User Tools

Site Tools



  1. A spoken text (i.e. words/phrases/sentences/paragraphs) is transcribed using a natural language processing API.
  2. Using the Wordnik API on each word from the acquired text groups of synonyms/hypernyms are found to replace their originals.
  3. The resulting text & found word groups are synthesized word by word using text-to-speech (TTS).
  4. The sonics of all words are analyzed using the Echo Nest API.
  5. Variations of the synthesized text w/ inserted word substitutions are analyzed based on various properties (rhythm, pitch etc).
  6. The resulting data is used to create a text w/ the same syntax as the initial input but with words substituted by synonyms/hypernyms based on desired rhythm, sound (harsh, mellow, screechy etc.), length, complexity (choice of words, amount of syllables etc.) or else.

The goal is to create a real time application that learns (collects metadata) while being used & can be applied in an installation environment to emphasize on rhythmical speech & miscommunication in language & technology.

Google Speech API

Chrome TTS API

Wordnik API

Echo Nest API

Contact Dennis at for more information

speech_synthesis-based_rhythm_analysis.txt · Last modified: 2013/09/25 15:17 by dheihoff