"Sander deWaal" wrote in message

"Arny Krueger" said:


How can I make useful sound files available to anyone of
my class A hybrid amps, in order to make a meaningful
comparison with another amp via PC-ABX?


Both amps should be driven by the same recording, the
same source, at relevant levels. A good source would be
a number of high quality digital recordings strung
together, with some test tones for level setting at the
front.

Get a really good computer audio interface, load the amp
with a relevant load, attenuate a portion of the amps
output to match the input of the interface, and record
the attenuated output of the amp with the audio
interface.

A pair of high quality 5K 2-watt carbon composition
potentiometers make a good power amp output attenuator.


OK, so for this to work, both Mike and I have to use the
exact same source, PC setup and dummy load (or speaker)?

Only if you want to compare an amp you have to one that Mike
has.

I am currently recommending the M-Audio Audiophile 24192
as a reasonably-priced but *overkill-enough* audio
interface. It works on PCs and I think Macs as well.

There's some work for me to be done in this regard, as it
stands I only have an on-board sound card on my mobo
(Asus whatsoever).

The PC that sits in the living room, will have to do the
work since the one in my shack and the notebook are only
connected via a wireless LAN network.
Both have mediocre sound cards, if not worse, so they
would be useless for this to work.

One you have made your digital recordings of the outputs
of the amps, use the test tones to match levels. Ensure
that the recordings have identical start and stop points
within a few milliseconds.

I use Adobe Audition for this, but the freebie Audacity
software looks like it has all the necessary tools as
well. I think Audacity runs on both PCs and Macs as well.

Hold on: you're saying that we should match levels
*after* the recordings are made?

Yes. Only rough matching is required while the recordings
are made. You do the matching with the audio editing
software mentioned below.

Most PCABX Compartors accept .wav files of any sample
type and rate that the local audio interface will handle.

It can also be useful to compare the input file to any
output files for the equivalent of a "Straight Wire
Bypass Test".

That I've done before, and that's why I know my amp will
sound different from most others :-)

You can re-record the output recordings again and again
to simulate the effects of passing the test music
through a string of amplifiers. These files also are
good for training purposes. Even the best power amps can
usually be heard goofing-up the sound after 10 passes,
for example.

But there's no point in doing so, IMHO.

Sure there is - for listener training.

That wouldn't resemble a real-wprld scenario, no one is
using 5 or 10 amps in a series string (unless one's name
is Peter Walker grin )

It's also a useful technique for listener training.

Recently, Mike McKelvy and I were discussing the various
topologies of our respective amps, and we thought it
would be interesting to compare my amps with his
Acoustat. If this could be done via PCABX, it would
save us a great deal of shipping costs for a comparison
IRL.

Thet's the big advantage of PCABX - you don't need to
ship equipment, just sound files. You can even do stereo
comparisons of amps for which only one channel is
available.

But it still seems to be a problem to compare two
amplifiers that are located in resp. SoCal and the
Netherlands, unless one is using the exact same source,
material, PC and dummy load.

Yes, some things have to be held constant. The source can be
be an inexpensive but good optical player, the material can
be on a CD, and the dummy loads can be built from the same
parts list and schematic. Still easier than intercontinental
amplifier shipping, no?

Synchronizing the files afterwards isn't a problem I
think.

However, I still don't understand how one can do a
meaningful comparison when one's to adjust levels *after*
the recordings are made.

Perhaps this will make more sense after you have some
practical experience.

That way, differences are sure to occur, especially when
loudspeakers are used as a load instead of resistance
dummy loads.

Not at all. Since you are comparing files, and the files
have been level-matched the comparison is level-matched even
if different power levels are used during data gathering.
Usually power amps don't change their sound quality with
only small changes in output level.

Or am I missing something here?

Does your power amp sound different at 30 watt output than
it does at 30.5 watts?