This disclosure describes techniques for integrating the provision of real time pronunciation feedback into a user’s daily conversations. With user permission, real time speech is received and transcribed to text based on a semantic understanding of the conversation, the user’s history of mispronunciations, etc. A pronunciation score is calculated for different portions of the user’s speech, e.g., syllables, words, phrases, or sentences. Real time feedback (e.g., warnings) for mispronounced portions is provided using color, vibration, etc., in a manner that causes minimal distraction. The correct pronunciation can be indicated using audio, text, etc. The techniques can be implemented on any device, in particular, on devices with small or no screens.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.