chop, chop, chop...Timbre

Dear Tuners,

Altho this may seem a little off subject, I don't really know where to turn
to, and who can awnser my question, and this is about, and related to light
and sound.

There has been recent posts to the alternative tunign list about this
phenominon of synasthesia, and I wanted to ask something about the light
sound relationship.

Sure, we know sound has a timbre, but does light have a timbre?

If light does have a timbre, then why do we always see pure colors?

---Sarn.

Hi Sarn,
> Sure, we know sound has a timbre, but does light have a timbre?

> If light does have a timbre, then why do we always see pure colors?

That's an easy question, or I think so.

We have thousands of ear hairs each sensitive to a different pitch.

However, we have only three types of colour sensor in the eye
(or two if colour blind, or four for some women, who can have two
types of red sensor, one from each parent, while men always have only
one)

So the colours have to be pure. It is as if one could hear only
three frequencies, and got all the colour timbres from the relative strenght
of those three frequencies.

Maybe others can fill in on the details, but I think that is the
basic idea.

Also suggests an idea. If one could turn colours into sounds by
doing an analysis of _all_ the frequencies involved, then one
could hear differences in colour that one couldn't tell apart
by sight - could actually be quite a useful tool perhaps.

Robert

Sarn wrote,

>> Sure, we know sound has a timbre, but does light have a timbre?

>> If light does have a timbre, then why do we always see pure colors?

Robert wrote,

>We have thousands of ear hairs each sensitive to a different pitch.

>However, we have only three types of colour sensor in the eye

>So the colours have to be pure. It is as if one could hear only
>three frequencies, and got all the colour timbres from the relative
>strenght
>of those three frequencies.

In order to answer Sarn's question, I'd go a little further, though. To each
point in the visual field, the eye and brain assign only one color. This
color is, as Robert says, dependent on the relative strengths of the
responses of the three color sensors. These color sensors have overlapping
ranges of sensitivity, so that any possible electromagnetic frequency within
the visible range will be seen as a certain color. However, combining lots
of frequencies will still result in only one color -- it's still only the
strengths of the responses of the three color receptors that affects what
color we see. By using lots of little red, green, and blue pixels
(corresponding to our three color receptors), our TV screens and computer
monitors can simulate almost any frequency or combination of frequencies
that occur in nature. Our eye won't know the difference.

With sound, the situation is completely different. We can hear multiple
instruments at the same time, and they don't necessarily combine to sound
like a single instrument (that only happens if they're playing simple otonal
chords including the unison). There is this phenomenon of timbre which has
no analogue in the visual realm. If the ear is sensing a bunch of
frequencies, the brain will parse these into approximate harmonic series and
assign to each a timbre based on the relative strengths of the partials in
the series, and a pitch based on the frequency of the fundamental (even if
physically absent) of the series. This probably evolved as a result of vocal
sounds being primarily composed of harmonic spectra, and an ability to
separately hear various vocal sounds provided an evolutionary advantage.

Visual spectra of chemical elements result from electrons in atoms jumping
between energy states. A given atom will be able to produce various
electromagnetic frequencies corresponding to the differences between these
states, so a given atom will have something roughly analogous to a "harmonic
series". However, (a) the visible range is very narrow, and most atoms won't
have more than one or two frequencies within the visible range; and (b) most
materials are composed of molecules, rather than isolated atoms -- and
molecules have a much more complicated set of energy states. So we see that
it would be really difficult to conceive of something like a "timbre" for
visual stimuli, and even if we could, we did not evolve in such a way as to
be able to see "timbrally" . . .

Synaesthesia is a completely different matter, though . . .

Hi Paul,

> Visual spectra of chemical elements result from electrons in atoms jumping
> between energy states. A given atom will be able to produce various
> electromagnetic frequencies corresponding to the differences between these
> states, so a given atom will have something roughly analogous to a "harmonic
> series". However, (a) the visible range is very narrow, and most atoms won't
> have more than one or two frequencies within the visible range; and (b) most
> materials are composed of molecules, rather than isolated atoms -- and
> molecules have a much more complicated set of energy states. So we see that
> it would be really difficult to conceive of something like a "timbre" for
> visual stimuli, and even if we could, we did not evolve in such a way as to
> be able to see "timbrally" . . .

Interesting.

Suggests the idea of listening instead to the Fraunhoeffer lines in a star spectra.

