Achieving 21st Century Standards of Excellence in Tuning

Aural Tuning Tests for 2:1, 4:2 and 6:3 Ty e !ctaves

"ill "re##er $%T

&adison, 'isconsin &arch, 2(()

There have recently been many questions posed about octave sizes, the differences between them and how to tune them. The fact is, that you can construct an Equal Temperament within any size octave, a 2:1, 4:2 or :! or even a wider than a :! octave or a sli"htly narrow octave if that is your "oal. #t is simply the consensus of opinion today that a compromise between a 4:2 and :! octave is the optimum. The very sli"ht beat you will hear in such an octave will ma$e your %& a little purer and your %4 and '! a little faster than what is theoretically correct and what all the old boo$s on tunin" say is correct. (one of these boo$s that # $now of, say anythin" other than to ma$e a pure soundin" octave. #n reality, an octave can only be truly pure, meanin" havin" no audible beat, between any two matchin" sets of coincident partials )e*planation to follow+ but there is indeed a small ran"e within which an octave may sound beatless or pure to the ear. ,rom what we $now today however, tunin" an octave from -4 to -! that sounds pure to the ear, when startin" the tunin" process with no rapidly beatin" interval chec$s, would most probably create an octave anywhere from a 2:1 to a 4:2 type. This article will e*plain in detail in how to tune any size octave you choose and how to $now the difference. .efore "oin" any further, let/s define 0ust what these numbers, said to be octave types, are. They are ratios, yes, but they are not the equivalent of each other as they would seem to be theoretically. This has been a source of confusion to some because ratios are used in definin" other intervals. - &th is a !:2 and a '! is a &:4 for e*ample. These are fi"ures which also point to partials which coincide and indeed a &th can often have more than one set of audible coincident partials )the definition to follow+ but a '! and a &th are not usually discussed as types the way octaves are, only whether they are tempered or not and to what de"ree. Theoretically, an octave has a ratio of 2:1 and any multiple thereof, such as 4:2, :! and beyond are the equivalent, merely fi"ures unreduced to the lowest common denominator. The ratio of 2:1 indicates simply that the upper note of the octave theoretically has twice the frequency of vibrations per second as the lower. .ut when discussin" octave types, these fi"ures indicate what are $nown as coincident partials. The words, partial, harmonic and overtone all refer to the same phenomenon and are "enerally interchan"eable but piano technicians usually prefer to use the word partial )used as a noun+ when discussin" tunin". - partial is one of the hi"her and usually fainter )but not always+ sounds that we hear alon" with the fundamental which is the pitch we thin$ about and reco"nize when any "iven strin" of the piano is struc$. The fundamental tone or pitch is also considered to be the first partial. Theoretically, an upper or hi"her partial is an e*act multiple of the fundamental pitch, such as twice the frequency, two and a half times the frequency, three times the frequency, etc. 'ost people, even musicians do not perceive )hear+ partials. Even many piano technicians "o their entire career sayin" that they do not hear them. 1thers consider hearin" partials as essential to tunin". (ot to worry however, because when we hear beats in any interval, we are hearin" partials which are mismatched or not e*actly in tune with each other. #ndeed, the very essence of aural tunin" depends upon the ability of the tuner to perceive and control beats. -ny strin" of the piano has a series of partials which, if transcribed to loo$ li$e music, appear to be a lar"e, spread out, dominant seventh chord with e*tensions. There can be as many as 24 audible partials from the lower wound strin"s. 2sually however, we only concern ourselves with the first 8 partials which ma$e up what appears to be when written out on a music manuscript "rand staff, a lar"e dominant seventh chord which covers three octaves. 3hen you play any particular note of the piano, you of course do not hear what sounds li$e a chord )one usually only perceives the fundamental+ but try this tric$ to actually hear them: press slowly enou"h one of the lower notes of the piano, ta$e 42 for e*ample, so that it does not sound and hold it open )usin" the sostenuto pedal if the piano has one+ and play in staccato fashion the partial series up to the ei"hth partial, you will clearly hear each note of the chord come eerily out of the open 42 strin". 5old open 42 and then stri$e 4!, 6!, 44, E4, 64, -7sharp 4 and 4&. 42 is the first partial or fundamental and 4& is the eighth partial in this e*ample. 3hen you play any octave, you will naturally have sets of partials from each strin" which sound side by side or coincide, thus you have what are $nown as coincident partials. .ut to complicate matters "reatly and "ive the tuner the ultimate and nearly unsolvable problem and challen"e, all piano strin"s bear a phenomenon $nown as inharmonicity. .asically, this means that the upper partials are not e*act mathematical multiples of the fundamental tone, as they are thou"ht of theoretically. They are always somewhat sharper than a theoretical partial would be and the hi"her the partial, the more e*a""erated the sharpness. #nharmonicity occurs because a %a"e 1

