Share via


Note

Please see Azure Cognitive Services for Speech documentation for the latest supported speech solutions.

Lexicons and Phonetic Alphabets (Microsoft.Speech)

A pronunciation lexicon contains words and their pronunciations, and defines all or part of the vocabulary for a speech application. Speech engines usually include a default lexicon, which you can supplement with custom lexicons that are specific to one or more of your applications. You can use lexicons to optimize the performance of speech recognition and speech synthesis (text-to-speech).

The pronunciations in a lexicon are specified as phonetic spellings that use combinations of letters, numbers, characters, and symbols to represent the sounds of spoken language. There are multiple phonetic alphabets, each of which uses a different character set to describe the pronunciations of words. The phonetic spellings in a lexicon typically use only one phonetic alphabet.

Follow the links below for introductory and reference information about lexicons and phonetic alphabets.