When we started building ShadowEdge, the easy path was a single number: say a sentence, get a score out of 100. It demos well. It’s also nearly useless for actually improving, and here’s why we went the harder route instead.
A score tells you that. Phonemes tell you what.
Imagine you score 78. Now what? You don’t know which words dragged you down, let alone which sounds inside those words. You’re left to guess and re-record blindly. That’s not practice — it’s a slot machine.
Phoneme-level feedback changes the question from “how did I do?” to “which specific sounds did I miss?” — and that’s a question you can act on. Miss the /θ/ in month three sessions in a row and the pattern is undeniable. You know exactly what to drill.
Measurable beats motivational
A vague sense of “I think I’m getting better” fades fast. Watching your accuracy on a specific sound climb from red to green over a week is a different kind of fuel — it’s evidence. Progress you can see is progress you keep chasing.
It’s what makes the loop work
Phoneme data is also what lets ShadowEdge hand off to WordEdge: the words you consistently miss aren’t a guess, they’re a measurement. That measurement is the seed of your next drill set. Without it, the “continuous loop” would just be a slogan.
So the four metrics — Accuracy, Fluency, Completeness, Prosody — and the per-phoneme detail underneath them aren’t decoration. They’re the whole point.