piano/s strin" is made of steel which has a certain stiffness factor to it which distorts what otherwise would fit a theoretical mathematical model. #f there were no inharmonicity as there is none in an instrument such as a pipe or"an, all coincident partials would match each other when an octave is tuned so that it has no beat but with piano strin"s, it is only possible for one set of coincident partials to match at any "iven time when tunin". 3hen one set of coincident partials match, all others must be and are off or out of tune with each other, to some de"ree or another, dependin" on the inharmonicity and how hi"h the partials are. Thus, we have the types of octaves8 the most commonly used are the 2:1, 4:2 and :! types, althou"h there are other possibilities. #n the 2:1 type octave, the second partial of the lower note is e*actly in tune with the fundamental or first partial of the upper note. #n the 4:2, the fourth partial of the lower note is e*actly in tune with or matches the second partial of the upper note and li$ewise, with the :! type, the si*th partial of the lower note matches the third partial of the upper. #t is also possible that none of the partials match e*actly, all may be a sli"ht mismatch but the octave can still have a pleasin" or desirable sound, dependin" on the "oal of the tuner. To be"in the fine tunin" process, you must first tune your -4 to the pitch source. 9ou start by simply tunin" a beatless unison to the for$ or other source such as a tone from a metronome that has an -744: pitch or any other -744: tone that you consider reliable. Then, to verify that the match is e*act, you "o down to the note, ,2. (ow, this is a fine tunin" procedure, not what you would do for a pitch raise or lowerin", somethin" only to verify the ultimate precision and perfection such as when attemptin" the Tunin" E*am or for a public performance or professional recordin" where the specification on the contract is -744:. 9our piano needs to already be nearly perfectly tuned for any rapidly beatin" interval chec$ or test to be of any useful or practical value. %lay the interval, ,27-4, a '1;. 9ou should hear a rapid beat but probably not very rapid and the ,27-4 seventeenth must be wide of bein" a pure or beatless interval, not narrow. ,latten 0ust temporarily the ,2 so that it creates a very rapid beat, as fast as you can discern without it becomin" an indiscernible blur. -lternately play the ,27-4 '1; on the piano, then the ,27-4 interval from ,2 on the piano and the -4 from the pitch source. 3hen both intervals beat exactly the same, it confirms that the piano/s -4 and the pitch source are e*actly the same. #t is possible to have an aural accuracy which would measure to within one cent electronically usin" this test. -ccuracy within one cent would score a perfect 1:: on the Tunin" E*am and would satisfy any contract obli"ation. #f you hear a sli"ht discrepancy, all you have to do is ad0ust your -4 up or down by the sli"htest amount to ma$e the ,27-4 test match e*actly. #f your piano/s ,27-4 is sli"htly faster than the piano<pitch source ,27-4, your piano/s -4 is sli"htly sharp and conversely, if the piano/s ,27-4 is sli"htly slower than the piano<pitch source ,27 -4, the piano/s -4 is sli"htly flat.

Tuning the various Types of !ctaves

(ow, tune the -! to the -4 and to start, ma$e it as pure or beatless soundin" octave as possible. Try sharpenin" the -! until you hear a sli"ht beat, indicatin" the octave is sli"htly narrow, then flatten it bac$ sli"htly until it sounds perfectly pure. This will probably be a 2:1 octave. To test for the 2:1 octave, "o bac$ down to the ,2 you previously used to test the -4 pitch. -"ain, it must be near where it will eventually be in fine tunin", a wide interval with a rapid beat. .ut flatten it temporarily so that it is very rapid, as rapid as can be without bein" an indiscernible blur. %lay the ,27-! '1: and then play the ,27-4 '1;. #f both intervals beat exactly the same, the -!7-4 octave is a perfect 2:1 octave. #f there is a discrepancy, it is not a perfect 2:1 octave but a wide or narrow one from that point of view. (ow, from a perfect 2:1 -!7-4 octave, flatten a"ain but very sli"htly the -! but still retain what sounds li$e a pure octave with no perceptible beat. This will be a very small amount. (ow play the ,! and -! to"ether and listen for a rapid beat. The ,!7-! third must be wide of perfect, not narrow and if your piano is already nearly perfectly tuned, it will beat about ; beats per second. (ow, temporarily flatten the ,! as you did previously with the note ,2 to ma$e the ,!7-! !rd beat very rapidly, as fast as can be discernible without it becomin" a useless blur. (ow, alternately play the ,!7-! '! and then the ,!7-4 '1:. 3hen both the '! and the '1: beat exactly the same, the -!7-4 octave is a perfect 4:2 type. #f the '1: is sli"htly faster than the '!, the octave is a little wider than a perfect 4:2 which will in fact, be the ultimate "oal but more on that later. To learn these differences, try to %a"e 2

