Sylber 2.0: A Universal Syllable Embedding

TL;DR

Sylber 2.0 introduces a universal, efficient syllable-based speech coding framework that captures detailed acoustic and linguistic features across languages, enabling high-quality TTS and improved low-resource ASR performance.

Contribution

It presents Sylber 2.0, a self-supervised syllable embedding model that achieves low-frequency, high-fidelity speech representation across multiple languages and expressive styles.

Findings

Performs on par with high-frequency models in speech tasks.

Enables TTS with competitive quality using only 72M parameters.

Provides effective features for low-resource ASR.

Abstract

Scaling spoken language modeling requires speech tokens that are both efficient and universal. Recent work has proposed syllables as promising speech tokens at low temporal resolution, but existing models are constrained to English and fail to capture sufficient acoustic detail. To address this gap, we present Sylber 2.0, a self-supervised framework for coding speech at the syllable level that enables efficient temporal compression and high-fidelity reconstruction. Sylber 2.0 achieves a very low token frequency around 5 Hz, while retaining both linguistic and acoustic detail across multiple languages and expressive styles. Experiments show that it performs on par with previous models operating on high-frequency baselines. Furthermore, Sylber 2.0 enables efficient TTS modeling which can generate speech with competitive intelligibility and quality with SOTA models using only 72M parameters. Moreover, the universality of Sylber 2.0 provides more effective features for low resource ASR than previous speech coding frameworks. In sum, we establish an effective syllable-level abstraction for general spoken language.