Almawave at the 22nd Conference of the Italian Association of Voice Sciences
13 January 2026
From November 22 to 24, at the University of Milan, the 22nd Conference of the Italian Association of Voice Sciences will take place, dedicated to the theme of experimental field phonetics for linguistic documentation.
As part of the satellite event, scheduled for January 21 and titled “Recent developments in speech technologies. The Italian perspective in research and industry”, Leonardo Badino, Head of Voice Engineering, and Enrico Zovato, Lead Engineer of Almawave’s Voice Engineering team, will present Velvet Speech 2B.
VelvetSpeech-2B is the first multimodal model of the Velvet family, designed to integrate text and speech within a single system and enable natural voice interaction with a Large Language Model (LLM).
Our experts will illustrate how Velvet Speech 2B can recognize and translate speech in real time, understand and respond in mixed Italian–English conversations, and integrate speech emotion recognition features, i.e. the classification of emotions in the voice.
