Re: [Phys-l] sound stuff

[John Denker <jsd@av8n.com>]

Anthony Lapinski wrote:
> I have a few questions for the group:
>
> 1) How do you add two different decibel levels?

> Must you convert it to intensity first? 

Yes, if the sounds add incoherently.

In the less-common but certainly important case where the sounds add
coherently, you must convert to amplitude and then add the amplitudes
(with due regard to phase).

> For example, 65 dB + 80 dB
> ? 145 dB. 

Not a chance.

> delta B = 10log(80/65) =
> 0.9, and so the answer is 65.9 dB?

You need to take the exponential of the dB number, not the log.

10*log(10^6.5 + 10^8.0) = 80.1


> 2) If a car sounds a horn. If you hear 450 Hz as it approaches and 350 Hz
> as it recedes, the actual frequency is about 394 Hz (not 400 Hz).
> Conceptually, why isn't the Doppler effect result symmetric for when the
> sound (or observer) is approaching/receding the observer (sound)?

Because 1/x is not a linear function of x.

Consider the extreme case of the car moving very near Mach 1.
The sound of the retreating car goes to frequency=0, a 100% change.
The sound of the advancing car goes to frequency=infinity, a much larger change.

> 3) All musical instruments produce overtones (harmonics) except tuning
> forks and some electronic pianos. How can the human vocal chords produce
> pure tones? Living things are both delicate and complex, yet it is easy to
> sing into a microphone and see a nice sine wave on an oscilloscope.

This illustrates what Bob Silsbee calls "the peril of the eyeball FFT".

You can't tell by eyeballing the oscilloscope whether the "nice" wave
is a pure sine or not.  You can easily detect any aperiodicity, but
not the presense of a harmonic if it is a few dB down.

A human _whistling_ is much more nearly a pure sine than anything coming
out of the vocal cords.