Thoughts and responses regarding the research at
Post Reply
Posts: 24
Joined: Sun Jul 10, 2005 12:04 pm


Post by Lyle » Sat Dec 01, 2012 6:58 pm

(Audio exercises to download are linked below, to the philosophically-impatient among you.)

Imagine that your infant ears & your newborn brain have heard the sounds of the world for no more than the few months that a baby has.

Your new ears, your new brain have yet to realize how unimportant the vast majority that most frequencies are: all are present, all the time, all are demanding attention… but the fundamental frequency of each pitch is fighting to grab your complete awareness to the exclusion of the extraneous "noise"… eventually, it will be the fundamental frequency that wins the struggle because that is what brings life to this wonderful experience we all call music. Oh, and the odds are also unfairly stacked on the fundamental's side: Mommy and Daddy might hum some melodies and recite some stories to us each day, captivating our little hearts and riveting our aural concentration toward what matters most to our ultimate human daily needs. Ignoring the higher frequency spectrum is quickly found to have no apparent consequence to the developing child -- if there is any meaningful cognitive content to be found at all "up there" it surely is as important as knowing the difference between a prime and a non-prime: so completely, utterly useless to a baby, and almost always of no utility to an adult, either.

Until that unconscious & semi-cognitive realization of Fundamental vs. Full Spectrum takes hold invisibly & silently inside every child's brain, some lucky few of us, although perhaps decidedly unlucky as many of them have expressed because their neural circuitry are working overtime against their will to ceaseless interrupt their concentration with mostly useless aural observation… their Perfect Pitcher brains electro-chemically cascading in overdrive, organizing every structural aspect of every coincidental part of a pitch's component with no regard to the brain owner's feelings about the matter. Well, the rest of us clueless cute kids just cuddled close and enjoyed listening to mommy sing about rowing boats and Winnie the Pooh.

All right. Enough facetiousness. I've set the imaginary scenario for now, truth or falseness not really important, the background is set for this post.

There is enormous "structure" in the simplest of a single sound, out there, always out there, all the time -- every kid *could* have picked it up, but most kids just don't. There's no point, no reason, there's no value, there's no payoff, there's no thrill or excitement or avoidance of danger. And as adults, we need MOTIVATION to bother, especially when every expert, every test, every instinct, everything we know in our gut already says that it just doesn't matter.

So then, the audio files I'm linking to here are an attempt to train your ear/brain to learn to hear all that structure you most likely do not hear and do not realize that you do not hear. In fact, you will NOT hear it, I bet. But I think with a little bit of time, you CAN hear it, and then with some more time, you CAN'T HELP but hear it, and then eventually (YEARS, mind you, YEARS) it will be a natural part of how you hear your own instrument. Not all instruments, not all music, but as an adult musician, I am pretty sure it will be part of how you handle your own instrument. Is it worth it? I think so. No guarantees. But hey! I'll even offer you a MONEY-BACK GUARANTEE! Yes, that's right! I'll give you back every penny you are going to pay me for this secret information!!! Oh right, you aren't giving me anything. Ha... sometimes, that means the information is worth more than when you pay for it... sometimes, it means it's worth less. You're on your own for figuring it out, this time. :-)

Download one or more of the following, depending on your interests, then continue reading:

All 160kbps MP3. (202MB) (114MB) (147MB)

A Perl script to produce these is linked to farther below.

Here's the overview of the plan: you are going to listen to one pitch, one overtone, in decreasing loudness, until it JUST DISAPPEARS FROM YOUR AWARENESS. That is the exact point you need to practice! Take the preceding track, that has the emphasis you need to hear it, and repeat it with the track that you can't hear it -- until you DO hear it. then, move on to the next lower pair, until you get to the lowest, non-elevated overtone. Your mission, should you choose to accept it, is to practice until you can hear the specific overtone in the original zero-EQ-modified audio file.

