The last edition of the German (snake-oil) HiFi-magazine AUDIO reported on some interesting recent research on hearing. They claim humanity is made up of two types of listeners: the fundamental tone hearer and the overtone hearer. They go on and say that the former is prevalently hearing with the left side of his brain and usually prefers melodic music and higher instruments (piano, violin, flute), while the latter is hearing with his ride side and favours instruments with rich timbre (voice, bassoon, cello).

There is a simple test, to find out to which class you belong. A pair of sounds is played and you decide whether the frequency is increasing or decreasing.

Here is the file: audio.de - overtone.flac
There is a total of 12 pairs. A German-speaking voice introduces each iteration and every pair is repeated once. For each sequence jot down, whether you thought the frequency to be increasing or decreasing.

Afterwards compare your results with this table:

The more your results agree with this list, the more you are a fundamental tone hearer. If your result is the exact inverse of the above, you are a pure overtone hearer.

While I don't agree with the alleged implications of the test (myself being an absolute fundamental tone hearer yet strongly preferring chromatic, timbre rich music), I find the basic idea (two types of listening) intriguing enough to warrant a discussion among a more technology inclined, sceptic crowd.

I figure posting a few test tones for an "academic" discussion should fall under fair use.

Interesting.

With the exception of the first pair I had the completely opposite result and hence must be overtone hearer.

Nevertheless I cannot say to prefer any of the mentioned instruments. However, harpsichord, one of my favourite instruments, tends to be in the first group, also contradicting the implications.

I guess I am more of a fundamental tone hearer

So, if you get totally identical results then you're using your left brain, and if you get exactly inverse results then you're using your right brain?

What does it mean if you get 7 identical results and 5 different results? Does it mean you're using both sides of the brain or does it mean you're using neither?

Have you performed the test?

At least Continuum and I had pretty consistent results.

what did they just pull this out of their ass or do they have any emperical evidence? I haven't performed the test yet, but according to this I would more than likely be more left side of the brain even though psychologically I am not left brained . I know stringed instruments have numerous overtones though. I don't understand how a noise increasing or decreasing in frequency can tell you what type of listener you are from a Psychoacoustical perspective. If I was taking a critical listening class in order to determine ascending or descending scales in music maybe.

Strange, all my results match exactly with the table, but cello has always been one of my favourite instruments because of its overtones

I have. I tried to repeat the tune of the tones by humming, and then I decided if my hums either went down or up. I got 7 vs 5.

9/12 consistency with the table - mostly fundamentals then, if that's to be trusted.

This raises an interesting issue though - I am a competent musician, and I play well by ear, but I've often been fascinated by the fact that I can hear chords very clearly, but not the constituent notes. I can recognise progressions and subtle harmonic changes, and find them quickly on the piano, but I find it very difficult to pick the notes of a chord out in isolation (ie. sing them back). I hear the 'colour' of the chord as whole but not the individual notes very well.

Wonder how this relates to the topic under discussion.

C

I'm a complete fundamental tone hearer it seems, but I can't seem to find any correlation between that and the suggested instruments I'm supposed to be in favour of.

This is a psychoacoustics issue. It's a human science, not a technology related study (unless one models a human as a deterministic machine, which is possible, although there are still black boxes).

It requires understanding of psychophysics and psychoacoustics.

At my current level of understanding, their argument sounds rather odd. If I understand your summary of what they are claiming, I think it's mis-phrased of probably just wrong.

Humans are overtone hearers, based on the research I've read.

All humans can be fooled to hear the virtual pitch with the missing fundamental. This implies that ability to reconstruct the fundamental pitch from the overtones is built-in to the central nervous system.

Now, if they mean that in normal situations, either overtones or the fundamental is more weighted depending on the hearer, this is an interesting argument.

However, I wonder how one can make a test for it: if you remove the fundamental, it's not a normal case anymore. Same goes if you tamper with the overtones.

Now, how do you discern with untampered musical tones (with fundamental and overtones intact), which of them is the more overriding factor in pitch detection for each hearer?

fMRI might produce some data, but I'm not up-to-date whether pitch processing has been located accurately on the cortex (it is a central nervous system processing issue, not an ear issue).



I'm assuming you are referring to the difference between having an accurate relative pitch detection vs. having an accurate absolute pitch detection?

They are different abilities (or skills). Relative pitch can be trained practically by anyone. Absolute pitch may not (by current understanding) be trainable by everyone, but seems to be more of a inhererited trait where early training may have some influence on whether the ability manifests or not.

