Yahoo Tuning Groups Ultimate Backup makemicromusic Re: Talismandala: Spectral Tuned Virtual Gamelan

In-Reply-To: <9pvtih+gvhm@...>
Oh, here it is!

> The 12 timbres that were used in this piece were tuned to
> concert "C", normalized and mixed together to create a composite
> wave. An FFT analysis of the composite was performed, and the tuning
> was derived from there. This process also included resythesis of the
> combined waveform to aurally check to see if the program got it
> right.

I suppose taking the most prominent partials from an FFT of each timbre
and putting them all together would have the same effect? Another
approach would be to find the most consonant interval for each pair of
timbres by ear, and combine them all to make a scale.

> Here's the resulting tuning/mode used in the piece and formed from
> the prominent partials found in the FFT and verified by hearing the
> resynthesized waveform:

And when you say "formed from" do you mean you put them all relative to
the tonic?

> The savvy reader will surely realize that because the composite
> timbre had several octaves of mixed inharmonic timbres, that the FFT
> should reveal octaves in the tuning - and yes this is true. Even
> though the partial structures of the individual timbres may not have
> contained exact octaves at 2/1, because they were combined, the
> octave was prominent in the FFT.

Um, what? You mean that because there are so many intervals, some of them
must be close to an octave?

> Here we are dealing with percussion sounds, but I propose that this
> would be highly interesting to explore with electronic timbres as
> well. Of course as Bill Sethares has thoroughly explored, resynthesis
> provides a way to create timbres with common partial structures - but
> what about if one doesn't wish to resynthesize the timbres, as in the
> case of the totally acoustic timbres used in Talismandala? Again, the
> *big question* is how to find a good tuning for disparate selections
> of timbres with radically different partial structures.

With real percussion instruments, the timbre doesn't scale with pitch,
which is something else you could take into account. In your example, the
low gong could be treated like a drone, and other instruments tuned
relative to it. You can do a full Setharian calculation like this, have
the drone constant and vary two other timbres. Which would mean the whole
scale transposing every time the drone changes.

Graham

🔗graham@...

In-Reply-To: <9q4a4r+2ujc@...>
Jacky wrote:

> An important aside: Are the timbres we consider to be harmonic, truly
> and *purely harmonic*, when we consider that resonators contribute
> formant - static frequencies, in sounds like the human voice and the
> violin?

There are two points here.

Are real world timbres purely harmonic? Not all of them, but a lot are
close enough that for musical purposes you don't need to worry about it.

Formants are filters rather than tones. It's like using a fixed filter
cutoff with an analog-style synth. The timbres still harmonic, but a
different one each time. Percussion instruments are very different. I
know my drum is. Every sound has the same fundamental present, but with
the "tak" sounds they're very much like different pitches once you filter
out that fundamental. I haven't checked the FFTs yet, but I wouldn't be
surprised if there were more fixed pitches higher up. It could be that
the idea of pitch is simply a different formant peak.

Graham

🔗graham@...

Jacky wrote:

> I didn't realize that you had/played a drum! May I ask what you have
> there??? Hope you'll post about any FFT results you find. I've got a
> selection of frame drums and I have FFT'd them before, and created
> tunings from the data for tuned samples. Usually I analize about 4
> different takes of the same stroke - dums, taks, snaps, etc.. to get
> a good overall idea about the partials of each.

Well, it's like a darabukkah, don't know how close in sound. I bought it
partly as a study aid for this Arabic music book I've got. I chose it
because I could get a variety of sounds out, including a good boom. It's
a relief after all these years of keyboard playing and microtonality to
have an instrument that isn't mapped by pitch.

I've had a look at some of the more pitched sounds. That fundamental is
steady at 132-4 Hz. Some partials as Kyma shows them vary by about as
much as a whole tone. It seems to be the same partials each time, so
it'll be different formants giving the impression of different pitches.
Also, the resynthesis isn't right at all, so the noise must be important.