See how this is different from normal ear training? I as your instructor can't say, "Was that a perfect fourth or fifth?" And you tell me, "Fourth!" and I say, "Wrong!" and make a mark in my book and adjust my lesson plan accordingly. Only you will know whether you can hear it or not. This we might call "intrinsic" ear training, versus "extrinsic." Or maybe not. Most (all?) ear-training is objectively verifiable from an instructor watching you from the outside. There's a different kind of ear-training exercise, much less explored & researched & talked-about, of which this post is about, that no one can look at you and decide whether you are doing it right or not. I'll come back to it in a bit if I remember to.

Back to the exercise -- just do it for a few minutes a day, don't overtax your ears/brain/sanity. Choose ONE NOTE and ONE OVERTONE to start, let's say, pick middle C -- note # 48 -- or middle E -- note # 52 -- and one overtone -- I suggest overtone # 5 because it's such a friendly fellow. Set up a playlist, or just sort the sounds so they go in reverse order. You will want to listen to the loudest one first. So using the piano as an example:


and let's say that BOOM you could faintly hear it at -12 but you can't hear it at -09. Set up a playlist and loop the two. Or loop the whole set, -18 -15 -12 -09 until you DO hear it at -09. The next day you might not hear -09. That is so completely normal you should smile and be happy that you are human and not a freak of nature.

Concentrate on a just a couple of notes. As I've said before, I picked C and E, specifically in one octave. I learned those two so well that it was incredible how differently they sounded to me. They sounded different in Character, in Timbre, in Quality, in Kind, in so many ways that had nothing to do with listening for overtones. And that's why I keep trying to get the message across -- your ear has naturally shut itself down for listening to those useless frequencies. Gradually open it back up. Then stop listening for them again but now, as an adult and a trained musician, you're going to be using that new information.

[So far and I'm not suggesting otherwise, but for the first few weeks, months, year(s)… don't bother trying to practice distinguishing notes from one another, just focus on the C-ness of C and the E-ness of E…]

Hopefully within a week or two, you will find yourself having easily dropped from -18 to -00 on several or all of the overtones of one of the pitches.

[NOTE: For all of the pitches, I have included a NON-MODIFIED file that ends in -00. That's for convenience so you can countdown from -18, -15, -12 etc to -00 no matter the harmonic you are working on. All of the -00 files for the same note are exactly the same, they do not have any equalization applied.]

Repeat with all notes, all overtones, until you have finished this timbre. Repeat a few times over a few weeks, because you're human, not a computer, and your ear/brain needs some time to absorb it all. But work on a new timbre -- piano, guitar, bass guitar, violin, flute, etc.

Also, OF COURSE, try to hear some/all of these overtones on your own instrument! Pick one pitch, one overtone. It isn't going to be easy, trust me. Re-read that Frequency Response Thread for my own difficulties over weeks/months.

I have plenty of other samples I can process & upload but what's uploaded he ought to be enough for discussion & experimenting. And I think you guys should (1) sample your own instruments -- the one you play & improvise & compose with every day -- because that is your fastest way to really get a great ear with your own instruments, (2) pick ONE timbre and really stick with it for a while, don't skip around, and (3) share with all of us, please.

I think you need to do this on one set of speakers / headphones. Don't keep changing environments. It's hard enough to hear these overtones as things are. They'll appear and disappear and you'll hear them one second and then not the next. (Which is why it was REALLY helpful for me to be hitting keys on my piano to draw my attention to WHERE I am supposed to be listening.) I do not know if unconscious/unaware listening of these tracks -- like just letting it run in the background -- would have any use at all, to let the overtones sink in. I do think that 99% of the value is in learning to consciously focus on hearing the specific overtone out of the mix of overtones, in the unaltered pitch. The EQ'd tracks are just like training-wheels to help you get to that point.

