Re: The word timbre: origins


Subject: Re: The word timbre: origins
gogins@pipeline.com
Date: Thu Sep 30 2004 - 12:06:31 EDT


Linear PCM has a dynamic range of 90 dB. Floats have 24 bit mantissas so
the actual precision is much much greater -- 138 dB. This is slightly more
than the dynamic range of human hearing (about 130 dB, as good or better
than the very best microphones), and top end pro equipment (few have it)
can almost reproduce this.

However, the dynamic range and precision have to accomodate a series of
multiplications, one or more at each stage of the processing chain. Each
integer multiplication adds a small amount of quantization noise, so the
total accumulated processing noise must be subtracted from the dynamic
range of the format to give the actual effective dynamic range. Using
floats or better yet doubles for these multiplications is much less noisy.

Original Message:
-----------------
From: Richard Wentk richard@skydancer.com
Date: Thu, 30 Sep 2004 16:47:17 +0100
To: cec-conference@concordia.ca
Subject: Re: The word timbre: origins

At 08:14 30/09/2004 -0400, you wrote:
>I think that this is where I have always had some difficulty with
>Schaeffer, and with spectromorphology -- the attempt to define the limits
>of spectrum without a corresponding (and simultaneous) examination of
gesture.

I think this was a very damaging idea. In reality *all* sounds are like
this. You're not going to find a genuinely steady timbral state anywhere
outside of a quartz-locked waveform generator.

>A bus drives by. Its spectromorphology is gestural. The spectrum (timbre)
>is not fixed and no one (small) window will provide enough information to
>establish the sonic identity, and by the time one has a large enough
>window, the timbre has become (IMV) gestural.

I don't think is different to a flute.

There's a huge difference between the amount of information needed to
recognise a sound as 'flutey' and the amount that's present in a real flute
performance - which will be a collection of disparate timbres connected by
all manner of complex articulations, and continuous and discrete pitch
*and* timbral changes.

The problem with a lot of synthesis theory is that it's based on these
false simplifications, and the mistaken assumption that you can detach
timbre from pitch from other elements.

In practice you can say 'That's a flute' from a single note played at a
single intensity. But it's a mistake to assume that this simplified example
offers a complete and final definition of the flute timbre.

But this is what you'll still find in many textbooks that are happy to
print a certain harmonic distribution with a note that says 'Flute'
underneath it.

For a complete definition of timbre you need a map of timbral changes with
both pitch and intensity, subject to gestural modifiers like the various
different kinds of articulation.

When people hear a flute I think what actually happens is they refer the
sound to the various timbral maps of this type that they carry around
inside their heads, probably aided by some kind of built-in smart-ish
interpolation and sonic pattern recognition. So what you get is a match
*within a map that includes gestural details and possibilities* and goes
far beyond the simple idea of a static, or even a simply varying, overtone
signature.

Richard

--------------------------------------------------------------------
mail2web - Check your email from the web at
http://mail2web.com/ .



This archive was generated by hypermail 2b27 : Sat Dec 22 2007 - 01:46:04 EST