<HTML> <HEAD> <TITLE>Re: [CAUT] Sacrifice (was tuners- technology)</TITLE> </HEAD> <BODY> On 3/2/05 8:31 AM, "Wimblees@aol.com" <Wi= mblees@aol.com> wrote: <BLOCKQUOTE>Doing unison tuning by= ear is described best by Virgil Smith. He says that the human computer can = analyze the beats much better than the ETD. This is evident when tuning octa= ves. As he put it. "...octaves were right on when only one string of th= e upper octave was sounding, but was flat when all three strings were soundi= ng". </BLOCKQUOTE>     Yes, this is what Virgil says. That a unison goes f= lat overall when it is tuned (ie, if one string is tuned “perfectly,&#= 8221; when the other two are tuned to it, the overall pitch of the unison wi= ll be flat of where the originally tuned string was). And about nine years a= go Jim Coleman confirmed this claim through measurements he made with a new = RCT (RCT had just come out). I questioned Jim when he asserted he had made s= uch measurements, as I had been unable to measure this phenomenon with my ne= wly acquired SAT. Jim shared his data (I think this discussion was on Pianot= ech), explaining in some detail his methodology. Bottom line: through carefu= l measurement he had confirmed that a tuned unison was between 0.1 and 0.2 c= ents flat of the average pitches of the three individual strings. In his exp= eriment he had tuned each of the strings to within a measured tolerance of 0= .1 cents of one another.     I more or less accepted that at the time, thinking = that RCT was perhaps more precise than SAT, hence you could read such a smal= l effect using SAT, but I had to say that a difference of such a small magni= tude was not going to be significant in the actual tuning of a piano. My own= take being that Virgil was using this purported “fact” to justi= fy stretching octaves. Remember he was asserting at the time (he has since r= ecanted) that he tuned by listening to the beat between the fundamentals of = the two notes of an octave, and made that utterly beatless. He also said (an= d wrote) that this produced pure and beatless triple and quadruple octaves.<= BR>     Your mentioning this unison/octave “apparent = phenomenon” led me to be curious. Having an RCT of my own now, I decid= ed to try to duplicate Jim Coleman’s experiment. First step is to tune= each string of a unison within a measured tolerance of 0.1 cents – no= t a real easy task. Among other things, it is difficult to get repeated read= ings for a single string that are within 0.1 cents of one another. It requir= es playing at an utterly consistent level of volume. To have any credibility= at all, one needs to be able to do at least three sample readings of each s= tring, and have all of them be within 0.1 cents of one another.     But, yes, I was able to do this, and proceeded to r= ead the unison, with the same care and the same number of samples. And then = I went back and repeated every step (re-measuring each string individually, = etc). My results: I did not confirm Jim’s data. I found what I conside= r to be completely random results. Sometimes the three strings played togeth= er would be flat, sometimes sharp, sometimes the same. I am by no means sayi= ng that Jim did anything but a very careful and credible job, as I know him = to be a very careful and utterly honest person. But my results were, shall w= e say, varied to such a degree as to lead me to believe that it would need a= great number of repetitions of the experiment to persuade me that there was= any measurable difference between the pitch of three strings sounding toget= her and the pitches of the individual strings.     I realize that Virgil has taught this in classes, a= nd that he has demonstrated, and that people have been persuaded by listenin= g to his demonstrations. I suggest that it is quite possible that, in many i= nstances, they heard what they thought they did. First, it is next to imposs= ible to tune a unison within a tolerance of 0.1 cents, and I would say that = it is utterly impossible without the use of a machine. It’s a problem = of resolution – 0.1 cents is at the threshold of where a pitch produce= d by a piano string can be measured. They just don’t produce pitch tha= t clearly defined. Variance in volume, and not that large a variance, will c= hange pitch more than that.     So my explanation of “how it works” in = Virgil’s demonstrations is that, in fact, the unison tuned is not &#82= 20;absolutely perfect.” That one of the strings is likely to be, say 0= .3 to 0.5 cents flat of the originally tuned string. And that the aural reso= lution of the pitch of three strings of slightly different pitches will be a= ffected by the factor of phasing (phenomenon where strings will tend to phas= e with one another, locking their pitches to one another just like PitchLock= does), so that it is quite possible that the perceived (and measured) pitch= of the entire unison would be lower than the original string, because of on= e string having a lower pitch. And the unison might sound very clean. A unis= on within a tolerance of 0.5 cents generally sounds “perfect” to= most everybody. But I know most if not all of us can hear a difference of 0= .5 cents in context of octaves, M3s and many other intervals.     At any rate, I would take Virgil’s assertion = with a very large pinch of salt. Maybe there’s some truth in there som= ewhere, but it isn’t what I would give the status of a fact. Regards, Fred Sturm University of New Mexico PS I would be interested in hearing the results of anyone else who replicat= es the described experiment.      </BODY> </HTML>