Why to do this, what might happen, why I think it's worth doing -- see that Frequency Response Thread I keep referring back to. Try the exercises I mentioned there, too, if you have any success hearing overtones with this thread's exercise, because I had so much rapid improvement by playing keys on my keyboard (equivalently, guitar) that were close to a lower key's overtone. I am sadly completely and thoroughly overwhelmed by some events in life that are may keep me from being able to join in very much conversation, but I will try when I have any opportunity. Do read that F.R. thread because that has many thoughts of mine on what you want to accomplish with all of this & why. Hence I am providing the scripts/source so if anyone decides there is any merit, someone else can pick up and run with it, even if you guys never hear from me again.

Lyle's Standard Disclaimer: Reminder, reminder, reminder -- while we all have gathered here today, day after day, to discuss and pursue our common interest and frustrations with perfect pitch, I like many others who have been at this for TOO LONG have realized that SCALE DEGREE RECOGNITION -- solfege -- is one thousand times more important than perfect pitch. Have fun with this and dedicate some time to it but DO NOT NEGLECT your DO RE MI's (all 12 chromatics in all 12 keys) or you will end up a 41-year-old frustrated composer like me. I repeat myself, for good reason! My ears are still .005% of what I wish they were and Darn It, I wish someone had known enough to teach me to sing/identify solfege when I was 12, or 8, or 6.

Undergoing targeted training like this overtone listening is not going to address real-time pitch recognition, within real music, within real harmony & chords & changing timbres. We've had many discussion over that… I both agree with Chris and disagree with Chris that "categorical perception" is the key to perfect pitch. So let me be confusing: I do not agree that "categorical perception" is the key to perfect pitch. Mathematically, that would certainly accomplish it! But, we're talking about wetware, and mushy old neurons, and this spaghetti of dendrites is a functional mess. We've had a quick discussion before, and maybe we just "agreed to disagree" without actually saying it -- in fact, neither Chris nor I nor anyone else said anything to each other on this if I remember. But I do believe that the premise said aloud, that there is exactly one dimension that underlies chroma: I believe this is misleading, it isn't true, and will send you on a wild-goose chase for the wrong reason. Chris did point out, if you have 2+ dimensions, the [brain] will auto-segment categories. So what if chroma had nothing to do with the fundamental, the brain auto-segmented on all this higher-level frequency stuff, the higher-frequency stuff was actually NOT used to identify pitch, but in the end, you're still left with a segmented single-dimension of the fundamental frequency line? I'm saying it could happen, but I'm not claiming any evidence for it, as it were. We already had that conversation, I believe. :-)

Having struggled for 20 years now with pitch recognition, having some success in identify pitches absolutely (on my piano) and similarly, with scale degrees, the issue I see is that it is universally isolated to outside of realtime music. How do we train for realtime, live music? What is the bridge for turning a skill, which some of have agreed *is* obtainable (identifying pitches in isolated but context-specific situations), into a skill that functions in realtime when there's a whole lot of other stuff going on? I don't think that categorical perception is the answer there, even though *by definition* categorical perception would accomplish that. (Yeah some of you will pounce on me for that strange admission, maybe I can counter-response some other time…)

I think the more difficult problem is that recognizing a musical "thing" out-of-context -- whether it be a specific pitch, a specific timbre (piano or guitar or flute), or a specific scale degree (DO RE MI) -- and then recognizing it at all times in real-time in the musical flow of a real song, all real songs… that's a difficult and different problem, and the one the requires (1) serious & significant musical practice and (2) unknown & unsolved theoretical problems in cognitive science.

(A semi-random example jumps into my head, of learning a martial arts skill -- you practice blocking and striking out of context, out of real time, but somehow, eventually, one hopes, you will know how to apply it in a scenario that you have never practiced, never planned, never saw coming, making identifications & decisions & applications in strict real-time. Same thing, sort of, right?)

