VOCAL RANGE

Telling you to move down by exactly a semitone would only work for people who already have training, which defeats the point. So the instruction is just "a little lower than before" — how far you actually moved is our problem to measure, not yours to control. Uneven steps are fine. Skipping three at once is fine.

You are also never asked whether you have hit your limit. Instead the measurement is allowed to break first: near the edge a voice does not stop cleanly, it wobbles, goes breathy and gets quiet — and that is exactly what the hold test is looking at. Two steps in a row that fail it ends the pass.

Down first, up second. Singing high notes tires the voice enough to spoil the low measurement afterwards; the reverse order does no such damage.

A note only enters your range if the pitch stayed inside a half-semitone band for long enough, loudly enough, and with a clean enough periodic waveform. Singing is not brushing past a note, it is catching one — a pitch you slide through is not part of your range.

This is measured twice with two different strictnesses, which is where the two ranges below come from. The strict one asks the note to be held twice as long and to drift half as much. Note that it penalises drift, not wobble: vibrato circles around a centre and drifts almost nothing, while a note at your limit sags downward. Punishing wobble would hand the smallest range to the best singers.

No attempt is made to exclude falsetto or any other way of producing sound. The only question asked is whether you can hold it, because plenty of songs are sung in falsetto. Vocal fry fails the test on its own — it is unstable and quiet, which is what the criteria already catch.

It does not recommend songs. Building that needs a melody-range database, and karaoke MIDI is copyrighted — the same wall we already hit once. Taste is the bigger reason though: two people with identical ranges want to sing completely different things, and range data cannot support a claim about what you enjoy.

It does not name a key either. A key says which note is home, not where the melody sits — an E minor song written low and an E minor song written high are different songs to sing. And the key button changes it anyway: press +2 and an E minor song becomes F sharp minor. Reporting a value that one button press invalidates would be self-defeating.

The whole tool rests on one piece of arithmetic instead. A key button shifts the entire melody by semitones, so the width of a melody never changes — that is the only hard constraint — while where it sits is free, and the machine can move it by six semitones in either direction. Six down is the floor, which is what makes the ceiling below a real limit rather than a suggestion.