MagiCodec: Simple Masked Gaussian-Injected Codec for High-Fidelity Reconstruction and Generation
This addresses a key limitation in audio codecs for generative AI applications, offering improved token expressiveness while maintaining high fidelity, though it is an incremental advancement in method design.
The authors tackled the bottleneck of neural audio codecs prioritizing reconstruction over downstream modelability by introducing MagiCodec, a Transformer-based codec with Gaussian noise injection, which achieved state-of-the-art performance in reconstruction and downstream tasks, producing tokens with Zipf-like distributions for better compatibility with language models.
Neural audio codecs have made significant strides in efficiently mapping raw audio waveforms into discrete token representations, which are foundational for contemporary audio generative models. However, most existing codecs are optimized primarily for reconstruction quality, often at the expense of the downstream modelability of the encoded tokens. Motivated by the need to overcome this bottleneck, we introduce $\textbf{MagiCodec}$, a novel single-layer, streaming Transformer-based audio codec. MagiCodec is designed with a multistage training pipeline that incorporates Gaussian noise injection and latent regularization, explicitly targeting the enhancement of semantic expressiveness in the generated codes while preserving high reconstruction fidelity. We analytically derive the effect of noise injection in the frequency domain, demonstrating its efficacy in attenuating high-frequency components and fostering robust tokenization. Extensive experimental evaluations show that MagiCodec surpasses state-of-the-art codecs in both reconstruction quality and downstream tasks. Notably, the tokens produced by MagiCodec exhibit Zipf-like distributions, as observed in natural languages, thereby improving compatibility with language-model-based generative architectures. The code and pre-trained models are available at https://github.com/Ereboas/MagiCodec.