I believe that if this hypothesis of full-spectrum frequency structure, so ineloquently stated above, is correct about structural overtone hearing as part of the basis of perfect pitch, then writing a computer program is a better approach than leaving everyone to fuss unguided with a set of thousands of audio files. Because it can keep track of your exact cutoff point of where you are hearing/not hearing each overtone for each pitch, for each timbre, etc. And even make it less discrete, allowing you to dynamically raise/lower the decibel level, etc. I've started writing such a program, but more than once. The likelihood of me actually finishing something grows ever more remote. I do have code in iOS (Objective-C) that plays audio files and applies a homegrown peaking EQ filter -- if any of you are brave enough I'll share by email, it's old enough I'm not sure how much use it will be. In the attached links I am sharing the actual Perl script that built the files that I uploaded today, and you will need to download the freely available SOX program that can highlight the harmonics.

As I've gone into in great length in other threads, but I'm not sure it matters how many times I say it, it is still not going to really sink in for another many many many conversations -- people with perfect pitch DO NOT IDENTIFY PITCHES BY THE OVERTONES. I'm not trying to be mysterious but I also recognize it totally sounds like I've lost my mind.

No overtones + no nonlinear amplitude frequency response curve + no phase-locking = nothing for the infant (or adult) brain to categorize sound into buckets, given the nature of ear's inclination to relative-pitchify the audio world. I think it the phase-locking of overtones that is what makes the identification of perfect pitch possible by true possessors. I mentioned that somewhere, but it's not terribly important and I see no way to scientifically validate it at this point anyway. Really, truly, in the end either some/one of you guys/gals is going to try it out for yourselves and in a few hours/days/weeks/months going to says "holy smoke, something's happened to my hearing!"… or you all will continue to nod your heads politely, as well you should while you just continue to ignore me. I'm ok with that. Carry on. :-) Scientific evidence will follow in 20-50 years, if enough anecdotal evidence is gathered that indicates that maybe, just maybe, just possibly, something could be happening here.

(By phase-locked of overtones I mean simply that the phases are always in a fixed relationship, between a fundamental and its overtones, regardless of the timbre, regardless of the amplitude of the overtones. The zero-point crossing is the same (picture component sine waves added together -- draw them, on paper, yourself). In contrast, playing two different pitches harmonically, they each individually have their own phase-locked component sine wave overtones, but compared to each other, they are not phase-locked, and in fact when you repeatedly play the two harmonic pitches over and over again, they will be slightly differently-timed every single time. For the individual pitch, the zero-point-crossing, the phase-locking, etc. of overtones with respect to the fundamental, completely ignoring the amplitudes, is something that can be detected mathematical (Fourier & company) and it can probably be a trivial neural circuit -- neither of which is something that is an *audible* part of music and therefore this is NOT something a person with or without perfect pitch *hears* -- does NOT hear.)

What we really could use is a good-old-fashioned academic conference, organized by Chris, financed by a kindly old wealthy gentleman, putting us up in a pleasant warm place for a week (I vote for a beach, somewhere in the Caribbean), and we all share our ideas and wave our hands wildly and shout and whisper and jump up and down in front of whiteboards and pianos and guitars and speakers and overhead projectors. I agree with Zach, I think with the talent dispersed among the group that (in)frequents this discussion board, we'd have a good shot at making some significant headway in a week, and I think we'd all end up agreeing on a lot more than we think we agree on by week's end. Maybe Mr. Aruffo can start planning ETC Week Belize 2015, eh?

Back to the "exercises" -- this is a set of intrinsic, not objective; subjective, in your own head, not objective or measurable by anyone else (until we get to brain-cap-brainwave-measuring devices). (P.S. a good friend of mine, another physicist, when I assured him I was "just fine" because I just had a brain MRI and it was all clear… he blurts out "an MRI? now THAT's a BLUNT INSTRUMENT!" laughter ensued… Point being, don't consider an MRI to be of much help in anything we're trying to do yet, it may be a miracle to 20th century medicine but it will be considered quite crude by 22nd century standards and has no bearing on how the brain accomplishes anything at the neural circuitry level…)

Obtain SOX from for your computer platform:

