Hello. Since this isn't really 'Music Theory' in the traditional sense, I thought audio experts would be better able to help (since it relates to EQ, frequency ranges and timbre). There's a discussion in the Music Theory section about vowel formants and their association with pitch...

I have always thought that the spectral (pitch) content of a vowel (or timbre) is relative to the fundamental frequency of a given tone. So, a clarinet produces (mostly) odd overtones 1:3:5,7, etc whatever the sounded pitch. Similarly, if I were to sing an "Ah" vowel, it involves a similar envelope of harmonics when sung at similar pitches (not in different octaves though).

This would mean that the spectral pitches associated with each vowel/timbre change depending on the fundamental. But there is a notion that vowels are associated with fixed spectral pitches:

Quote Daniel Davis:

---

Vowels are precise pitches (or formants to be more precise), which suggests that everyone has not just good relative pitch, but actually perfect pitch else we couldn't perceive vowels.

---

(This is also echoed elsewhere online and wikipedia.)

What I'm not clear on is, if I twang an "Ah" vowel on my jaw harp (for instance) how can the pitches in the spectrum be absolutely fixed for this vowel? I mean, how can an "Ah" on a smaller harp produce the same exact (fixed) pitches, because it's a different spectrum coming from each jaw harp? Also, if you re-pitch a jaw harp recording, shouldn't we hear different vowels if this were true?

Is there really an exact set of frequencies that enables us to recognise each vowel? I can sort of understand that frequency ranges might be involved (i.e. the skill required for EQing) for complex timbres but precise pitches for vowels?

I don't know if I've got a mental block, but something isn't right. I'm not really clear about the association between a vowel formants and timbres either. Can any acoustics experts shed any light on this please?

Thanks

Your mouth acts as a filter. Play with an old MS20 for a while and you'll get the idea!

I know that. I'm asking about something a lot more difficult.

The question is about fixed pitches for vowels, regardless of the fundamental pitch.

Okay, let me put it another way...

You press a key on a synthesizer, then you play with the filter and it produces vowels that you can hear are associated with frequencies in the harmonic spectrum of that tone. These are precise pitches, yes...

But if I press a different note, surely the vowels produced by the filter on that tone will be caused by emphasising different 'precise' frequencies corresponding to the spectrum of that tone?

So, (if this is true) I can't see how specific frequencies could relate to vowels regardless of what note we press on a synthesizer. Is it because of how close harmonic frequencies become as we ascend higher in the spectrum of a tone? Are these 'micro' bands/ranges we're talking about here rather than "fixed pitches"? I do understand that when we EQ something, the 'nasalness', 'chestiness', etc, derives from attenuating fixed frequency bands. So, I'm really trying to grasp how these two aspects relate.

Thanks

Filters can be additive, as well as subtractive. The MS20 for example has strong resonant filters and you do the same, when you change the shape of your mouth.

And resonant filters, brings us back to formants. Woopee!

Thanks, I really do understand that.

It's really about how different vowels occur from different fundamental tones. Surely, if I hear the vowel "ah" from the note G, I will also here the vowel "ah" when I press the E below because it's the same envelope of harmonics that forms the vowel, not because of fixed high frequencies that are the common to the spectrum of both notes.

Cheers.

IMO, a great deal of formant theory is just theory looking for a subject!

The human ear is more than the sum of all frequencies and a bit of maths bolted on the back of it. Behind the ear, there is a massive brain, the largest brain in the animal kingdom!

And we do not know properly how that brain actually works!

We use aural filters to know in part where sounds are coming from, who is making a noise and what, if anything that noise means. We do that by using our brains. We hear a sound coming from another human and we immediately can picture how they are forming their mouths. We can judge (to some extent) age, sex, size and even disposition, simply by the sound. We can actually hear if they are smiling - and in fact tests have shown that we can hear if they are smiling very, very easily!

Knowing the difference between Ah and Oh is vital to our very survival as a species. The filters in the mouth tell the listener what is being said, who is saying it and if we should get ready to shag them or club them to death. Even in total darkness, if we hear that they are smiling and they are in the club-to-death group, we stay our hand and put the club down. If they are in the shag group, we lift the kilt. The filters in the outer ear tell us if they are behind us or in front of us.

All this is made possible, not by the frequencies held within the sounds, but by the way our brains process those frequencies, based on experience.

I think it clicked somewhere in the back of my mind when I wrote: "Are these 'micro' bands/ranges we're talking about here rather than fixed pitches?"

I'm used to dealing with the lower harmonic series as musical notes (e.g. on paper or instruments) and the higher harmonic series as frequency bands using EQ.

My fuzzy noggin has now formed a link (made out of brain-meat) between the two...

As you go up further and further up into the harmonic series, the distance between the spectral frequencies become so close as to fall within the 'just noticeable difference' (JND). The ear cannot distinguish between the highest frequencies of an E fundamental and the closest equivalents in a G fundamental spectrum (for instance). Rather it is narrow bands of these high frequencies (that cross through the spectrums of all fundamental tones or keys we press on a synthesizer) are associated with vowel formants.

So technically, we can only talk of 'precise pitches' when we deal with an individual fundamental, but because the narrow frequency bands that determine vowels encompass almost identical frequencies (inside the JND) people refer to the bands as "precise pitches" as though they are independant of the fundamental notes.