Find your R.
Hold the target. We’ll stop the recording.
- 1Connect
- 2Practice
- 3Listen back
Let’s hear your R.
Turn on the microphone, then start a try. We’ll take it from there.
Your target Exploration modeSet up auto-stop
Use a target chosen with your speech professional or a confirmed good “er”. The starting line is only a visual reference.
Drag to tune. Tap − or + for finer control. A lower number isn’t automatically better.
Acoustic estimates, not a pronunciation diagnosis. Whole-word accuracy still needs listening.
One try at a time.
Your listening judgment and the meter measure different things. Use both.
How the acoustic feedback works
This independent experiment estimates formants with linear predictive coding. It uses volume, voicing, agreement between two model orders, and stability checks to avoid rewarding unclear sound. It has not been clinically validated. It does not identify words or diagnose a speech condition.
F3 and the F3−F2 gap can help visualize R-like resonance, but appropriate targets vary across voices and sound contexts. A match to the adjustable line is not proof of correct pronunciation. Read the acoustic-measurement research · LPC method reference.
Inspired by the idea of visual practice in staRt. Independently built by Arjun Gerald Ramachandran; not affiliated with NYU, BITS Lab, or staRt.