These are example SOX commands that create the 5th overtone emphasis for C5, for example:

sox guitar/guitar-60.aif outindguitar/guitar-60-5-18.aif equalizer 1308 10q 18
sox guitar/guitar-60.aif outindguitar/guitar-60-5-15.aif equalizer 1308 10q 15
sox guitar/guitar-60.aif outindguitar/guitar-60-5-12.aif equalizer 1308 10q 12
sox guitar/guitar-60.aif outindguitar/guitar-60-5-09.aif equalizer 1308 10q 9
sox guitar/guitar-60.aif outindguitar/guitar-60-5-06.aif equalizer 1308 10q 6
sox guitar/guitar-60.aif outindguitar/guitar-60-5-03.aif equalizer 1308 10q 3

So you guys can get SOX all by yourself and just fiddle with it on your own from the command line, use a calculator or a chart to get overtone frequency values, etc. SOX is a seriously useful program in its own right even if you have no reason to ever investigate what I'm talking about in this post.

The script is written in Perl. It's a mess, but it's functional or at least functional enough that I can run it.

*** MISTAKES are not just possible, but very likely, so you are hereby warned ***

If you clean it up, please share your improvements! One immediate problem is that sometimes SOX will chirp to the console that it detected clipping when I bumped up the harmonic too high, but I have not attempted to detect that -- and so some of the audio files are probably trash-worthy, and I'd rather it be handled cleanly.

Here is the Perl script I wrote that takes in .AIFF files (if you're on Windows you'll probably want to do something about AIFF/WAVE right away but SOX will be happy to oblige):

I run everything on my Mac so you will have some tweaking to do if you aren't. Someone with open source experience who wants to take a little responsibility, please feel free -- github or whatever, I am not picky and I give you permission and I'm unlikely to take the time/effort to figure it out, too busy with life's calamities and catastrophies.

The command 'afconvert' is an Apple command that converts an audio file to an Apple lossless file. If you aren't on a Mac… use LAME or something if you prefer. I think I commented it out anyway having had some new problems with OS 10.8.

I used the program AutoSampler to get sounds from my keyboards, soft synths, sampled sounds, etc. You can just records your guitar, piano, flute, whatever you play, as evenly a velocity as possible into recorded WAV files and probably get this script to work with minimal effort (I hope, at least someone here probably can help you).

If this ever leads anywhere… 2 people train 4 train 8… a new program… a new audio CD course… someone does a serious research study and voila! eureka! 9 out of 10 kids have preliminary live perfect pitch! (or don't!) please don't forget to acknowledge Bryce Alexander. or send him a donation, poor guy, born too late and after a marketing genius like Burge. Bryce planted an idea in my head to explore the overtones. I took a different direction, choosing to focus on a single overtone at a time rather than all at once (though you can see in the Perl program how to bring out all of them at once if you like).

Posts: 145
Joined: Tue Oct 23, 2012 3:51 pm
Location: United States

Post by zacxpacx » Sat Dec 01, 2012 10:46 pm

Two quick points:

1. Perceptual differentiation will teach someone to pick out what a pitch fundamentally is, whether or not it's the overtones of base frequency.

2. MRIs may be crude, but they do consistently show areas of the brain activated in perfect pitch listeners that aren't activated in relative listeners when interpreting sound stimulus. Relative listeners must activate these areas of the brain if they are truly using a perfect pitch strategy.

Posts: 52
Joined: Mon Nov 10, 2008 4:39 pm
Location: Dallas

Post by Nikolaus » Mon Dec 03, 2012 4:29 am

research can help, but sitting around for research to be scientifically validated would be like edison sitting around for science to catch up before inventing the light bulb. craftsmen throughout history have NEVER had the luxury of waiting for science to catch up with their discoveries. fact of the matter is i've never seen an ear training course that made even RP truly functional (much less PP). and the reason APA has very little (if any) practical value is because Aruffo never intended the program to be about decoding structure, which is the primary skill required for dictation (PP or RP). and i don't think MRIs would tell us much (beyond a point), because fact of the matter is that we don't know how (or how long it would take) the adult brain to develop the faculties for AP, and i feel that RP is ignored when considering the abilities of absolute pitchers. ditching a method just because it doesn't develop AP perception from the outset i don't think is correct, because for all we know AP (in adults) could be an incremental thing, and not something that just happens all at once. most people give up after only a few hundred hours training for PP, when in reality it merely could be a matter of practice. (thousands of hours of hard work listening to and dictating scores).