Because of the high temperature, these are mainly single atoms or very simple
molecules.

They are absorption lines, but one could invert the spectrum first to make them
the ones one hears.

Or one could do the same thing for emission nebulae, and the region close to
the sun where atoms are at a very high temperature and emit instead of
absorb.

Also as you say, one could map a much wider range than our visual range into the
hearing octaves.

Listening to colours we normally perceive might give a fair amount of "coloured noise"
because of the tendency of surfaces to reflect a wide range of frequencies, e.g.
white I suppose would sound like white noise.

I wonder if anyone has done the experiment, surely they must have. I know that
radar is often turned into sound, but haven't heard of it being done for a
region of the spectrum including light.

I wonder, perhaps as you say, few of the colours we see would be perceived as a kind of timbre
However, reason I think would be that most will have a large continuous component.

I think if the lines were discrete, would be perceived as a timbre, either
harmonic or inharmonic.

I'm not sure I'm convinced that only harmonic series can form a timbre. What about
bell sounds. Bell makers use the sequence
1/12, 1/6 1/5 1/4 1/3.
http://www.oakcroft13.fsnet.co.uk/lehr.htm

That looks as if it has been selected from the utonal rather than the otonal series.

Plus, one just does perceive instruments with inharmonic timbres as having an identity
as an instrument does one not. Here is anohter suggestion: if one hears a note,
then hears the same pattern of partials translated to another pitch, maybe somehow
one keys into that pattern as a timbre? Probably with the harmonic series pattern
being favoured as it is so prevalaent, and as you say, it would have evolutionary
signficance for picking out voices, but I imagine it isn't the only one that
will work.

After all, one also wants to pickout other kinds of sounds, say, the sound
of a branch or twig breaking under foot, which has a kind of characteristic
"timbre", while varying in pitch, and I don't think it is particularly harmonic.

Or, the sound of the wind, and other natural sounds, which may have been highly
significant for early hominids!

Robert

Hi Robert.

I guess what I mean by timbre is specifically a "pitched tone".

A harmonic series will sound like a single pitch, and a specific timbre.

An inharmonic series like a bell may have a specific timbre, yes, but it
will be much more indefinite as to pitch. You'll hear several pitches mixed
together -- some real frequencies, and some created by the brain's effort to
find harmonic series in the frequencies.

>Bell makers use the sequence
>1/12, 1/6 1/5 1/4 1/3.
> <http://www.oakcroft13.fsnet.co.uk/lehr.htm>
http://www.oakcroft13.fsnet.co.uk/lehr.htm

>That looks as if it has been selected from the utonal rather than the
>otonal series.

Did you get that series directly from the article or did you do a lot of
octave-fudging to get it??

>if one hears a note,
>then hears the same pattern of partials translated to another pitch, maybe
somehow
>one keys into that pattern as a timbre?

Maybe after a significantly long training period.

>Probably with the harmonic series pattern
>being favoured as it is so prevalaent, and as you say, it would have
evolutionary
>signficance for picking out voices,

Some think it's learned in the womb.

>but I imagine it isn't the only one that
>will work.

Maybe you're right, but you'd have to filter out the sounds of voices,
reeds, strings, and brass out of your listening diet for a long time, and
even then it might not "happen" . . . anyway, we're agreed that there is no
visual analogue to timbre since a combination of many frequencies looks just
like one frequency with a certain saturation and luminance . . . right?

Hi Paul,

> Did you get that series directly from the article or did you do a lot of
> octave-fudging to get it??

Yes, directly fromn the article, and the author says that bell tuners specifically
target those ratios:

"
This latter idea turns out to be possible. After the introduction of the template in the beginning
of the fourteenth century, bell science gradually developed and made it possible in the beginning of
the seventeenth century to give the lowest eigenfrequencies of the bell the ratios 5:10:12:15:20. It
is interesting that these can also be expressed as 1/12: 1/6: l/5: 1/4: 1/3-a remarkable fact that
indicates that the minor third is dominant in the bell's sound.
"

See the section TUNING THE BELL

> Maybe you're right, but you'd have to filter out the sounds of voices,
> reeds, strings, and brass out of your listening diet for a long time, and
> even then it might not "happen" . . . anyway, we're agreed that there is no
> visual analogue to timbre since a combination of many frequencies looks just
> like one frequency with a certain saturation and luminance . . . right?

