A New Approach

Thoughts and responses regarding the research at acousticlearning.com.
TS
Posts: 168
Joined: Sun May 07, 2006 4:58 am

Re: Naming Pitches

Postby TS » Sun Nov 04, 2012 1:46 am

zacxpacx wrote:One of the problems a relative listener faces in the development of perfect pitch is their perspective of the staircase. A purely relative listener views the staircase from the side. All they can see is the ascension of the stairs, not the repetition of colors. To them, the staircase is a vertical line. It is a vertical cross section from top to bottom of the staircase. To begin developing absolute pitch, the relative listener must change their perspective. Luckily, APA already accomplishes this.


This is not actually the case. Most (all?) relative listeners perceive colours on the spiral, and also perceive that the colours repeat. It may take a little training for them to become aware of this fact, but the mechanism is there. Most (all?) relative listeners even have categorical perception of the tones. The difference between relative listeners and absolute listeners is that for the relative listener the colours are relative to each other, so if they hear a certain chord sequence they may find that suddenly the colours have all moved to different places (a modulation, a.k.a. a key change has happened).

The Interval Loader component of ETC trains this kind of categorical relative listening.

Here's the entry from 2003 when Chris first became aware of this phenomenon: http://www.aruffo.com/eartraining/resea ... #intervals

zacxpacx wrote:As for relative pitch -- it should not be a problem. Those with relative pitch don't automatically know the name notes in a piece after given the starting pitch. Using relative pitch to name notes requires focus. ... As long as the player doesn't actively try to use relative pitch to label tones, it shouldn't be too much of an issue.


The problem is that most (all, as far as I know) people who don't have absolute pitch already have a strong mechanism for perceiving tones categorically, and if you give them a task where they have to place tones in categories, they will automatically, unconsciously and without effort use this existing relative mechanism to accomplish the task. The task will therefore improve their relative listening, but it probably won't develop their absolute listening at all.

zacxpacx wrote:I see memorization as a prerequisite for categorical perception.


Memorisation is probably important when training for absolute pitch, but it may be important in the same way breathing is: obviously if you don't breath then you're not going to learn absolute pitch, but mostly you can just take it for granted.

It may be that to develop categorical perception you need to learn some concepts / linguistic tokens / labels / names and the categories will follow from that process, and if you're able to remember what you had for breakfast then you will memorise the categories with no effort in the process.

On the other hand, if you're not training the right concepts or labels etc., then you won't learn to perceive pitches categorically with any amount of repetition.

zacxpacx
Posts: 157
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

Re: Naming Pitches

Postby zacxpacx » Sun Nov 04, 2012 12:43 pm

TS wrote:This is not actually the case. Most (all?) relative listeners perceive colours on the spiral, and also perceive that the colours repeat. It may take a little training for them to become aware of this fact, but the mechanism is there. Most (all?) relative listeners even have categorical perception of the tones. The difference between relative listeners and absolute listeners is that for the relative listener the colours are relative to each other, so if they hear a certain chord sequence they may find that suddenly the colours have all moved to different places (a modulation, a.k.a. a key change has happened).

The Interval Loader component of ETC trains this kind of categorical relative listening.

Here's the entry from 2003 when Chris first became aware of this phenomenon: http://www.aruffo.com/eartraining/resea ... #intervals


I agree with the point you make here. Let me clarify what I had said before. Relative listeners naturally do not perceive the chroma colors of the tones. But, as you said, with a bit of training, they can learn to hear the different colored chroma and even notice that the chroma repeat themselves. Like I said, this is what APA does. It shifts the relative listeners perspective above the staircase, where they can view the cyclical spectrum of chroma. I also agree that relative listeners can develop categorical perception for tones. What I was trying to get at is that relative listeners do not have categorical perception for pitch.

To have categorical perception:
1. Differences between items in the same category are compressed.
2. Differences between items in separate categories are accentuated.