Posts: 52
Joined: Mon Nov 10, 2008 4:39 pm
Location: Dallas

Post by Nikolaus » Mon Dec 03, 2012 4:32 am

thanks so much for sharing this, Lyle. you took a subject that has frustrated and overwhelmed me on too many occasions to count and made it surgical and precise.

Posts: 52
Joined: Mon Nov 10, 2008 4:39 pm
Location: Dallas

Post by Nikolaus » Mon Dec 03, 2012 4:54 am

with the (correct) understanding that using a height strategy is not AP, how do we know that implementing it on a large scale wouldn't lead to it somewhere down the line? and how do we know that children weren't using a height strategy at some point (before their AP developed), and that it might be a precursor and prerequisite to AP development? just thought i'd throw the possibility out there.

Posts: 52
Joined: Mon Nov 10, 2008 4:39 pm
Location: Dallas

Post by Nikolaus » Mon Dec 03, 2012 7:15 am

tryin' to think of a good way to handle this...

you say get to the point where you can easily hear one particular overtone (the "target" one, i suppose you could say...) so maybe the loudest track is supposed to indicate which tone i should to be listening for, and then ideally i should get to point where:

piano-60-5-18 (to indicate the overtone)
piano-60-5-00 (to test if i can hear it)

then i add in the other overtones one at a time (into the same playlist), until they're all easy. with the sixth and seventh overtones, usually more baby steps are required:


and then ideally, i should get to the point where i can jump from 18 to 00, and then still hear the overtone. am i wrong? then once i can hear them all, i could do a playlist of only the 18 and 0 tracks, so that i could cover the entire series in about 42 seconds (at 3 seconds per track), before either repeating or moving onto another playlist with another pitch.

but why do this? so (like you said) i will be able to eventually hear note color, and then be able to combine it with scale-degree recognition. in the past i never knew what enabled me to hear note color (other than just sheer exposure to one timbre), so i'm eager to see how it goes with a timbre i'm not very familiar with. and the test to see if i can? i'll use the RP tracks (with random key centers) you so generously provided in the previous thread. if i can get perfect accuracy with those, then you might be onto something here. after that, i think the basic triads are the most important (by far), as opposed to random atonal structures. then from there, it will be a systematic progression til i can easily tell you what chords and pitches are being used in basic church hymns and bach chorales. how's that for a battle plan? :p i'll let you know how it goes (if i can get around to it, because RP ear training occupies most of my time). but i'll throw this out there in case anyone wants to try.

Posts: 52
Joined: Mon Nov 10, 2008 4:39 pm
Location: Dallas

Post by Nikolaus » Mon Dec 03, 2012 7:22 am

must say, though - i question the necessity of listening for certain overtones, because they get completely swamped by the neighboring ones (such as the 6th of middle C)

Posts: 24
Joined: Sun Jul 10, 2005 12:04 pm

Post by Lyle » Mon Dec 03, 2012 2:04 pm

Your statement “get completely swamped by the neighboring ones” gets at what I am talking about. Practice hearing that one overtone which disappears in the mix, the one that gets lost in the crowd. It’s definitely still there; you just lose it when it’s not amplified artificially.

You are not practicing overtone listening so that you will actually listen for overtones. It’s more like, you are stretching before a martial arts workout. Or, you are memorizing your multiplication table before taking calculus. Or… someone come up with a better analogy, please… because you will *not* be listening for overtones when you set about to identify a pitch. Rather, my perception of pitch just “opened up” as Burge would say. More information came to my ear, more sound was present, simply more was available. I figure that the overtone circuitry of my young kid self just masked itself out with time, being rather useless for most daily tasks; it was used so infrequently it just stopped being used. And you don’t notice what your brain no longer processes.

