Text-To-Speech (speech synthesis) is a solution to convert text transcripts automatically into speeches. With CALLR, TTS can be used to produce media files that are played at the moment you chose. The read text can contain variables (such as $name) corresponding to the entry of your directories to produce personalized vocal prompts. You can call each customer by his name or translate into speech any other customer-level entry of your database you might need (order ID, order price, the amount due...).
List of TTS voices and how to access it
A catalog of TTS voices is put at your disposal to offer the service in a range of language (English, French, German and Dutch) with masculine and feminine voices covering different accent.
For instance, Dutch is offered in a feminine voice with a Belgium accent (NL-BE_Ellen) or a feminine and masculine voice with a Dutch accent (NL-NL_Claire and NL-NL_Xender).
Voices are sorted using two codes: the first one is the language code while the second is the country code.
A full list of the available TTS voices can be fetched with the API using the media/tts.get_voice_list method (doc). It requires no parameter.
Snapshot of TTS voices (03/06) fetched using Python 3.5
List of available TTS voices per country:
- Belgium (NL): Ellen
- Germany (DE): Yannick, Anna
- France (FR): Thomas, Sebastien, Audrey, Viriginie
- United Kingdom (EN-GB): Daniel, Serena
- Netherlands (NL): Xander, Claire.
Set the pace of speech with TTS (API)
While setting the content of a media using the media/tts.set_content method (doc) the options object taken as a parameter, Media.TTS.Options object (doc) allow to set up the rate with an integer ranging from 0 to 100, 50 being the default value.