I took a lot of samples, and I've mapped them across the keyboard for
further selection. I did take 4 of each, mostly to ensure one of them
would be good, partly so I can double them or crossfade between them.

> I've been frame drumming now for 3 years, playing Tar, Riq, Bodhrans -
> love the frames!
>
> Let's hear some of that drumming Brother!

It's only recently I got a proper microphone. I might do a live, acoustic
backing sometime. But these samples have got real potential. I'll keep
on with the grunt work of selecting and processing them.

Graham

🔗Paul Erlich <paul@...>

--- In MakeMicroMusic@y..., graham@m... wrote:
> In-Reply-To: <9q4a4r+2ujc@e...>
> Jacky wrote:
>
> > An important aside: Are the timbres we consider to be harmonic,
truly
> > and *purely harmonic*, when we consider that resonators
contribute
> > formant - static frequencies, in sounds like the human voice and
the
> > violin?
>
> There are two points here.
>
> Are real world timbres purely harmonic? Not all of them, but a lot
are
> close enough that for musical purposes you don't need to worry
about it.

The human voice and violin are purely harmonic, to the extent that
such a specification even makes sense for a sound of finite duration
and variable pitch.

> Percussion instruments are very different. I
> know my drum is. Every sound has the same fundamental present, but
with
> the "tak" sounds they're very much like different pitches once you
filter
> out that fundamental. I haven't checked the FFTs yet, but I
wouldn't be
> surprised if there were more fixed pitches higher up. It could be
that
> the idea of pitch is simply a different formant peak.

Graham, surely you know that you're merely exciting different
vibrational modes of your drum? One can do similar things with a
guitar string -- if you hit it a certain way, you get a higher note
because you're suppressing, say, all the odd-numbered partials . . .

🔗graham@...

In-Reply-To: <9q5f92+di0g@...>
Jacky wrote:

> > Well, it's like a darabukkah, don't know how close in sound. I
> bought it partly as a study aid for this Arabic music book I've got.
> I chose it because I could get a variety of sounds out, including a
> good boom.
>
> This is a beautiful drum. Is yours metal, ceramic or other?

It's wood, with a skin that looks like it used to belong to some animal.

> > I've had a look at some of the more pitched sounds. That
> fundamental is steady at 132-4 Hz. Some partials as Kyma shows them
> vary by about as much as a whole tone. It seems to be the same
> partials each time, so it'll be different formants giving the
> impression of different pitches.
>
> Very interesting. Isn't that around C?

I think so, mathematically. I don't have anything approaching perfect
pitch.

> A very good point. When I do resynthesis with Wavanal - just one
> program I use for this - it is not able to recreate the noise
> components, but one can tell by its resynthesis that it is getting
> the partials correct (in the sustain stage). I know that this app
> will let you re-enter different values to create another spectrum, so
> it might be cool to try to paste some attack transients onto the
> beginning of the sustain stage rendered by it, to create new sounds -
> but I've not tried it yet. One would have to create a new decay stage
> for this though - easy to do really.

One good thing about Kyma's spectrum editor is that you can hear the
timbre at any time as a sustained note, and drag the pointer backwards and
forwards. This does give some interesting results. Also hear each
partial in isolation, or any combination thereof. So I might be able to
grab a good drum spectrum and use it for another purpose.

> This is what I like to do, is create a velocity cross-fade sample
> instrument. Typically, I'll do something like this:
>
> Midi Velocities
> 0-90
> 91-100
> 101-110
> 111-127
>
> For the four layers of the Akai, which seems to respond well to my
> keyboard and playing force (this is the one used on Talismandala).
> Sometimes I may roll this down 10 though, depending on the samples
> used. Somtimes I like to overlap them as well.

Yes, I might try that with these "tak" samples, as it looks like they can
be treated as a single pitch. Also combine similar sounding mono samples
to get a stereo one. I remember you saying you recorded the whole drum
part twice to do this kind of thing.

Graham