Jason Kanter asked: >>> So let us suppose, *hypothetically*, that we had two recordings on CD. Suppose these recordings are of: the same professional pianist/artist, playing the same program of classical music of the last century or so, on the same *model* of piano (though a different piano in a different hall), recorded with the same equipment... but one of the recordings is tuned in ET by one tuner, and the other recording is tuned in a very bold WT by another tuner. How would we conduct a *truly scientific* test to examine the musical reality of this *hypothetical* set of CDs? <<< If, by "musical reality" you mean "which do listeners prefer", then a scientific test would go as follows: Solicit a pool of qualified volunteers to listen to the CDs. Do not tell the listeners what the differences are between the two CDs. (i.e. don't mention tuning.) Just label them #1 and #2. The labelling should be randomized so that the ET and WT tunings are equally likely to be labelled #1 as #2. Of course, the actual meaning of the labelling must be maintained on a list known only to the testers. Just ask the listeners which CD they like the best. We can improve the reliability of the test even more if we are allowed to change some of the conditions you set. If the chosen performer has a particular bias for or against WT vs. ET, then it is possible that this bias will be expressed in the performance as she listens to herself play. You could remove the performer as a factor by using one of the high-end automated player systems to do the performance for recording. In fact, I don't see why you need to even take a chance on using two different pianos in two different rooms. You might end up measuring the slight difference between individual pianos or individual room acoustics. Simply use one piano, tuning it as appropriate before each recording. -Bob Scott
This PTG archive page provided courtesy of Moy Piano Service, LLC