Mind Your Ps and Qs

September 13, 2022

Make great impressions with synthesized speech

The English language is filled with homonyms: two words that sound alike and may even be spelled alike but have different meanings and pronunciations such as close, lead, live, read, row, sow, wind, and wound. Pronunciation is determined by the context in which the word is used.

This makes things confusing especially for speech synthesizers that can easily mispronounce these words. Speech synthesizers interpret spoken languages differently. It is recommended you become familiar with your voices before you allow fully dynamic situations to yield awkward results. You can detect and address these issues before you customers hear them.

Chant Developer Workbench arms you with key tools and workspaces to generate pronunciations, markup text to use the correct pronunciation, and validate solutions that yield the best speech clarity. The results can easily be deployed in applications with Talk&Listen, SpeechManger, and SpeechKit.

There are four articles in the Knowledge Base that discuss approaches for handling these issues: H072206, H072221, H072223, and H072225.