"bob" wrote in message
...
On Apr 2, 9:28 am, "Arny Krueger" wrote:
"Harry Lavo" wrote in message
One of the principles of testing in any field of human endeavor is
to try to emulate as much as possible the conventional
context of the variable under test.
Exactly. And that is exactly the path we followed while developing ABX.
That
you would mention that concept and the OOhashi test and all of the
technical
gyrations that it imposes on the normal listening experience in the same
post, is a true wonder!
Yeah, the standard seems to be that an negative ABX test conducted in
someone's living room is too unfamiliar to be reliable. But a positive
listening test conducted in an MRI tube, well, that's the gold
standard!
I can't let this piece of fabrication stand. Either you are deliberately
distorting the facts for effect, or you learned or remembered little from
the test.
The listening test was not done anywere near an MRI. A special room was set
up with an easy chair, soft lighting, plants, and a window wall looking out
at a peaceful scene. No equipment was in evidence. Music was played...in
ABBA order (order known to subject, but not assignment of variables.....also
not known to scientists conducting test....thus double blind). After the
music ended, the subject was asked to RATE THE MUSIC provided by A and then
B on a scale, using normal musical attributes. After a brief rest of about
10 minutes (a palate-cleansing, if you will), the test was repeated. Order
bias was controlled blindly and randomly, as mentioned. The test was
repeated several time among each of seventeen subjects.....who were music
lovers whatever their professions. The variables: a CD recording of
Gamelon music lasting 3 mins 20 secs, in one case played with the standard
CD cutoff frequency of 22khz, and in the other played with a separate
"super-tweeter" added (but electronically separated and amplified) that
reproduced the ultrasonic frequencies of the Gamelon above 26khz.
Statistical analysis showed that the sound quality was rated higher when the
ultrasonic frequencies were present. During this whole proceeding the
subjects did not know what was being tested.
The test you are disparaging did not involve an MRI tube and nothing to do
with listening evaluation. It was a separate test that involved EEG and PET
scans of the same 17 subjects in the same room. It was a completely
different test, and all the respondents were asked to do was stay
awake...the data being sought was generated by their bodies and recorded by
the lab equipment. The were then exposed to four test conditions at random,
and then in reverse order. These were a different set of Gamelon music of
approximately the same 3min length, once with and once without ultrasonic
sound present, another a 3min period of silence, and yet another three
minutes with only the ultrahigh frequencies playing (silent to the
subjects). The test monitoring recorded the test subjects physiological
reactions to the four stimuli, which were tested in random order and then
reverse order, and after a short break, repeated several more times with
order randomized. The result of this test was that the music selection
activated the portions of the brain active in listening to music, with the
ultrasonic-added variable eliciting more response, especially in the
pleasure-centers of the brain. Whereas the ultrasonic-only portion of the
recording, divorced from the audible sound, meant nothing and elicited no
response as did the silence variable. Again, these were statistically
significant results.
So....the results of the study. The addition of ultrasonics led to greater
listening pleasure as experienced and recorded by the respondents, and as
mesured by the lab equipment independent of the respondents control. And
notice this was without any direct comparison or choice to disrupt
concentration on listening during the test itself, in either of the two
tests.
The scientists noted two important things, in their opinion. One was the
creation of a relaxed listening environment that duplicated to some degree
the ideal home listening environment. The other was the use of musical
excerpts that lasted a bit longer than three minutes and more time between
musical excerpts than normally used. They explicitly stated that they felt
the 20 second snippets of music used in the testing done for Sony twenty
years earlier was a possible major flaw in the work that established the
22khz CD cutoff (in which case it is also a flaw in most ABX testing, as is
the quick switching). This latter conclusion was based on preliminary work
with the EGG system wherein they determined that there was a substantial
"ramp up" and "ramp down" in brain activity after the start and stop of
musical selections, suggesting that short musical excerpts and
quick-switching both had the potential to distort the musical experience.
I'm going to this length because it is obvious that Arny and Bob are once
again trying to disparage the test, as they did when I presented the results
several years ago. For newcomers to the thread I think it is important to
set the record straight. For anybody who wants to read the entire article,
it is he
http://jn.physiology.org/content/83/6/3548.full