LLM-Codec: Neural Audio Codec Meets Language Model Objectives
For researchers building spoken language models, this work addresses the mismatch between codec reconstruction and autoregressive prediction, improving both token predictability and reconstruction fidelity.
LLM-Codec augments neural audio codec training with language-model-facing objectives, improving token predictability for spoken language models. It achieves 61.6% accuracy on SALMon speech coherence (+12.1 points over AUV) and reduces perplexity by 35, while also improving speech Mel distance by 5.0% on Codec-SUPERB-tiny.
Neural audio codecs are widely used as tokenizers for spoken language models, but they are optimized for waveform reconstruction rather than autoregressive prediction. This mismatch injects acoustically driven uncertainty into the discrete token space and increases language-model perplexity. We propose \ours, which augments codec training with language-model-facing objectives while keeping both codec and LLM architectures unchanged. \ours introduces (i) future token prediction with Medusa-style multi-step heads to encourage multi-step predictability, and (ii) semantic alignment that matches audio and text representations via a memory-bank contrastive loss. A differentiable Gumbel bridge enables end-to-end gradients from these objectives to the codec encoder. On SALMon speech coherence, token LMs trained on \ours reach 61.6% accuracy (+12.1 points over AUV) while reducing perplexity 35. On Codec-SUPERB-tiny, \ours improves speech Mel distance by 5.0% over AUV while simultaneously achieving the learnability gains, demonstrating that reconstruction fidelity and token predictability can be improved together.