TS wrote:Memorisation is probably important when training for absolute pitch, but it may be important in the same way breathing is: obviously if you don't breath then you're not going to learn absolute pitch, but mostly you can just take it for granted.

It may be that to develop categorical perception you need to learn some concepts / linguistic tokens / labels / names and the categories will follow from that process, and if you're able to remember what you had for breakfast then you will memorise the categories with no effort in the process.

On the other hand, if you're not training the right concepts or labels etc., then you won't learn to perceive pitches categorically with any amount of repetition.


My reason for bringing up the issue of memory is this: APA teaches you to perceive the chroma of pitch. It goes through all 12 pitch categories having you compare and learn what each pitch sounds like. Why is it that after a lot of repetition a listener can't ingrain the chroma of C, D, E, F, etc. into their mind? Or can they? I haven't played APA long enough to know myself.

zacxpacx
Posts: 157
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

Bad Me

Postby zacxpacx » Sun Nov 04, 2012 7:52 pm

I guess I should've read further into Chris's research before giving the staircase analogy. Though it was something I thought of on my own, it turns out that the helical modal of pitch has already been disproved.

http://www.aruffo.com/eartraining/research/phase13.htm

The explanation can be found here in the entry "Slide some chroma to me".

Axeman
Posts: 151
Joined: Wed Jul 20, 2005 4:20 pm
Location: New Zealand

Re: Naming Pitches

Postby Axeman » Mon Nov 05, 2012 6:50 am

zacxpacx wrote:My reason for bringing up the issue of memory is this: APA teaches you to perceive the chroma of pitch. It goes through all 12 pitch categories having you compare and learn what each pitch sounds like. Why is it that after a lot of repetition a listener can't ingrain the chroma of C, D, E, F, etc. into their mind? Or can they? I haven't played APA long enough to know myself.

I have played a lot of APA in the past and have experienced just what you are saying Zac. At first the C chroma stuck out like the proverbials after a short while of playing but after entering into the other pitch levels the fixedness of the c chroma began to fade and by the time I got up to the B note had still not solidified for me. Although I did hear the G pop out informally while listening to music.

Axeman
Posts: 151
Joined: Wed Jul 20, 2005 4:20 pm
Location: New Zealand

Re: Naming Pitches

Postby Axeman » Mon Nov 05, 2012 7:28 am

TS wrote:
It may be that to develop categorical perception you need to learn some concepts / linguistic tokens / labels / names and the categories will follow from that process, and if you're able to remember what you had for breakfast then you will memorise the categories with no effort in the process.

On the other hand, if you're not training the right concepts or labels etc., then you won't learn to perceive pitches categorically with any amount of repetition.


I definitely agree TS

The Trouble I found with APA is that it zeroed in on the chroma of each note but there was no way of solidifying that chroma into the memory. Actually the idea of chroma (the experience of the reality of it) is still kind of hazy for me for this reason. I recall someone posting way back something like ' Isn't this just like repeating the note over and over? ' And even though there was good reason behind Chris design of the tests for each note it still felt like this comment was close to the real experience in the long run because the chroma was never concrete in the perception. i.e. retainable in the memory.

I was thinking that maybe because this phenomenon exists then it may be that we need to help our ears to categorize the pitches by at first assigning a different timbre to each note while still maintaining the same testing methods of APA. As higher levels are reached new timbres are added to each pitch category. so the beginning levels will be super easy - C is trumpet, G is violin (across octaves too). But by adding new timbres for each note as the game progresses the game gets harder. By the third level you are naming the notes using three different timbres for each pitch class. Eventually you will be naming the notes with equally shared timbres for each pitch class.
Another way may be to start with different sounding timbres for each pitch class and slowly morph them into the same timbre at the higher levels.

All this is to say that we might be able to impose on the ear an obvious categorization to guide the deeper categorization.