We all have audio “blind spots” that we don’t realize we have, well, because we’re blind to them. By specifically listening for each of (say) the first seven overtones of (say) one octave of pitches, you’re going to discover which ones you hear quite well, and which ones you didn’t realize you don’t hear. When they are amplified, you’ll notice them; as they fade, you’ll have a harder and harder time picking them out from the rest of the pitch’s components. But once you’ve practiced this a little while, you won’t have any difficulty picking out that unamplified overtone. And in my own personal experience, the point at which I could pick overtones out, is exactly the point that the pitches each had an “aura,” a “chroma,” a unique extra-sensory-timbre that made them sound different from one another – C no longer sounded like E. The chroma is part of the whole, the gestalt, the being… you get where I’m going with this, I don’t want you to think that overtones *are* the chroma or vice-versa. I don’t listen for overtones specifically when I’m listening to pitches while I’m playing my piano. In fact I haven't really listened to overtones specifically in many months, until I sat down to make this script & and files. But I still hear C-ness and E-ness.

I suspect us non-perfect-pitchers have blind spots in 99.9% of our spectral hearing. It simply doesn’t matter for music, or speech, or anything in daily life. The exercises here are intended to awaken the neural circuitry that let you focus & become aware of frequencies above the fundamental, while the fundamental is playing, keeping both in focus simultaneously.

As far as height as a strategy for perfect pitch, it just makes no sense to me anymore. The minute that C had such a radically different perceptual character from E… height comparisons would be so much more work. You can appreciate that from a scale-degree standpoint – figuring out a melody interval-by-interval, instead of by scale-degree against a key-center… or by “this much higher” than “that much higher.” The scale-degree-guy expends almost no effort, the other guys’ heads are going to explode…

I wouldn’t go so far as to think of the thousands of audio tracks provided here as a sequence you have to go through comprehensively. In fact I’m sure there’s a better way to do it, via a personalized computer program, which I’m unlikely to ever write – someone else, please, get creative and invent something better. I do think that starting small & constrained is important, and the baby-steps approach is important, and picking the sound closest to your own instrument is best. I’d pick just one note (or at most two) and do the louder-to-softer harmonics until you do hear them in that one single pitch, and work hard to hear them in the same pitch on your own instrument. Then you’re really making progress. You could do what you are saying, make a play list to jump from -18 to -00 as soon as you’ve got the ability to pick out the specific overtone on the specific pitch easily. That’s probably not a bad idea to brush up once in a while. For practical purposes, I’d really make the move to your own instrument as soon as you’ve got the idea down, meaning: learn to play the “guide tone” overtone on your instrument while listening to a lower pitch on your instrument; for piano, play E3 and alternate playing/not-playing B4 & Ab5, etc. and similar for the guitar. Like I outlined in the Frequency Response thread, but with so little responses there I don’t think people got it, so these audio examples ought to make it very clear how you are trying to re-sculpt your hearing.

This kind of ear training is perpendicular to scale degree training, and so completely different, you could put literally a minute a day into it and not interfere with your relative pitch training. Pick two favorite notes in one octave… run those notes in the audio exercises but also on your own instrument, and report back in a month. :-) The first few weeks I was doing it, I couldn’t force my ear to hear a the particular overtone I wanted to hear – sometimes I could, sometimes I couldn’t, I just had to learn how to listen in a particular spot. It’s frustrating, be prepared, you know the frequency is there, you just can't hear it, but you KNOW it's there.