"et these Equal Beating tests absolutely perfect so that you will $now and understand the difference, one from the other. (ow, from a perfect 4:2 octave, flatten sli"htly a"ain the -! until there is a very slow beat between the -! and -4, 0ust a little more than you mi"ht feel comfortable hearin". (ow play the -!744 m!. #f your piano is already nearly perfectly tuned, the -!744 m! will beat very rapidly at nearly the limit of what may be discernable. - m! is normally about 1 cents narrow in an optimally tuned Equal Temperament and at that spot on the $eyboard8 the beat is close to bein" an indiscernible blur. Therefore, you may temporarily sharpen the 44 sli"htly, to ma$e the -!744 interval easier to hear. (ow, alternately play the -!744 m! and the 447-4 ' . 3hen both the -!744 m! and the 447-4 ' beat exactly the same, the -!7-4 octave is a perfect :! octave. (ow, to achieve the optimum compromise between a 4:2 and :! octave, sharpen the -! very sli"htly so that neither chec$ for a :! or a 4:2 octave tests perfectly. #n other words, the test for a 4:2 octave should reveal a sli"htly faster ,!7-4 '1: than ,!7-! '! and the test for a :! octave should reveal a sli"htly slower 447-4 ' than -!744 m!. 3hen you have found the spot for -! which reveals this sli"ht discrepancy between the 4:2 and :! tests, you will hear that the -!7-4 octave has a very slow beat to it, about one beat in every two seconds. This is now considered the optimum width for the initial -!7-4 octave in Equal Temperament and what is used by most 4TE=s to set up the 'aster Tunin" for the Tunin" E*am. #t is also the equivalent to within a very small and ne"li"ible de"ree of a 4:2 octave plus one cent. This is the width of octave that >r. -l ?anderson used to obtain the amount of stretch needed for optimal piano tunin" when he created the calculation for his Electronic Tunin" >evice, the ?anderson -ccu7Tuner. #n this $ind of compromise, note that none of the coincident partials are in tune or match e*actly with each other. This may also be an e*ample of the whole octave sound of which @ir"il ?mith often spea$s. This, almost but not quite perfect tunin", is an e*ample of the $ind of compromise which is necessary to defeat the problem and challen"e of inharmonicity, the way to achieve the finest tunin" possible from the modern piano, throu"hout its entire ran"e. -lso a note of importance, :! type octaves are considered optimum for .ass tunin". The m! with lower note and ' with upper note test is useful down to about the note ,1. 3hen the octave has a reasonably "ood sound with 0ust a sli"ht and very slow beat when a sin"le octave is played and both m! and ' interval chec$s beat exactly the same, the octaves in the .ass are optimally tuned. To tune the seventh octave in the hi"h treble in perfect 2:1 octaves with the si*th octave below it, as is required for the Tunin" E*am, use the test for the 2:1 octave as described, a '1: below the bottom note and a '1; below the top note. #n most other situations however, the seventh octave tuned so that it matches other partials from further down the $eyboard is considered ideal, dependin" upon the perception and taste of the pianist whom we serve. - future article is planned which will discuss and e*plore all of these options.

#n summation, the tests for the 2:1, 4:2 and :! octaves are as follows:

2:1 use the test note a '1: below the bottom note and a '1; below the top note. 4:2 use the '! below the bottom note and the '1: below the top note. :! use the m! above the bottom note and a ' below the top note. The Equal .eatin" techniques are attributed to 1wen Aor"ensen B%T. The understandin" of coincident partials is from Aim 4oleman, ?r. B%T and his Coleman Beat Locater charts from ,red Tremper B%T.

%a"e !