Official Code
https://github.com/DigitalPhonetics/IMS-Toucan
Interactive Demo
Audio Samples
For these samples, we adjust the language embedding such that each speaker speaks with an accent that matches them. Speaking without an accent is also possible by choosing a language embedding that matches the text rather than the speaker.
-
English Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
German Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Dutch Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Finnish Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
French Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Greek Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Hungarian Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Italian Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Polish Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Portuguese Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Russian Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish -
Spanish Speaker
English German Dutch Finnish French Greek Hungarian Italian Polish Portuguese Russian Spanish