Work in the open.
Projects, observations, and the questions behind them. A growing record of what we’re learning about audio.
Current project001
RivetText-to-speech
In developmentExpressive speech.
Expressive speech.
A smaller footprint.
Our core text-to-speech project. A study in expressive speech, compact models, and practical inference.
Read the project briefLab notes & perspectives002
A different scale
Why we’re starting with smaller models, and what we want to make open.
The note that isn’t there
The auditory phenomenon behind our name: a pitch can survive the absence of its fundamental frequency.
Papers000
The work comes first.
Our first papers are ahead of us. Technical reports and their supporting materials will appear here when they’re ready to share.