Normally stretched treble sounds flat

[Fred Sturm]

John Baird wrote:
> 
> Alternate Subject Line:   Outrageously stretched treble sounds great
> 
> RE: The phenomenon where a normally, RPT-exam-passing, stretched treble
> sounds flat, especially when playing a slow arpeggio up to the top
> octave, but an extremely stretched treble sounds very good. The single
> octaves, 10ths & 17ths don't pass an inspection by tuners, but the piano
> does sound great when it is played.
> 
> I believe this has been discussed before--does anyone remember when or
> what the subject line was?
> 
	There was an article in the PTJ some years ago entitled (I think)
"Picasso Tuning" and the thread may have had that title. I remember a
discussion along these lines on Pianotech about four years ago. Jim
Coleman contributed quite a bit, including his "pure 5ths" temperament,
which he later turned into PTJ articles.
	I personally tune a triple octave stretch on concert grands and any
other piano whose inharmonicity allows (a judgment call, and the variety
is quite broad). I find this stretch, which produces rather fast beating
"single" (2:1) octaves, and 3rd/10th, 10th/17th and 3rd/17th tests with
pretty wide differences in beat rates in some parts of the scale (again,
varies tremendously depending on inharmonicity curve) creates a more
brilliant and generally pleasing sound overall. Pure double octave
stretch, which is very reliably produced by SAT FAC tunings, is quite
nice and clean, but not as "interesting" and "vibrant." But frankly, I
don't think triple octave stretching  is very "radical" in sound. None
of the octaves or double octaves sound bad unless you focus the ear on
listening for beats. I suspect wider than triple octave could be quite
acceptable in many circumstances.
	Trouble is, it seems most discussion focuses on anecdotal evidence,
with no quantification to back it up. It would be interesting to do a
series of tuning comparisons somewhat like what Jim Coleman and Virgil
Smith did a couple years back, varying tunings by perceptible and
measurable stretch constants consistently applied.
Fred Sturm
University of New Mexico