Text-to-speech allows a computer system to speak any text using either a synthetic or natural voice.
The text can be taken from documents, a website, text messages or integrated software system (where it’s not possible to show text on the screen).
Free and open Welsh language text-to-speech resources are available from the Language Technologies Portal, along with other resources that help developers use voices in various deployment environments. Resources are available as downloads or via an online API. The resources are based on two open source text-to-speech systems, namely
MaryTTS based resources
The portal’s MaryTTS resources enable you to create your own Welsh language synthetic voices. The new voices are naturally sounding and are very easy to create. Here are two examples from recordings made with two members of the Language Technologies Unit:
All the necessary documentation and scripts to get you started with creating your own Welsh language text to speech can be found in the following GitHub repository:
Festival based resources
A ready built Welsh language voice for Festival can be downloaded or used online via an API. Here is a demo of the voice:
The voice itself can be downloaded from the following GitHub repository:
You can download the voice as a zip file with the following direct link:
Festival TTS API Service
If it’s not possible for you to embed Festival and the downloaded voice into your software project, then the voice is available via an online API.
The voice can be used via an online API service that we host so as to voice web based text and/or use in devices connected to the internet but with limited hardware.
Further Festival Resources from GitHub
Our Welsh language text to speech voices rely on the Festival Speech Synthesis System. Both the data and code that we have developed bring you Welsh language text to speech in Festival are open source and available for you to download, adapt and run locally: