Posts: 1692
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Post by aruffo » Mon Nov 21, 2005 5:13 pm

I do have recordings of that somewhere on line, here.. let's see if I c'n find it...

Ah! And, in the meantime, I am reminded of why I compiled the research pages in the first place. Check out Separating pitch chroma and pitch height in the human brain. My summary of this paper (from Phase 11), which is where the files came from:

The authors manipulated the spectral envelopes of tones so that their subjects heard either changes in pitch chroma or changes in pitch height, but not both. Their results showed that each quality was mapped to a different area of the brain: "...chroma change is specifically represented anterior to primary auditory cortex, whereas height change is specifically represented posterior to primary auditory cortex." The authors suggest that this might indicate tone height as a function of auditory scene analysis, while tone chroma may be described as an "information stream."


Let's see now.. where are those darn files..? Ah! Yes, here they are.

etaxier
Posts: 136
Joined: Tue Oct 04, 2005 1:57 am
Location: NYC

Post by etaxier » Mon Nov 21, 2005 8:40 pm

Thanks for that link. Maybe you meant "octave" instead of semitone when you said, "the same chroma can be only one semitone 'higher' and still be the same pitch." That's probably what confused me.

Maybe it's been APB or just my own training, but I heard each of those sounds as being in the same octave and instead having different timbres. If you listen to a bassoon and then a horn play the same note, you'll sometimes hear the same illusion! It's still a very interesting point, that timbre plays such a huge role in the way we process pitch 'height' (or luminosity, or whatever).

Two more interesting things: 1.
Pitch phenomena can also be explained, however, in terms of the time structure, or temporal regularity, of sounds.
Contemporary composers have used this notion to create some interesting determinant relationships between rhythm and pitch. A neat idea (and I'm sure the effects are neat), but it seems like an arbitrary relationship to me: most of the studies I've read have shown that rhythm and pitch are processed in different parts of the brain (no matter whether the physical difference between the two is merely one of frequency).

2.
my hypotheses of musical sound as movement-- octave equivalence explained as constant acceleration, melodic contour described as the movement of an implicit object, et cetera
That's my hypothesis too! but the theorist Joseph Schillinger beat us to it when he wrote, “music is capable of expressing everything which can be translated into form of motion."

In any case, i apologize for completely disconnecting from the original discussion. I'll, uh, stop doing that.

Guest

Post by Guest » Mon Nov 21, 2005 8:46 pm

aruffo wrote:In the We Hear and Play system, a new octave is introduced by telling the student "red, but a brighter red."


I think once you do that, you are teaching relative pitch of the octave. Whether you call it a brighter red, a hotter lemon, or a bigger widget, you are basically noting the relationship between two pitches an octave apart. IMHO that's not a bad thing. But be clear that's teaching relative pitch, not absolute pitch.

I think that in terms of absolute pitch, the closest pitch to any given pitch is a microtone up or down from the pitch. Harmonically, the closest pitch is an octave, followed by either a 4th or 5th which are equidistant, followed by 3rds/6ths, followed by 2nds/7ths which are the furthest from the given pitch. And in terms of the scale and melody, the closest pitch is the next pitch on the scale. These are all different ways of organizing and listening to pitches.

Regards,
Lou

LouMan5
Posts: 12
Joined: Thu Sep 08, 2005 6:31 pm

Post by LouMan5 » Mon Nov 21, 2005 11:05 pm

Ok I read the research you referenced at Separating pitch chroma and pitch height in the human brain. The following quote is my own synopsis/paraphrase of this article into human language.

Amazingly, even though people generally think of pitch height as strictly a function of frequency, there are at least three ways one can alter spectral characteristics which lead to a subjective sense of height difference even though the fundamental pitch remains the same (ie it's still a "C"). In particular, (1) one can change the intensity of alternate harmonics (ie reduce the intensity of odd harmonics), or (2) one can change the phase of alternate harmonics, or (3) one can change the spectral envelop of the sounds (Shepard Tones). This article focuses solely on method (1) and in particular, by reducing the intensity of odd harmonics.

Height difference or equivalence is determined by presenting two tones to a person and then asking them which one is higher. Most of the time, if we present a C and a C#, people will say that C# is higher. Now here's a trick. Let's say we present two tones C and C#. Then we apply a "pitch height" increase on the C tone by reducing the intensity of the odd harmonics. At some point, we will get people guessing 50% of the time that the C is higher, and 50% of the time that the C# is higher. At this point, we can say the two tones are at the "same pitch height" even though they are at different chroma. We have thus changed chroma independently of height! Now we think it's not enough to fry their brains with these weirdo pitch changes, so let's fry their brains with some extra MRI radiation as well to see what's going on inside there!!!!! Then when we get the results, we'll go to a bar and see if we can impress some chicks with how smart we are because the results are incomprehensible to anyone unless they happen to have triple majored in Physics, Brain Surgery, and Statistics.

Anyway, the results are that when the chroma changes but the height is kept constant, it affects areas of the brain in front of the Primary Auditory Cortex (roughly, on the side of the head but slightly in front of the ears) but when chroma is kept constant and pitch height is changed, it effects areas behind the Primary Auditory Cortex (roughly, on the side of the head near the ears. Oh yes, also did we tell you that we are the smartest group of people on the planet Earth?


I'm not quite sure what to make of this, but I do have some reservations on using this to make assumptions about perfect pitch. First of all, they only tested one of the three known ways to alter pitch height independently of chroma. Secondly, their subjects were most likely not absolute pitch listeners, who often cannot tell if one pitch is higher than another because that's a relative pitch function. And finally, in reference to the following quote:

Statistical parametric maps were generated by modeling the evoked hemodynamic response for the different stimuli as boxcars convolved with a synthetic hemodynamic response function in the context of the general linear model.


I mean, common.... boxcars!?!?! What do railroad cars have to do with anything? And hemodynamic !?!?!? Let's leave the weird sex stuff out of this, ok???? These guys are whack. Case closed.

Regards,
Lou

aruffo
Site Admin
Posts: 1692
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Post by aruffo » Mon Nov 21, 2005 11:55 pm

Um.. I confess I'm not sure exactly where you're going with that, or how much of what you're saying is meant to be a joke, but at this point I guess I shrug and say that the research is there, and the results are what they are, and I'm intrigued by your objections to their conclusions so I'll probably take another look at it. F'r my part, as I see it, I suspect that etaxier has extracted the most potent statement
It's still a very interesting point, that timbre plays such a huge role in the way we process pitch 'height' (or luminosity, or whatever).

and very handily summarizes what I've come to understand, which is that "height" is a timbral issue. "Distance between tones" does not exist for absolute listeners because there's no such thing as "distance". Just timbre. I find it most useful to think of "relative pitch" as structural thinking-- which is why I generally try to avoid using the term "relative pitch" these days.

LouMan5
Posts: 12
Joined: Thu Sep 08, 2005 6:31 pm

Post by LouMan5 » Tue Nov 22, 2005 1:01 am

Well I just gotta add some humor to these dry studies or I'll go bonkers. Did I get the idea right - if you drop out the odd harmonics it raises the pitch? I wasn't sure on that detail.

Ok back to the main thread.

what I've come to understand, which is that "height" is a timbral issue.


Please help me understand this. They way I see it, pitch can be a timbral issue when the tones are manipulated artifically, thus exposing some unusual characteristics in our auditory system. But is this really the dominating factor for height perception in a natural setting? I think that frequency shift is the dominate cause of height perception in natural musical settings, not timbral changes. This is a pretty fundamental understanding, even among those very very smart people who wrote that study. :wink:

One thing that is interesting is if I understand correctly, timbre changes can only increase the height by a maximum of one octave. Also, it seems to me that this timbre changing is not the same kind of height change that fundamental frequency provides. Timbre changes are just a height change relative to the original timbre.

This is why I contend that there is no such thing as perfect pitch outside your experience of timbre or outside your experience of pitch in octaves. For example, a female voice sounds higher than a male voice because of timbre. But if you take the time to learn both timbres absolutely, you will learn to be able to hear the fundamental and associate them as being equal pitch. If you only learned pitch in an octave on one instrument, you will not have perfect pitch for other octaves - you might be able to fake it using relative pitch to figure out those other octaves. Also, you won't have perfect pitch for other timbres that are significantly different from your learned repetoir of AP. AP doesn't automatically transfer to other contexts. It takes a lot of learning. That's what makes it so difficult to achieve.

"Distance between tones" does not exist for absolute listeners because there's no such thing as "distance". Just timbre.


I don't understand that one. Are you saying that there is no such thing as distance for APP or are you saying there is no such thing as distance flat out? People who learn AP early sometimes have trouble with relative pitch because they calculate it in their head ("C to F that's a 4th") rather than feel it. However, they instinctually know when a note is "a little sharp". That is, they don't calculate that... they perceive it directly. To me, that indicates there's directional distance that is separate from timbre.

Regards,
Lou

LouMan5
Posts: 12
Joined: Thu Sep 08, 2005 6:31 pm

Post by LouMan5 » Tue Nov 22, 2005 1:32 am

Bass Geek wrote:I also had a theory student once who was a very talented organist with "naturally" acquired AP. (She said that most of her life she just assumed everyone could hear distinct pitches). Anyway, not only could she not determine octaves, but she couldn't even tell which note was higher or lower than another! This presented great difficulties in RP testing - she could identify what chord type she heard, but not the inversion ...


This is exactly the prediction of my theory. She learned perfect pitch but then associated all C pitches to be the same class. She probably had to practice that quite a bit, thinking "C" whenever she heard a "C" regardless of what octave it is in. Just goes to show how the mind can get into the trap of being too good at something. Well now when she hears "C3" followed by "F3" her mind is shouting C!!!! F!!!! and she has no idea if it went up or down in pitch.

Regards,
Lou

aruffo
Site Admin
Posts: 1692
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Post by aruffo » Tue Nov 22, 2005 5:25 am

I was glad to be reminded of that research article... to be reminded that I've already written about this at length, in detail, and with specific scientific references in "Phase 11" and various other places around the main site. The prospect of repeating it all is wearying; so let me offer a desultory "please re-read that stuff" and then volunteer the briefest of recaps.

"Height" is a perceptual concept that is learned to compensate for the average listener's lack of harmonic perception. Not only do absolute listeners not perceive "height", but neither do normal listeners until they are trained to it.

Two pitches create a harmonic interaction. This pattern, when heard, can be interpreted simultaneously as the movement of an implicit object ("distance") and as a tonal event ("harmony"). I would argue that harmony is the more musical interpretation.

Two voices, or perhaps two differently-fashioned tuning forks, seem to be "higher" and "lower" because they are. They have the same one-dimensional chroma, but they are not the same two-dimensional pitch. Because they are nearer "apart" than an octave, and we don't have a musical model to explain different "heights" of the same chroma that are not octaves, we insist that they really are the same pitch. But they aren't.

Guest

Post by Guest » Tue Nov 22, 2005 8:31 am

Woops. I'm very sorry if I've offended or wasted your time. I've read your entire site and some places more than once. I'm trying to keep an open mind. Thanks again for the recap and I'll go check it out again.

Regards,
Lou

aruffo
Site Admin
Posts: 1692
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Post by aruffo » Tue Nov 22, 2005 11:22 am

No fear about either, I'm sure... I'd forgotten that the reason I "know" this stuff is that I've written about it (and, most recently, also cited the publications). Although I suspect the research demands a re-definition of "absolute" and "relative" pitch, I prefer to be challenged where the evidence seems lacking. I'm "weary" only because to fully (re)explain those statements I just said flatly-- which are not true but merely supportable conclusions-- I would have to re-write all the support that I have been gathering in these recent "phases"... which shouldn't be necessary, because I've already written it, but at the same time I can't really hold up my end of a discussion on this topic without it... which means I have to ask you to read it.. in which case I appreciate your patience.

Post Reply