Personalized. Friendly. Quality training.

This site hosts free, open-source tools that give a numerical and visual assessment of the prosodic characteristics of singing and speech.

New IntonTrainer 2.0 for Chinese

Records your voice, aligns your pitch contour to a native-speaker reference using Constrained Dynamic Time Warping (CDTW), and gives an instant similarity score (0–100%) with a Unified Melodic Profile — so you can hear, see, and track how closely your melody matches the target. An Android version is coming soon on Google Play.

Statistical analyzers

Intonation trainers

Users

The geography of users who have accessed this site since the beginning of 2019 is shown on the site map.

Authors

Downloads

Windows builds, user guides, and source code for each tool. The latest release is IntonTrainer 2.0 for Chinese — desktop now, Android coming soon on Google Play.

Analyzers

Trainers

How it works

  • The system compares the melodic (tonal) portraits of a reference phrase and a phrase spoken by the learner, with active learner–system interaction.
  • The main algorithms used for analyzing and comparing intonation features are described in the papers and the advanced user guide.
  • A set of reference sentences represents the basic intonation patterns of English speech and their main varieties.
  • A current linguistic idea is that a foreign accent is more evident and stable in intonation than in segmental sounds.
  • IntonTrainer 2.0 uses the same comparison of unified melodic portraits, with an updated interface, automatic recording, and instant similarity scoring. See the Chinese 2.0 download (desktop now, Android coming soon) and the project on GitHub.

Papers

In English

In Russian