Adaptive Oscillatory Inductive Bias for Modeling Sharp Prosodic Dynamics in Diffusion-Based TTS

Sandipan Dhar, Nirmesh Shah, Ashishkumar Gudmalwar, Pankaj Wasnik

Media Analysis Group, Sony Research India, Bangalore

Email:pankaj.wasnik@sony.com

Diffusion-based text-to-speech (TTS) models have achieved significant improvements in speech quality. However, modeling sharp prosodic transitions and rapid pitch variations in expressive speech remains challenging. Existing diffusion-based TTS decoders commonly utilize periodic nonlinearities such as Snake activation function to capture harmonic structures, but this activation funcation provides limited adaptability when modeling abrupt amplitude and frequency variations. In this paper, we investigate the role of oscillatory inductive bias in diffusion-based TTS decoders and introduce an adaptive oscillatory nonlinearity that enables controllable periodic modulation while maintaining signal stability through a linear bypass component. We refer the resulting TTS system as OscillaTTS. Experiments on the LJSpeech and Emotional Speech Dataset show consistent improvements across objective and subjective evaluations, indicating improved modeling of expressive prosodic dynamics.

Accepted at INTERSPEECH 2026

Figure 1: Schematic overview of the proposed OscillaTTS architecture based on StyleTTS2.

Demo Samples

Text StyleTTS2 Oscilla-TTS GradTTS GlowTTS FastSpeech2 Ground Truth
In Fort Worth, there occurred a breach of discipline by some members of the Secret Service who were officially traveling with the President.
The Secret Service had received from the FBI some nine thousand reports on members of the Communist Party.
Might have been more alert in the Dallas motorcade if they had retired promptly in Fort Worth.
It has also used other Federal law enforcement agents during Presidential visits to cities in which such agents are stationed.
Three of these agents occupied positions on the running boards of the car, and the fourth was seated in the car.
A formal and thorough description of the responsibilities of the advance agent is now in preparation by the Service.

Emotional Demo Samples Proposed Oscilla-TTS

Text Angry Happy Sad
The nine the eggs, I keep.
I did go, and made many prisoners.
That I owe my thanks to you.
Clear than clear water!.

Emotional Demo Samples Baseline StyleTTS2

Text Angry Happy Sad
The nine the eggs, I keep.
I did go, and made many prisoners.
That I owe my thanks to you.
Clear than clear water!.

Citation

@inproceedings{Oscilla2026,
    title={Adaptive Oscillatory Inductive Bias for Modeling Sharp Prosodic Dynamics in Diffusion-Based TTS},
    author={Dhar, Sandipan and Shah, Nirmesh and Gudmalwar, Ashishkumar and Wasnik, Pankaj},
    booktitle={INTERSPEECH},
    year={2026}
  }