Or then I completely misunderstood what you were trying to say

6/12...what am i

I matched all 12 with the original poster's.
My favorite instrument is piano/organ, I like melodies, and don't like bass as much.

As a matter of fact, the boundaries between the two extreme types of listeners are reported as being gliding.

Additionally, fundamental tone hearers are believed to be more sensible towards proper signal phase, impulse precision (transients) and "time-right" playback with low runtime differences.

Overtone heares, on the other hand, are believed to be more sensible towards tonality and more 'forgiving' towards runtime differences (e.g. between woofers and tweeters of loudspeakers) that fundamental tone hearers are particularly critical towards.

I find these results highly interesting because it involves some serious research among a group of purposedly mixed listeners like pro musicians, hobby music enthusiasts and everyday people. Measurements include EEG analysis to assess how e.g. intensely the listening cortex of each participant is working during the listening experience.

At this time, it is nothing more than enhanced basic research ... but research is pointing in the right direction IMO.

Yes, you are right. The word technical was meant to allude to the means of studying this effect, that is by things like FFT-analysis or MRI, as I am less concerned with the statistical soundness or the psychological or sociological ramifications.


I'm afraid that's what they did: remove the fundamental tone. Maybe you are right in that all humans can reconstruct a missing fundamental. The conclusion of this test, however, is, that this is done to a quite varying degree.

Without doubt, this is the more interesting question, although it appears to be quite difficult to implement a test for it. The current test can be taken by anyone in just two minutes.


Yes, I should have pointed that out more clearly. They give no exact numbers, but the graph is rather flat with peaks at both ends, so quite different from a random binomial distribution.

I got 11/12. On foobars spectral view the band narrows from the opposite direction of where the tone is said to go, so if it "goes lower", there's actually less energy on the lower end.

I have the same ability to hear the"colour" of a chord, and I scored 12/12.

Yeah, that's not quite what I meant; sorry for the confusion. I meant that if I listen to, say, two or three voices singing in harmony, I can hear very well the intrinsic harmonic relationships and play them on the piano quite easily, yet I couldn't sing you back any of the parts in isolation (except the melody line) very easily. Oh, and the bass line - that sticks out too.

I think that relates to this test in that I could hear that the test tones played two notes at a time, yet I could only 'hear' one part. I guess this tests which one you most clearly hear when there is no clear 'lead' part.

C

I don't feel qualified to judge my own hearing.

I scored 0/12 based on the table (exact opposite)

Audigy 2 NX, AKG K240 headphones (not that it matters)

Pure overtone hearer (2/12).

I pay too much attention to voice and high frequencies. That's almost my notion of "fidelity", odd.

Interesting. With headphones (tried two types) I match the given table 100%. With speakers I tend to say number 7 is increasing rather than decreasing. With a little concentration I can now make 7 sound either way, while from the rest I clearly favor one or the other.

I do like cello very much though.
I am not conscious of liking or not liking music the way the article claims.

6/12

12/12 match with the table, yay! I can hear fundamental tones!

Here is a spectral view of the first pair:



Obviously, they cut off more than just the fundamental!

Here is the sonogram of the sample :



Every trial except the seventh one is made of harmonics with not only the
fundamental missing, but some early harmonics too. The highest is always at the
same frequency.

The harmonics go up or down, while the missing fundamental goes in the opposite
direction.

I got 4 times the fundamental answer, 5 times the harmonic one, and three times
I heard distinctly several notes playing together, some going up, while other go down. I'm quite the opposite of you, Carlos

I got 100% exact answers to the table, so I guess I'm a fundamental tone listener. But I let some of my colleagues try the test, and for a good proportion of the samples, they made the opposite conclusion or were left guessing (ie. they couldn't hear the difference).

I find it weird, that someone who thinks number 1 is decreasing, is a fundamentalist, while in fact the fundamental changes to a higher pitch. Same goes for the other samples.

P.S.: Please excuse the silly pun.

Every sequence is basically a chord in the normal position, followed by another (different) chord in an inverted position. e.g. C-E-G (I), B-D-G (V). As others have pointed out, the "center of energy" has moved down, but the root has gone from C to G.

When I tried the test using my desktop speakers, I tested as a 0/12 (overtones). When I tried it again 10 hours later with my Sennheiser 850's, I tested 9/12 (more fundamental).

What is up with THAT?

JeanLuc pointed out an interesting fact:
I recall not being able to distinguish between my speakers in phase and out of phase (one pair of cables crossed).

This would correlate with my result.

Edit: Typo

It doesn't. The pitch of the (removed) fundamental is the same as the distance between two harmonics, so as they get closer the frequency decreases.

