ASR Language Identifier

A language identifier is part of the grammar header that specifies the language to be used when performing decodes.

The format of the language identifier follows the convention set out by RFC 3066. The identifier is a language and country pair, like en-US for United States English. The DNN engine is accent-agnostic, so there is no need to specify en-US or en-GB for English — you can simply specify en. If you need dialect-specific spelling, then specify en-US or en-GB for regional spelling.

The speech engine currently supports the following languages:

Language CodeDescription
en-USAmerican English acoustic model and dictionary
en-AUAustralian English acoustic model and dictionary
en-GBU.K. English acoustic model and dictionary
en-INIndian English acoustic model and dictionary
es-MXMexican Spanish acoustic model and dictionary
es-COSouth American Spanish acoustic model and dictionary
fr-CAFrench Canadian acoustic model and dictionary
pt-BRBrazilian Portuguese acoustic model and dictionary
de-DEGerman acoustic model and dictionary
it-ITItalian acoustic model and dictionary

To specify the language in a grammar, use the following syntax:

ABNF

language en-US;

GrXML

<grammar language="en-US" ... >

Was this article helpful?