I think that the categorical perception of pitch has to be latent or otherwise we are never going to develop the ability without being genetically 'gifted'.
It might be that children who get early training in music achieve categorical pitch labeling because their brains are still developing left and right brain tendencies for specific jobs. So the labels themselves are perceived in a kind of right brain way and vice versa and back and forth for the perception of pitch categories. Whereas adults may have well defined tendencies assigned to right or left brain activities and they don't cross over much anymore.

TS
Posts: 168
Joined: Sun May 07, 2006 4:58 am

Postby TS » Mon Nov 05, 2012 2:03 pm

About the memory thing, I guess it's been mentioned many times here in the forum and in Chris' notes, but perceiving meaning in the thing you're trying to remember will help in remembering it.

Here's a demonstration that I usually have in mind when thinking about this issue, I'm not sure where this comes from or if I just invented it myself.

Here is a sequence of letters, look at them for 5 seconds and try to memorise the sequence as best as you can:
achryeikdtba

Now here are the exact same letters, but ordered differently, again look at them for 5 seconds and try to memorise:
birthdaycake


So the point here is that it's easy to remember something if it has some meaning, but almost impossible to remember the same thing if it's presented in such a way that it doesn't convey any meaning (like a sequence of letters rearranged in a different order).

So, to come back to APA, the sounds in the game are single tones in the beginning, then chords with some tome added, then arpeggiated chords, etc. To me it seems that the sounds in the game are all musical equivalents of "achryeikdtba", and there aren't really any "birth-day-cake" sounds in the game. If you have the imagination you can imagine how a chord or arpeggio is part of some melody, and maybe then you can get actual results with APA, but otherwise it seems that the sounds in APA are not conveying any emotional message, which supposedly is what music is for.

I've been thinking that maybe the only thing that's needed to make APA actually work is to switch the melody words and the egg sounds around, so that the egg sounds are recognisable and meaningful melodies, and the melody words are single tones.

zacxpacx
Posts: 157
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

On Memory

Postby zacxpacx » Tue Nov 06, 2012 8:05 pm

That is a good point you bring up on memory, TS: percepts and sensory information that conveys an idea or concept is easily identified and remembered.

I don't know if I agree with where you go from there, though. You say that you want a sound in game that isn't random, something that sounds familiar and has meaning like birthdaycake. Certainly, there are plenty of melodies and sounds that have meaning in all of our lives, but we are trying to find a concept of pitch. Even a well-known melody won't do the trick. You'll just be learning that the first note of the melody can be called "c" or "d".

We have to come up with a meaningful concept for "c" and "d" on their own.

aruffo
Site Admin
Posts: 1688
Joined: Tue Dec 14, 2004 12:09 pm
Location: Evanston, IL

Postby aruffo » Tue Nov 06, 2012 11:49 pm

Hm.. I'm considering the timbre-pitch association.. not because I think that timbre necessarily would be effective for the purpose (you want to be able to recognize the pitch directly, and not infer its identity), but you make me wonder about dimensionality. Robert Goldstone's lab has demonstrated that we can naturally and unconsciously form stable categories when the categories are multidimensional, and this made me scratch my head because I didn't know how to induce categorization of a unidimensional continuum.

But what if there were some second dimension attendant upon pitch? That would increase the amount of information available to make a judgment. In the case of people whose absolute pitch perception locks them into C-major, the second dimension is its membership in that key, but obviously we relative listeners aren't locked into anything.

Another idea-- and this is a new thing, both generally in the scientific literature, and to this site-- has upended the idea of how we learn phonemic categories. It was previously thought that we learned them by direct comparison: f'rexample, that we learned to abstract a "b" sound by realizing that bat and pat are different. But this is the Chordhopper mechanism, and while Chordhopper does improve "phonetic" awareness by forcing our brains to automatically start picking out pitches, it doesn't induce categorical perception. This other process, apparently, does.

The process.. oh, drat, I can't remember what it's actually called... suggests that we learn different phonemes because we constantly hear different examples of each clustered around its prototypical "center". Before long, our minds start to average out the samples to find the "center", and subsequently create a boundary midway between centers, and consequently associate items on either side of the boundary as "belonging" to that stimulus area.

