Balsu talka ("Voice Gathering")

The "Balsu talka" ("Voice Gathering") project was created to allow technology to better understand the Latvian language. In order to create a diverse, open and accessible Latvian speech data set, we invite you to join forces to collect at least 1000 hours of spoken data.

Although speech technologies are developing rapidly, many devices still do not sufficiently understand the Latvian language. For example, it is not possible for us to provide voice commands for televisions, cars or telephones in Latvian. There are also few apps which you can use in Latvian with your voice instead of typing on the keyboard. Such shortcomings are exacerbated by the availability of open and freely available speech data in the Latvian language. This hinders the technological development of the Latvian language and makes it difficult to include the Latvian language in speech recognition research. The "Voice Gathering" project uses Mozilla's Common Voice platform to collect voice samples that are available to everyone, especially companies developing technological solutions and language researchers around the world.

Help implement this project and make technologies available in Latvian by going to https://balsutalka.lv/!

The Language Technology Initiative is financed by the European Union, as well as the National Development plan.