Enhancing into the codec: Noise Robust Speech Coding with Vector-Quantized Autoencoders
This addresses the issue of noise robustness in speech coding for applications requiring reliable audio compression in real-world environments, representing an incremental improvement.
The paper tackled the problem of neural audio codecs producing unintended outputs in noisy conditions by developing compressor-enhancer encoders and decoders based on VQ-VAE autoencoders with WaveRNN decoders, showing they operate well in noisy conditions and even outperform models trained only on clean speech on clean inputs.
Audio codecs based on discretized neural autoencoders have recently been developed and shown to provide significantly higher compression levels for comparable quality speech output. However, these models are tightly coupled with speech content, and produce unintended outputs in noisy conditions. Based on VQ-VAE autoencoders with WaveRNN decoders, we develop compressor-enhancer encoders and accompanying decoders, and show that they operate well in noisy conditions. We also observe that a compressor-enhancer model performs better on clean speech inputs than a compressor model trained only on clean speech.