The catch, of course, is that the center itself is a fixed point (albeit a synthesis of multiple components), and something has to reinforce its fixedness.

Which gives me an idea. Hm... I need to figure out how to program random waveforms on demand.. MIDI does do microtones, but those only seem to work when played on a Macintosh. Maybe I'll test this on a Mac and see how it goes. (Where'd I put that old laptop? I don't think I've touched it for five years or so...)

TS
Posts: 168
Joined: Sun May 07, 2006 4:58 am

Re: On Memory

Postby TS » Wed Nov 07, 2012 11:21 am

zacxpacx wrote:You say that you want a sound in game that isn't random, something that sounds familiar and has meaning like birthdaycake. Certainly, there are plenty of melodies and sounds that have meaning in all of our lives, but we are trying to find a concept of pitch. Even a well-known melody won't do the trick. You'll just be learning that the first note of the melody can be called "c" or "d".

We have to come up with a meaningful concept for "c" and "d" on their own.


I'm thinking that what we really want is a concept of pitch within the language of music. A running car engine or a refrigerator has a pitch, and a cat walking on a piano keyboard produces pitches, but I think that learning those won't help in learning absolute pitch.

The difference is do we want to learn 'a' as in 'achryeikdtba' or 'b' as in 'birthdaycake'.

When you see the letters arranged as birthdaycake, you don't really see the letters, you try to ignore the letters as quickly as you can and instead you see a word and maybe an image and some association, you see meaning in short. When you see achryeikdtba, on the other hand, there is no meaning so you must focus on the letters and try to make something out of it.

Now, when you're playing APA the sounds are mostly meaningless, so you have to focus on the sounds themselves and pick out all kinds of qualities from them. Then later when you're listening to music the sounds you hear hardly even register because you're paying attention to the emotional meaning and not the sounds, so it's like a completely different brain circuit gets activated. If this is the case then it's no wonder that nothing transfers from the APA practice to the "real world".

So maybe instead of having a game where you hear sounds and get asked "is there something in common between these sounds?", we need a game where you receive emotional meaning through music and get asked "is there something in common between these emotional messages?". And maybe the concepts for the 'c's and the 'd's will come naturally when the focus is on meaningful messages instead of meaningless sounds?

zacxpacx
Posts: 157
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

The Meaning of Pitches

Postby zacxpacx » Wed Nov 07, 2012 4:16 pm

Hey TS, I agree and disagree.

I do think that it is important to learn the concepts of pitches in musical context, I don't think the parallel you've drawn between musical and language meaning is accurate.

Language spelling is repetitive and constant, and people learn a fixed number of words. Birthday cake, giraffe, forest -- all are "words" composed of "letters" attached to "concepts". People learn the concepts for which these words are attached and learn to interpret them in chunks, instead of letter by letter.

In music, there is a theoretically infinite number of "letter" combinations attached to an infinite number of emotional "concepts". Furthermore, once someone knows how to listen to music, they can extract meaning from any musical phrase. There isn't a learning of words as there is in English. Words aren't assimilated piece by piece in music. Everything in music is inherent and natural.

What people need to learn are the pitches that construct musical words and the concepts they embody, so that we can deconstruct musical words and phrases. The teaching isn't of musical words. We know how to interpret those. We need to teach pitch categorization. Disassembling and integrating this perception with music is a step to be taken after.

TS
Posts: 168
Joined: Sun May 07, 2006 4:58 am

Postby TS » Thu Nov 08, 2012 11:55 am

zacxpacx wrote:In music, there is a theoretically infinite number of "letter" combinations attached to an infinite number of emotional "concepts". Furthermore, once someone knows how to listen to music, they can extract meaning from any musical phrase. There isn't a learning of words as there is in English. Words aren't assimilated piece by piece in music. Everything in music is inherent and natural.


