Paul wrote:

The book's standards of good performance are superlative, i.e. the
"good"
designs here are probably comparable to the best designs commercially
available, in terms of raw audio quality.

The author comes from the Scientific School of Audio System
Performance
Analysis (SSoASPA). He believes that if two amps with similar specs
sound different, it doesn't indicate the presence of subjective,
unmeasurable attributes --- it merely means that we are not
performing the right tests for the right parameters

..

Agreed. Specification sheets often do a spectacularly poor job of
illuminating performance advantages and differences.

Chapter 2, "Beginning at the beginning", focuses on balanced to
unbalanced signal connections, and then discusses stepped
attenuators. Both these are among the latest "purist" fads, with
questionable
benefits in most cases. enormously to the value of the book.

Both of these issues have positive aspects. Balanced I/O is widely
used
with audible benefits in so-called professional audio equipment and is
used for audio production. Stecpped attenuators provide precise,
repeatable gain settings and exceptional channel balance.


Without a stepped precision attenuator you are playing with yourself
in terms of repeatability.

Could I have asked for anything more from a book which wants to cover
all aspects of the audio home-building scene?

Yeah, you could, in fact. A lot more.


4. Some circuits for testing audio equipment, e.g. a sine wave
generator, a high-Q notch filter for harmonic distortion analysis, a
capacitor meter, etc., would have been useful.

The author probably knows that all of those pieces of legacy test
equipment have been replaced by computer audio interfaces and
analytical software.

A PC is -not- a piece of test equipment. National Instruments
http://www.ni.com/pxi/ has an in between, but it's expensive. PCs,
commodity PCs, are noisy and unbenchworthy.

There will be HP200CD oscillators running long after the last
Microsoft desktop OS is booted for the last time. The Tek scopes will
be dead from CRT failure but someone will offer a real oscilloscope
with a real electrostatically deflectred jug. The old chopper
stabilized audio voltmeters will still run too.