Continuum,

your interpretation is not completely right (at least as long benski's analysis is correct): The fundamental (base tone) is not removed but one octave higher.

You know, that reminds of how speech recognition works for vowel sounds. You measure the harmonics, and pick the harmonics that are louder than adjacent ones. The first of these is the first formant, and the second of these is the second formant. You call those harmonics "formant frequencies."

When you hear [i], [a], or [u], the main difference between the sounds is the difference between the formants. You can hear the distinct vowels very clearly, but not the individual constituent formants.

I wonder how *this* relates to the topic under discussion.

The formants aren't really the individual harmonics. They are simply peaks in the power spectral envelope caused by the resonances in the vocal tract. The individual harmonics are related to the pitch and represent the glottal excitation (the vocal cords vibrating) that gets 'filtered' by the vocal trac.

One of the issues of speech recognition is to recognise words, sub-words, or phonemes, rather than the person speaking. It is known that pitch and related measures are often speaker-specific (and gender-specific, of course) so it is desirable to remove this information for speech recognition (called speaker normalisation). That is why LPCCs (linear prediction cepstral coefficients) were one of the earlier features used in speech recognition, since linear prediction separates the power spectral envelope from the underlying harmonic structure (contained in the residual). Also, MFCCs (Mel frequency-warped cepstral coeffs) which are now the most popular features used in speech recognition are calculated based on the output energies of triangular filterbanks warped on the Mel scale. Hence, nearly all of the harmonic structure has been removed (smoothed out by the filters) and only the envelope/shape is kept.

5/12 - I guess that edges me into the overtone category.

I'm unable to pass this test consistently. When I took it the first time, I percepted it as chords and I couldn't really decide if given test is inc or dec (depends on which tone I focus on). Then I took it the second time and I tried to "sing" these sounds in my mind, then they got somehow more consistent but different than the first time.
So, the results seem to depend on how you come to this test - if you come to it technically, like trying to figure how would you play these sounds on some polyphonic instrument (or how does its spectrum look like), then the results will be different than if you come to it purely aesthetically, like you do while listening to the music.

There is a misunderstanding. I'm talking about the fundamental frequency in the harmonic series, benski talks about the root of a chord (apparently sometimes called fundamental as well).

Cool thing. I am a fundamental tone hearer with a bug. Tone 7 is inverted to me!

hawkeye : continuum's interpretation is correct.

In fact, what happens in the 12 sound couples present in this sample, is that they are made of the last N (N= 2, 3, 4, ...) harmonics of different fundamentals, up to the LCM of the two fundamentals. But they never include the fundamental harmonic.

So the lower frequency the fundamental, the lower the distance between its harmonics.

Because of this, as the LCM (the higher band on the spectrum view of both sounds) is the same, the frequency of the N harmonics just before LCM are higher for the lower frequency fundamental than for the higher frequency fundamentals.

So the sound which has the lower fundamental (the fundamental is the GCD of all the harmonics) has the higher N harmonics before the first common harmonic (which is LCM of the fundamentals of the two sounds).

(LCM = lowest common multiple, GCD=greatest common divider)

For example, take couple number 12.
(I used audacity, hence the screenshots are from it.)
Its spectrum is

The first sound (left) is made of two bands: the lower, around 1328 Hz, and the higher, around 1660 Hz. They are separated by 332Hz.
1328=4*332 and 1660 = 5*332, so these two bands are respectively the 4th and 5th harmonics of the fundamental 332Hz (332 is the GCD of 1428 and 1660). This corresponds almost to a E2 (E2=329,63Hz).
The second sound (right) is also made of two bands: the lower is around 1245 Hz,and the higher is around 1660Hz as for the first sound.
1245 = 3*415 and 1660 = 4*415. These two bands are respectively the 3rd and 4th harmonics of the fundamental 415Hz (415 is the GCD of 1245 and 1660). This corresponds almost to a G#2 (G#2=415,3Hz).

The harmonics of the fundamental of first sound (332Hz) and of the second sound (415Hz) join at 1660Hz, as 1660 is the LCM of 332 and 415.

A fundamental tone hearer interprets the first sound as the 4th and 5th harmonics of the fundamental at 332Hz, a somewhat detuned E2, and mentally reconstitutes the missing fundamental. Try hearing the first sound of the 12th couple and playing an E2 on an instrument.
After that the fundamental tone hearer interprets the second sound as the 3rd and 4th harmonics of the fundamental at 415Hz, an almost perfect G#2, and mentally reconstitutes the missing fundamental. Try hearing the second sound of the 12th couple and playing a G#2 on an instrument.
As you see the fundamental hearer hears the second sound higher than the first in the 12th couple.

An overtone hearer interprets the first sound as the combination of two sounds: a pure harmonic at 1328 Hz (a slightly higher E4) and another pure harmonic at 1660Hz (an almost perfect G#4).
After that the overtone hearer interprets the second sound as the combination of: a pure harmonic at 1245Hz (an almost perfect D#4) and the same almost perfect G#4 at 1660Hz.
As you can see the overtone hearer hears, in both sounds, two notes, one of which is lower in the second sound (the D#4) than in the first sound (the E4).


--- A second exercice: how to hear like an overtone hearer when you're fundamental tone hearer and vice versa ?

A) you're an overtone hearer and wanna understand.

Create a sound which implements all the harmonics of the fundamental, including the fundamental and, of course, the two that you hear as distinct sounds.

Here, I created a sawtooth signal of frequency 332 Hz (E2), followed by the first sound in the couple 12 of the test, then a sawtooth signal of frequency 415Hz (G#2), followed by the second sound in couple 12.

Sorry, no flac, just a mono WAV file.
The spectrum is as follows


You have the kind of signal that a fundamental tone hearer reconstitutes (timber excepted), down to the fundamental tone (lowest bar in the "ladder").

B) you're a fundamental tone hearer, you wanna understand (as me)

Create a sound separating the two components of the sound couple 12.

Here is a sine signal of frequency 1328Hz (the E4), followed by a sine of frequency 1660Hz (the G#4), followed by the first sound of couple 12, which is their combination.
Then, a sine of frequency 1245Hz (the D#4), followed by the same G#4 at 1660Hz, followed by the second sound of couple 12, which also is their combination.

Here again, no flac, just a mono WAV file.
The spectrum is as follows


You have the decomposition that an overtone hearer gets.

Amusingly, when performing that second experience, I had difficulty returning back to my 'fundamental tone hearer mode' on the sample ! (talk me about rewiring the brain)

Note : as I'm a fundamental tone hearer (12/12), I'm not sure the example sound in A is well suited for overtone hearers. I myself don't hear the different sounds composing the harmonics. But I can, with exercice B. One of my coworkers heard the distinct frequencies, and still managed to get 12/12 when I asked him to search the fundamental tone. Great !

Hope it helps a little. All in all, I lost a nice day .

Vincent

If your an fundamental tone hearer and want "hear" it like an overtone hearer, concentrate on the tests long enough, and you'll soon get the "overtones only". (It's like that 3d cube thing - it can be a cube or a corner).

Wombat : note that the 7th couple, on which you fall, is exactly the one having a difference in that the harmonics of sound having the lowest fundamental are taken at LCM and LCM+1*fundamental instead of LCM and LCM-1*fundamental. This raises the energy in higher frequencies, and thus confuses more the listenner.

It must be because you were not exactly at the same distance of both. In a shop, I did not noticed it until I moved in front of the listeners in order to have the speakers positionned symetrically before me. The inversion is then obvious.

I did not make myself clear enough I guess ...

So-called 'Fundamentals' are believed to be very sensible to phase shifting of a kind that is e.g. caused by steep crossover networks in the crossover frequency range ... that's why these people might prefer 2-way speakers with a simple 6dB/octave crossover.

An inverted phase of one channel will most likely lead to unknown interferences and frequency band cancellations (especially in the low-end beneath 200 Hz) which would be a rather tonal issue - a plain lack of bass ...

Thanks everyone who helped me understand. Especially vinouz for the detailed explanation.

Well, it seem I'm a 100% fundamental tone hearer which probably true because I prefer piano to cello.

Since some people hear differences with speakers or headphones i tried it also and can´t repeat it. I always have the same scores and tone 7 is still inverted of course.
Maybe i drunk something away on the left part of my brain

Edit: Brain parts

same as the table except for #5 (11/12)
apparently I'm a fundamental tone hearer

I just find it interesting how different people hear things differently (or interpret what is heard differently)

looks like I'm a fundamental one too

nr. 5 is more a guess for me, I do hear a difference but I can hardly say if it's going up or down. Well, my ears aren't trained at all, so this might be the problem here. Nr.7 is inverted of course, everything else is as in the Table.

0/12 so according to this I'm a pure overtone hearer. The notion of the preferred instruments though, is only somewhat correct. On few of the samples I wasn't sure and I made the choice based on the first impression. On those samples I could easily think that it was the opposite really.