Perhaps it can happen with no training period if the sounds already have some
close resemblance to natural sounds? For instance, percussion instruments
will resemble natural percussive sounds?

Just an idea, and a development of your evolutionary suggestion to
inharmonic timbres.

Are there any natural sounds that generate utonal series I wonder,
to sort of follow on from that?

Interesting extra thing about colour. One can also see colours that
are outside the visual range of single frequencies in a way. I'm thinking
of magenta, which one sees if one has red and blue, but very little
in the way of green.

So colour isn't quite linear, more like a circle, as one
sees in the colour wheel.

Then as you say, one has saturation and luminance. The luminance
just corresponds to volume, while saturation is like the
percentage of white noise mixed in with the frequency.

Except, since one is only sensitive to three frequencies,
playing all three frequencies at the same volume will
sound the same as white noise.

But perhaps this is pushing the analogy a bit too far?

Robert

Robert wrote,

>Perhaps it can happen with no training period if the sounds already have
some
>close resemblance to natural sounds? For instance, percussion instruments
>will resemble natural percussive sounds?

There's no regularity to these spectra, in the way that all harmonic series
are the same.

>Are there any natural sounds that generate utonal series I wonder,
>to sort of follow on from that?

No. There are natural phenomena (like throat singing) that general utonal
series _successively_, but not as simultaneities (you may want to search for
"subharmonic", etc. in the archives).

>Interesting extra thing about colour. One can also see colours that
>are outside the visual range of single frequencies in a way. I'm thinking
>of magenta, which one sees if one has red and blue, but very little
>in the way of green.

>So colour isn't quite linear, more like a circle, as one
>sees in the colour wheel.

Yes -- that's because one of the three color receptors (either blue or red,
I don't remember which) actually has two response peaks, one at each end of
the visual spectrum. So this essentially closes the frequency range into a
circle.

>Then as you say, one has saturation and luminance. The luminance
>just corresponds to volume, while saturation is like the
>percentage of white noise mixed in with the frequency.

Exactly.

>Except, since one is only sensitive to three frequencies,
>playing all three frequencies at the same volume will
>sound the same as white noise.

Right.

The cool thing is that there are colors in nature that you can't quite
reproduce with an RGB monitor. Do you remember that discussion?

-Paul

According to recent theories this is how color perception (in first
aproximation) works:

There are three types of (color sensitive) cones in the retina. These have
single peaks in their wavelength sensitivity. They are mostly referred to as
L, M and S (for long- middle- and short wavelength sensitivity. The peaks
are aprx. at 580, 540 and 440 nm.
The first order neural signal processing in the retina combines the cone
signals in three signals according to : L+M+S, L-M+S and L+M-S
L+M+S gives you the anachromatic lightness response
L-M+S gives hue information along the red-green axis
L+M-S hue along the yellow-blue axis
The red-green response has the two peaks Paul referred to.
This ends the controversy about which three colors are 'the' primary colors
(red,green,blue) or (red, yellow,blue), because human vision is opponent
based on four primary colors, being two pairs of complimentaries.

Kees

----- Original Message -----
From: "Paul H. Erlich" <PERLICH@ACADIAN-ASSET.COM>
To: <tuning@yahoogroups.com>
Sent: Friday, April 06, 2001 12:39 PM
Subject: RE: [tuning] Re: chop, chop, chop...Timbre

> >So colour isn't quite linear, more like a circle, as one
> >sees in the colour wheel.
>
> Yes -- that's because one of the three color receptors (either blue or
red,
> I don't remember which) actually has two response peaks, one at each end
of
> the visual spectrum. So this essentially closes the frequency range into a
> circle.
>
> >Then as you say, one has saturation and luminance. The luminance
> >just corresponds to volume, while saturation is like the
> >percentage of white noise mixed in with the frequency.
>
>

--- In tuning@y..., "Kees van Prooijen" <kees@d...> wrote:
> According to recent theories this is how color perception (in first
> aproximation) works:
>
> There are three types of (color sensitive) cones in the retina.
These have
> single peaks in their wavelength sensitivity. They are mostly
referred to as
> L, M and S (for long- middle- and short wavelength sensitivity. The
peaks
> are aprx. at 580, 540 and 440 nm. >>>>

Hi, the following should be helpful:

http://www.cs.unc.edu/~majumder/color/paper.html

Haresh.