Posts: 146
Joined: Wed Jul 20, 2005 4:20 pm
Location: New Zealand

Post by Axeman » Thu Dec 06, 2012 7:27 pm

Hey there guys I have been trying to follow along. This idea of yours lyle is similar to bryce alexanders program. I found that frustrating (the program) because you had to connect your instrument to the computer. My question is to do with the overtones. I can quite easily hear overtones on my guitar and certain ones come in and out of my perception as I focus on them. (You can play the harmonics above any fretted note by touching the string at certain nodal points to hear what they should be then play the fundamental note again to see if you can hear it in there).
My Question; is the overtone signature the same for each note or do some pitches have certain ones that are accentuated (peaks on a graph of overtones). I know that certain instruments tend to accentuate particular overtones ( I remember reading something about the clarinet in particular in that respect). The reason for my question is that if every pitch has the same amount of overtones then how will this lead to a perceived difference between pitches even if you do hear all of the overtones of each one.
I do like the idea of trying to open up your 'child's ear' (my nomenclature of your method) and I think my post about timbre was along the same lines.
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.

Posts: 52
Joined: Mon Nov 10, 2008 4:39 pm
Location: Dallas

Post by Nikolaus » Sun Dec 09, 2012 8:59 am


how difficult would it be to learn how to use sox to create tracks in the timbre of my own instrument? as it stands now, the documentation is utterly vague and confusing, and i can't even figure out how to open the command line (i click on the sox icon, the command line comes up and then disappears for whatever reason). ideally, what i'm hoping for is that you could give me step by step instructions, but only if it wouldn't be an absolute chore for you to explain.

Posts: 24
Joined: Sun Jul 10, 2005 12:04 pm

Post by Lyle » Mon Dec 10, 2012 6:16 pm


Yes, it was Bryce's work that brought overtones to my attention. I think the way to answer your question is that even if the "overtone signature" is identical for each note for a given timbre, your ear does *not* hear them the same. That's the nonlinear frequency response curve in action (see the Wikipedia article I linked to somewhere). It's effect is ridiculously subtle, though, almost intangible by nature's intent. In a nutshell: what enters your outer ear, is not identical to what ends up in your brain.

I tried to draw a picture once, and might try again someday. I alluded to this in that Frequency Response thread, that this leads to a kind of "anti-timbre" where the result is not just that some overtones that are enhanced, but also that some overtones are slightly *muted*. I think somehow the ear compensates so that you are generally unaware of it, like your visual blind spot where your optic nerve attaches but you don't notice unless you trick your visual system, and yet with training, you *can* become aware of the slight timbral-like quality this gives to the individual pitch class.

To toss out a half-baked notion: It's the anti-timbre that's consistent across pitch classes and timbres, and each pitch class has a unique anti-timbre.

So I suggest you do exactly what you're doing to play/listen for overtones on your guitar, but do them with two and only two (fretted) notes in a low octave. Listen for all seven of the overtones individually, and also (separate exercise) try to just "absorb" the entire spectrum at once as you listen to the two notes back and forth. And see if over the next few weeks something doesn't emerge slightly different from the two notes.


I can try to help. For one thing, you'll need to open up a command line window first. Hit the Start button, and (assuming you're on Windows 7) type 'cmd'. However you're going to probably need to add the path to sox to your search path first, because I just tried this on a windows 7 machine and it didn't find it when I then typed 'sox'. You can google for how to do that, but that might be overly technical, I don't know. Rather than clog up the forums (and because my attendance online here is going to continue be very spotty so you're more likely to get a response) please send me an email and we'll figure it out offline for you, lyle at vaden dot com.

Posts: 146
Joined: Wed Jul 20, 2005 4:20 pm
Location: New Zealand

Post by Axeman » Mon Dec 10, 2012 11:58 pm

Cheers Lyle,
that clears things up for me. I have read about the non linear frequency response thing now that you mention it. Will try your suggestions. I have been listening to some of the mp3 s you created and can hear at leas two of the overtones in most of the 00 level notes. Will keep practicing to expose my ear to more.

Posts: 1
Joined: Wed Feb 20, 2013 8:25 am
Location: United States

Post by Adamml » Wed Feb 20, 2013 8:30 am

I run everything on my Mac so you will have some tweaking to do if you aren't. Someone with open source experience who wants to take a little responsibility, please feel free -- github or whatever, I am not picky and I give you permission and I'm unlikely to take the time/effort to figure it out, too busy with life's calamities and catastrophies.

Post Reply