Lesly Widner, PhD

Helping your TTS sound more natural – with linguistic expertise in prosody and speech acts.

I’m a linguist specializing in prosody and speech acts. My work focuses on how rhythm, pitch, and intensity convey meaning in real-life communication — and how these patterns can be described, modeled, and applied to improve natural-sounding TTS and other voice-based systems.

I don’t build AI models myself, but I provide the linguistic insights that help make them sound more human.

Understanding how prosody shapes meaning in human speech gives you a powerful edge when developing voice-based systems. With my work, you can identify the specific prosodic patterns that align with different communicative functions—like requesting, insisting, confirming, or expressing urgency—and apply this knowledge to design more natural, context-aware TTS outputs. Whether you're refining how your system responds or improving how it speaks, these insights help bridge the gap between sounding robotic and sounding human. You can learn more about how to apply prosody and speech acts to TTS here.

What I do

What Sets My Research Apart

How You Can Apply My Work to Your Projects

What makes my work unique is that, as far as I know, no one has yet offered a comprehensive overview of how prosodic features—like pitch, rhythm, and intensity—work together in real-life communication. This integrated perspective simply isn’t available elsewhere.
At the same time, while the idea of communicative functions may sound simple, there is still no typology that’s both cross-linguistically applicable and systematically linked to prosodic patterns. That’s the gap I’m addressing—bringing together functional and prosodic analysis in a way that’s clear, structured, and useful for real-world implementation.