eess.ASMay 28, 2026

MELD: Mel-Spectrogram-Based Speech Language Modeling with Discrete Latent Variables

Authors: Sung-Lin Yeh, Wei Zhou, Gil Keren, Duc Le, Zhong Meng, Hao Tang, Jay Mahadeokar, Ozlem Kalinli, +1 more

Organizations: University of Edinburgh · *Work done at Meta · 2Google DeepMind · 3Meta Superintelligence Labs

Abstract

Recent speech language models rely on encoders that are optimized separately from autoregressive models. Since these encoders are unaware of the downstream objectives, the extracted representations may not be optimal for downstream tasks. To address this limitation, we introduce a discrete latent variable model on mel spectrograms that jointly optimizes the encoder and the speech language model. Joint optimization not only brings improvements over codec-based and other mel-spectrogram-based baselines on zero-shot Text-to-Speech (TTS) and Speech-to-Text (STT) tasks, but also effectively alleviates common issues in autoregressive mel spectrogram modeling, such as prolonged silence generation and word omissions.

Explore similar work

CardsList