See it listen
The whole idea is honesty about what changed. Below is real engine output on a controlled pair — one singer, one note, sung two ways.
01 — The diffTwo takes of one note
Same singer, same song, same sustained note — recorded twice, once in straight tone and once with vibrato. The brass line is the straight-tone reference. The grey line is the second take, aligned to it. Nothing here is drawn by hand: both contours are torchcrepe output, DTW-aligned by the engine that runs the Studio.
Real engine output · 0¢ is the reference's own pitch centre · 1.43s sustain, selected by rule · VocalSet, CC BY 4.0
02 — Train the earWhich one scoops into the note?
You improve fastest when you can spot the gesture yourself. Both takes reach the same high note. One slides up into it from below; one lands straight on it. Which one scoops?
Schematic contours — a perception drill, not measured audio.
03 — Progress, not a scoreThe same phrase, three weeks
There is no single number to chase. There is only whether the thing you worked on actually moved. Here is one leap, measured across three weeks of the same drill.
Leap arrives 40¢ flat, 180 ms late.
18¢ flat, 90 ms late — onset tightening.
Within 8¢, on time. Next frontier: vibrato onset.
Illustrative — a worked example of the loop, not a recorded learner's data.
Try it on your own voiceFree while it's young — three analyses a day, no account, no email. New here? Start with the curriculum →