This seems to me a romanticised and naive view of music. For start, I don't think there are an infinite number of emotional concepts, for example Paul Ekman's list of basic emotions only contains 7. And only relatively few musical "letter combinations" are attached to any emotional concept, exactly like there are theoretically millions of ways to rearrange the letters "birthdaycake" but only a handful of those arrangements have any meaning.

And you can really only extract meaning from musical phrases that you are familiar with. You do have to learn words in music like you do in English.

And when we take the perspective of reading and writing music it becomes even more like reading and writing language. You absolutely must learn "words" in music, and you need to know which "letters" they are made of before you can read or write music with any proficiency.

zacxpacx
Posts: 157
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

Infinite Meaning

Postby zacxpacx » Thu Nov 08, 2012 4:11 pm

I respectfully disagree. I do think there are a theoretically infinite number of musical concepts. I don't want to flesh out an argument as to why right now, maybe later. I do agree with the role you suggest absolute pitch must play in music as a language, and I think this is what's most important, as the goal of this site and forum is to teach adults perfect pitch.

Whatever the function and meaning conveyed by musical words, pitches are the "letters", the basic unit upon which everything else is built. To learn perfect pitch is to perceive the spectrum of sound frequency in chroma and to divide it into 12 musical pitch categories.

Let's get going on THAT!

zacxpacx
Posts: 157
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

Game-Making

Postby zacxpacx » Wed Nov 21, 2012 3:55 pm

Other than Chris, do any forum members have programming skills for making ear training games? Beyond basic programming this would include generation of specific pitches, timbres, chords, etc.

zacxpacx
Posts: 157
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

Postby zacxpacx » Mon Nov 26, 2012 2:10 am

The Importance of MRI

Since this thread was my introduction to perfect pitch research that included some generalizations on my direction of attack, I will continue to post updates to how I plan on tackling perfect pitch here. For this particular post, I want to clarify the criteria that for me satisfy true perfect pitch.

Previously, I had two criteria: either being able to name all the notes of a melody or activating the same areas of the brain as an AP listener while listening to music. For now, I'm cutting it down to JUST using the same areas of the brain as an absolute listener. Why? There are too many ways to learn to name notes (some which allow great speed) and despite all of Chris' research there is still ambiguity as to what perfect pitch really is.

Two theories as of now (and I could probably find more with enough digging) are categorical perception of chroma dependent on fundamental frequency and absolute judgement based on the internal relationships of overtone series. There are arguments for both. Nothing is definitive.

Which got me thinking: what would be a way to correlate one or the other to true absolute pitch? If an MRI could show that using one of the strategies -- this is glossing over the issue of implementation, since that is a whole problem of its own -- activates the same areas of the brain as an absolute listener, then that would be strong supporting evidence that it was the correct strategy. Activation of the same areas of the brain seems the most definitive test to me right now.

zacxpacx
Posts: 157
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

Postby zacxpacx » Tue Nov 27, 2012 7:23 pm

The Importance of MRI pt. 2

Today I thought about just how hard, if not impossible, learning perfect pitch as an adult may be. As adults, our brains are automatically activated in specific areas when we hear sound or music. Any sound, any context, the areas of the brain that are lit up (relative-pitch areas/language areas etc.) remain relatively constant. If the sounds we hear already activate fixed areas of the brain, how are we ever to learn perfect pitch?

Taking a step back from perceptual learning and cognitive psychology: To gain absolute pitch we must activate the areas of the brain associated with it. As of now, we don't know how on-track or off-track we are to really pinning down perfect pitch. Getting MRIs of individuals playing APA seems to me the best way to confirm whether or not anything meaningful is happening in the brain. If an individual playing APA uses the same areas that they normally use when listening to music/sound, then perfect pitch could very well be impossible to learn as an adult.

Perfect pitch could be dependent on a child to use different areas of their brain to make sense of the sound they hear. If adults cannot be prompted to use those areas of the brain associated with perfect pitch when listening to sound, then it's hopeless.


Return to “Your thoughts about music”

Who is online

Users browsing this forum: No registered users and 1 guest