If all you mean by "speed of sound" is "how fast this particular

disturbance travelled from A to B", then you're right, but that isn't

the usual (or useful) definition, because it only applies to the one

case. The usual definition of a shock wave is one where the entropy

density is significantly increased by its passage.

I was merely complaining that Wikipedia presents the small-signal

approximation of the speed of sound (and the rest of the small signal

sound properties) without properly explaining what the approximation

is and what its limits of validity are. Not to mention a complete

lack of explanation as to why the speed of sound is what it is.

A drawback of the standard convention of referring to the small-signal

speed of sound as the speed of sound is that there is a tendency to

apply it (and the other small-signal sound properties) to all sound

with the possible exception of shock waves or sound loud enough to

clip at zero absolute pressure. This is like assuming that the small

signal response of an amplifier applies right up to clipping at the

power supply rails - it just isn't so.

Real finite amplitude sound has losses. The primary losses, or

"increase in entropy density" if you prefer, are due to the conduction

of heat from the higher pressure, hotter part of the wave (peaks) to

the lower pressure, cooler part of the wave (troughs), converting

mechanical energy into heat. These losses are insignificant over

short distances within the frequencies of human hearing and at

comfortable listening levels, but they become significant at high

frequencies because the peaks and troughs get close, putting an upper

limit on the frequency of ultrasound which will propogate a

significant distance in air. At high enough frequencies a transducer

will simply heat the air in front of it, increasing entropy density.

Over long distances this effect selectively attenuates higher

frequencies even within the range of hearing, again increasing entropy

density. Likewise where the sound pressure approaches atmospheric

pressure the temperature differences between peaks and troughs

increase, losses increase significantly - and the speed of sound

becomes significantly non-constant.

So your "usual" definition of a shock wave applies to non-shock finite

amplitude sound also, and is probably why no source I consider to be

authoritative on the subject uses it (e.g. Shapiro - The Dynamics and

Thermodynamics of Compressible Fluid flow, Thompson - Compressible

Fluid Dynamics, Blackstock - Fundamentals of Physical Acoustics). All

of these sources use what I regard as the usual definition,

essentially a "large" change in state variables (pressure,

temperature, density ..) in a very "thin" layer (or "short" time

depending on frame of reference). Entropy density also increases, not

because that is a defining characteristic of shock waves but for the

exact same reasons other finite amplitude waves increase entropy

density - primarily due to the conduction of heat.

The thickness of the shock layer is not independent of the magnitude

of pressure change across it; as a shock wave weakens with propogation

(and usually expansion) its thickness increases until it is no longer

a shock wave but rather an oridinary finite amplitude sound wave, with

no clear dividing line between the two, and no sudden change in the

significance of the entropy density increase.

Yes, all of which are very very constant on the time scales of sound

waves (see subject line).

But not constant between NYC and Denver (see original question and my

response re: air impedance, the significant change with altitude.)

On a slightly longer time scale, the atmospheric CO2 increase has

decreased the small-signal speed of sound at STP from 1126.91 Feet/Sec

in 1976 to 1126.89 Feet/Sec in 2003. Retune those instruments

.