Klarinet Archive - Posting 000698.txt from 2000/11

From: Tony@-----.uk (Tony Pay)
Subj: [kl] Does spectral analysis help?
Date: Tue, 21 Nov 2000 11:08:16 -0500

On Tue, 21 Nov 2000 07:21:00 -0800 (PST), Bilwright@-----.net said:

> Tony Pay wrote:
>
> > What makes me so sure? Why, because I've tried it myself, of
> > course;-)
>
> Aiyeee! Don't leave me hanging, Tony!

My point was rather that I'm not against trying things.

But I really think you're barking up the wrong tree here. It was the
*other* part of my post that was important.

> For example, when you played a purposefully <choose your metaphor>
> tone, was it obvious that the even-numbered harmonics went up or down?

No, it wasn't.

> did the slopes become steeper or flatter?

The slopes of what? You just get a graphical readout of the Fourier
analysis. It's a series of equally spaced vertical lines of differing
heights. Unless you're talking about something different.

> did the ratios between the peaks take on the same relationship each
> time that your ear believed you were playing the same tone, etc?

There seemed to be very little difference. That's probably to be
expected, since a lot of the changes that we are sensitive to are quite
high frequency ones.

> Did you do this on your own PC...

I don't own a PC. I might buy one sometime to run linux.

> ...or in a university lab with megabuck equipment?

It was in the Department of Applied Maths and Theoretical Physics at
Cambridge, but in the 80's, so even if it was megabuck equipment, it was
pretty primitive computer gear by today's standards.

> Was it just a one-time experiment during a cook's tour, or did you
> repeat it until you decided there was nothing more to see? Did you
> keep printed spectra and cogitate about them afterwards?

We were looking at what might be true if a note seemed to sound flat and
sharp at the same time. They subsequently wrote a paper about something
else, but looking at the problem in the time rather than frequency
domain, which is probably sensible, given that transients are very
important, and not captured by steady-tone frequency analysis.

The paper was, 'On the oscillations of musical instruments', McIntyre,
Schumacher and Woodhouse, Journal of the Acoustical Society of America
74 (5), November 1983.

> In your other post, you asserted that the ear delivers better data. Is
> this really true?

It's the *only* data, isn't it? What we want is to play music well, and
our decisions about what constitutes doing that are made by *listening*.
The fact is that there aren't very good empirical correlates of things
like 'good tone' and so on. But surely that's not surprising, if you
think about it.

There may be empirical correlates of 'bad tone', I suppose. For
example, I think of a 'bright' sound as one that has stronger high
harmonics than does a 'dark' sound -- that's the way round it is for me.
Some unremittingly bright (dark) sounds aren't useful at all for playing
music [translate as, some sounds with unremittingly strong (weak) high
harmonics aren't useful at all for playing music] because we need the
variation -- hence our long-ago conversation about what playing music
*needs*. But that's a difference you hear, as in unremitting 'eee', or
unremitting 'ooo'. It's not a property of one steady sound, but a
statement about the range that's available to a player.

By the way, since there's been some discussion about my own recordings,
and I was invited (though hitherto declined) to comment on them, there's
a sort of illustration of what I mean in the CD of Arnold's music
with Jimmy Galway that Walter mentioned, in particular the Trio for
flute, oboe and clarinet. Because these three instruments can imitate,
or 'get close' to each other, and because of the nature of the music
itself, we played the trio with quite a lot of attention to tonal
modulation, both of attack and sustained notes. So you can hear the
variation between 'bright' and 'dark' on that CD particularly.

It's a nice CD, anyway, even if I'd rather have had the slow movements
of that trio a little simpler. Hans Keller once talked approvingly
about Arnold's 'profound triviality', and I remember amusing myself by
telling Heinz Holliger how this particular trio, which he didn't know,
was a 'masterpiece', and he should definitely get hold of it>-)

> Think about all the illusions and dependencies that have been
> described --- for example, pitch perception depends on loudness, and
> so forth.

But all that means is that to get the right perceived pitch you have to
play so that the pitch is perceived to be right:-) If you played it so
that the *frequency* was 'correct', then you'd be *wrong*.

> I don't intend to demean a musician's ear when I say this, but
> examining a graphic representation *has* to provide additional
> information, I would think.

But not about anything that's helpful to you, in learning to play a
clarinet.

> Think about golfers who (presumably) have direct neural connections
> between their brains and their arms, but when they see their stroke in
> a video, they say, "Oh! I didn't realize I was doing that."

But that's feedback to do with the *method*, not feedback to do with the
*result*.

What you're trying to do, on that analogy, is to find a 'better' way to
know whether the ball has gone in the hole.

Tony
--
_________ Tony Pay
|ony:-) 79 Southmoor Rd Tony@-----.uk
| |ay Oxford OX2 6RE GMN family artist: www.gmn.com
tel/fax 01865 553339

.... The surest way to be late is to have plenty of time.

---------------------------------------------------------------------
Unsubscribe from Klarinet, e-mail: klarinet-unsubscribe@-----.org
Subscribe to the Digest: klarinet-digest-subscribe@-----.org
Additional commands: klarinet-help@-----.org
Other problems: klarinet-owner@-----.org