[Harp-L] comb testing

[karen mass <kdmass7@xxxxxxxxx>]

okay, as an engineer, here's the only way i can see to prove/disprove the
theory that comb material affects the actual and perceptible sound of
different comb materials.

EQUIPMENT REQUIRED:
   sine wave generator
   audio amplifier
   studio monitor speaker
   spectrum analyzer
   microphone
   compressed air source
   regulator valve
   test fixture to connect output of regulator valve to harmonica
   set of reference reed plates and covers
   digital audio recorder
   anechoic chamber
NOTE:  all test equipment must be of high laboratory or recording studio
quality

UNITS UNDER TEST:
   harmonica combs made of various materials
      NOTE:  all combs must be made to the same machine drawing, to the
same tolerances, and machined using best commercial practices
      NOTE:  comb design must reflect reed plate and cover configuration to
provide for minimal leakage and identical mechanical fit of reed plates and
covers
      NOTE:  harmonica assembly for each comb type must be performed using
the same fasteners and other hardware
      NOTE:  all fasteners shall be tightened to specific torque numbers or
other measurable forces using suitable tools

TEST PROCEDURE:
in an anechoic chamber perform the following:

1.   CALIBRATION:
1.1        using sine wave generator, amplifier, monitor speaker,
microphone and recorder, and spectrum analyzer, measure spectral response
of the test set up across 0 Hz to 25 kHz.
1.2        record audio baseline response (reference spectra). record
output of spectrum analyzer and audio output.

2.  INSTRUMENTED TEST
2.3        remove sine wave generator, replace with test fixture connected
to output of regulator
2.4        place harmonicas assembled using reference reed plates and
covers and TEST COMBS in front of microphone
2.5        set regulator to output air at TBD pressure and rate
2.6.       allow tone to play for 1 second to settle
2.7        measure and record spectral and audio output of harmonica for
all notes
2.8        repeat 1.5 for all test articles
2.9        compare all test run output spectra to reference sine wave
spectra.  subtract reference spectra.  (this should eliminate the effects
of the test system from the results)

3.  SUBJECTIVE TEST
3.1        using the audio recorder, amplifier, and monitor speaker, play
recorded harmonica test samples for individual listeners
3.2        record listeners opinions of differences (not quality) between
recorded samples (including reference recording), without identifying the
materials used in each sample
3.3        compare listener results to instrumented test results and the
actual identity of test combs

using this procedure it should be possible to determine if there are any
actual spectral differences between combs and if there are any perceptible
differences between combs.
i think the test design should remove any possible differences that could
invalidate the test results.
the procedure may need a bit of tweaking, but it should be pretty close to
correct.

for what it's worth

iowaplayer