Steady Pitch / for choir directors

How to run vocal range checks on a choir, and what to do with the numbers

Twenty singers, one rehearsal, and a card you tape inside every score for the rest of the year. The measuring is the easy part. Reading the result is where directors get stuck.

Why you are guessing at keys

Ask a director why a piece sat badly and you usually get an answer about the arrangement. Ask them for the lowest comfortable note of their fourth alto and you get a shrug, which is fair, because nobody carries twenty numbers in their head.

That shrug is the whole problem. Choosing a key against a mental picture of a choir is guesswork dressed up as judgement, and it fails in the same direction every time. You pick something that sits well for the voices you hear most, which are the confident ones, and the piece quietly excludes the third of the room that cannot reach the top or cannot sustain the bottom. Those singers do not tell you. They mouth it, or they drop out on the hard bar, and what you hear is a thin section rather than a range problem.

One rehearsal of measurement removes the guess for good. Ranges move, but they move slowly enough in adults that once a year is plenty. In adolescents, twice a year, and a re-check on anyone who suddenly sounds different.

What to measure, and the mistake almost everyone makes

Measure the lowest note the singer can HOLD comfortably, and the highest note they can HOLD comfortably. Two numbers per singer.

Comfortably is carrying the entire method. The instinct is to take the extreme, the lowest croak and the highest squeak, because it feels like more information. It is worse than useless, for two reasons.

Nobody sings their extremes in tune. The last two or three semitones at either end go weak and unstable before they disappear, so the extreme note is one the singer cannot repeat, cannot tune, and certainly cannot hold inside a chord at bar forty of a concert. A key chosen from extremes falls apart the first time somebody is nervous.

And the extremes are the least stable thing about a voice. They move with sleep, hydration, time of day and confidence. The comfortable band barely moves. Build your year on the figure that changes and you will keep rediscovering that your numbers were wrong.

So the instruction is not sing as low as you can. It is go down until the tone stops being yours, then come back up one note and hold that. Same at the top.

Running it in one rehearsal

Twenty singers takes about forty minutes if you do it yourself and about fifteen if you do not.

Put them in pairs. One sings, the other writes. Same instruction and same order for every pair, low first then high, because doing the top first tires the voice and drags the bottom figure up.

Ask for a held note, roughly a second, on an open vowel. A note that flashes past is not a note anybody can use in a score, and it also defeats every measuring method there is, including your own ear.

Two passes at each end, and keep the more conservative of the two. A first attempt at the bottom of a range is almost always short, because the singer has not settled into the register yet.

Reading twenty pairs of numbers

Write the lows in one column and the highs in another, then sort each column. You are looking for three things and none of them need any theory.

The common comfortable band. Take the highest of the lows and the lowest of the highs. That interval is what your whole choir can sing in unison with nobody excluded. Do not expect it to be generous. On a mixed group with adolescent boys mid change it is routinely narrower than anybody in the room would have guessed, and whatever it turns out to be, that number is not a failure of your choir. It is the constraint you have been fighting blind. I am deliberately not printing a figure to expect here, because the whole point of the exercise is that yours is the only one that matters.

Write it on a card and tape it inside your score folder. Every question about a key becomes arithmetic instead of anxiety, and you stop finding out in rehearsal that a tune sits a third too high.

The outliers at each end. Usually one or two singers whose low sits far below the group and one or two whose high sits far above. Those are not problems, they are assets you now know you have. That is where your descants and your bottom octave doublings come from.

The narrow ones. A singer with a range of a fifth is telling you something, and it is rarely talent. Often a changing voice, sometimes a technical knot in one register, and occasionally somebody who has been singing a part that does not fit them for two years and has stopped using the rest of their voice.

Changing voices, and why the ordinary parts break them

If you have adolescent boys, the range checks will show you something that looks alarming and is not. One boy has four or five usable notes and very little either side of them.

The naming matters here, because getting it wrong sends you to the wrong shelf. In the Cooper and Collins classification the cambiata is the FIRST stage of the change. That voice has not dropped an octave yet, it sits higher than people expect, the tone goes woolly and the range narrows. The boy who is stuck low, with a middle that will not phonate and cracking between registers, is in the SECOND stage, which the same classification calls the baritone. Both are normal, both are temporary, and both want music written for the range they have this term rather than the one they will have next year.

What does not work is the obvious thing. Give the low boy the bass part and he cannot reach it, so he pushes, and pushing is the one thing that stops low folds vibrating at all. Give him the alto part and he sings it an octave down into a register he has not got yet, or in a falsetto he cannot control. Give either stage five good notes written for five notes and they sing them well, which does far more for the voice, and for whether they come back next term, than a wider part they can only survive.

There is repertoire written for this rather than adapted to it. Don Collins built a whole method around it, and the Cambiata Collection is on JW Pepper.

Re-check every couple of months. These stages move, and they move fast.

The singer who cannot match pitch, and the thing to rule out first

Every choir has one, and in a good half of cases the diagnosis is wrong.

Before treating it as a pitch problem, check whether they are matching perfectly an octave away and being told they are wrong. A man, or a boy whose voice has just dropped, handed a note in the middle of the piano will often land the same note an octave below. That is a correct match, musically, and if nobody says so he concludes he is tone deaf and stops trying.

Test it two ways. Have them match another singer in their own octave rather than the piano. And drop into their octave yourself and see whether they land.

There is a real reason a voice works better than a piano here. A piano note starts with a percussive attack, its partials sit slightly sharp of exact multiples because real strings are stiff, and the whole thing decays from the moment it speaks. A sung note is steady, harmonic, and it carries formants, which is to say it carries the shape of a vocal tract making it. Sing the target and you hand the singer something to imitate rather than something to measure, and imitation is a much older skill than pitch matching.

If they match a voice and not a piano, you do not have a pitch problem. You have a referencing problem, and it goes away with practice on unisons.

Doing the measuring without a piano

You can run the whole thing at a keyboard, and directors have for a century. The two costs are that it takes your attention, so you cannot supervise the room, and that you have to trust your own ear at the bottom of a range where everybody's ear is worst.

I built a free browser tool for this because I wanted the numbers without the keyboard. It is mine, so weigh the recommendation accordingly. The singer sings down to their lowest held note and up to their highest, and the page names both ends, the width in octaves and the nearest voice type. It runs in a browser tab on a phone, nothing installs, there is no account, and nothing anybody sings leaves the device, which matters when your singers are minors and the alternative is sending their voices to somebody else's server.

It also draws the raw pitch curve it heard underneath the result, so when it misreads somebody you can see that it misread them, instead of writing down a number that was never real.

Any method that gets you twenty honest numbers is fine. The numbers are the point, not the instrument.

What to write down and keep

Name, date, lowest comfortable, highest comfortable, and one word about the quality at each end. Weak, breathy, solid, pushed. That word is worth as much as the number by the second measurement, because a low D that was breathy in September and solid in February is the clearest evidence of progress you will ever hand a parent.

Keep it in something you can sort. The value of range data is almost entirely in being able to sort it, and a pile of paper slips cannot be sorted in the ninety seconds you have before a rehearsal starts.

Related, and it comes up in every staffroom, why your tuner says you are singing the wrong note. If you hand students a tuner app, this is the reading that makes them think they cannot sing.