Deep Vocoder: Low Bit Rate Compression of Speech with Deep Autoencoder
This work addresses speech compression for low bit rate applications, offering incremental improvements over conventional codecs.
The paper tackled low bit rate speech compression by proposing Deep Vocoder, an end-to-end method using a deep autoencoder and analysis-by-synthesis vector quantization, resulting in improved frequency-weighted segmental SNR, STOI, and PESQ scores, with PESQ scores of 3.34 at 2400 bit/s and 3.08 at 1200 bit/s on the TIMIT corpus.
Inspired by the success of deep neural networks (DNNs) in speech processing, this paper presents Deep Vocoder, a direct end-to-end low bit rate speech compression method with deep autoencoder (DAE). In Deep Vocoder, DAE is used for extracting the latent representing features (LRFs) of speech, which are then efficiently quantized by an analysis-by-synthesis vector quantization (AbS VQ) method. AbS VQ aims to minimize the perceptual spectral reconstruction distortion rather than the distortion of LRFs vector itself. Also, a suboptimal codebook searching technique is proposed to further reduce the computational complexity. Experimental results demonstrate that Deep Vocoder yields substantial improvements in terms of frequency-weighted segmental SNR, STOI and PESQ score when compared to the output of the conventional SQ- or VQ-based codec. The yielded PESQ score over the TIMIT corpus is 3.34 and 3.08 for speech coding at 2400 bit/s and 1200 bit/s, respectively.