Klarinet Archive - Posting 000145.txt from 2002/11

From: "Jeremy Yager" <bomber@-----.com>
Subj: [kl] dB (was tuners)
Date: Sun, 3 Nov 2002 01:22:09 -0500

To expand on Mark C.'s arguement:

The dB scale is a logarithmic scale--it can represent great changes in
values in a given context without having huge numbers (or exceedingly small
ones).

The human ear percieves sound energy on a logarithmic scale--making the dB
ideal for describing human hearing. This is why you can play a stereo (or a
clarinet) at a given volume, and adding a second acoustically identical
source (another identical stereo or clarinet) and it will not sound much
louder. There is actually twice the energy present, but since your ear
hears on a logarithmic scale, it does not sound significantly louder as
'twice' on a log scale is very little difference--3dB, as stated before.
Think of the 120dB jet engine and the 60dB conversation that most folks are
familiar with.

The only difference between that and dB in the hands of recording engineers
(and electricals, too) is the definition of the baseline. On the scale used
by hearing health folks (and most people are familiar with), 0dB is nearly
perfect silence--almost no sound energy, yielding values >0 for any audible
sound.

When used by recording engineers, 0dB (unity gain) is different--recording
equipment systems (IIRC) are defined to be the ideal recording level, and
consistantly setting the equipment to record at >0 dB levels can result in
distortion in the loud sections, and recording levels <<0dB, system noise
(i.e. tape hiss) becomes a problem.

The dB scale can actually be used with any baseline--it is a purely relative
scale--you can sensibly say that one signal is at -10dB compared to another.
What Dr. Pyne seems to have done is set 0dB to the RMS amplitude of the
fundamental frequency, and all of the harmonics are merely given in relation
to the fundemental--they are all less than the fundamental, so the dB values
are all negative. RMS is 'Root Mean Square'. The average value of a pure
acoustic signal is 0, and thus meaningless, where RMS gives a positive value
in all cases, and thus better describes how much energy is in a signal (the
harmonic, for example) of interest.

----- Original Message -----
From: "Mark Charette" <charette@-----.org>
Subject: RE: [kl] Tuners

>
> Perhaps someone else here does?
>
> ---
> You need to read carefully. The scale is dbVrms, where 0 db references the
> microphone nominal output. 0 db is just a reference level.
>
> If you also check carefully, you will find that some even partials are at
a
> higher level than neighboring odd partials.
>
> "1/2" or "3/8" are irrelevant fractions on a log scale ... the whole idea
of
> using decibels and log scales is to amplify small differences. A 3 db
> difference mean twice the power, a 10 db difference 10 times the power. A
40
> db difference (as in the 1st chart) means a 10,000 times difference.
>
> In a different type of scenario, and extremely simplified, if a trumpet
puts
> out a maximum 90 db sound level and a clarinet a maximum 70 db sound
level,
> you'd need 100 clarinets to equal the sound level of one trumpet ...
>
> Mark C.
>
>
>
>
> ---------------------------------------------------------------------
>
>

---------------------------------------------------------------------