Robust Low Rate Speech Coding Based on Cloned Networks and Wavenet

By Felicia S. C. Lim, W. Kleijn, Michael Chinen, Jan Skoglund. 2020

13 citations recorded by the source corpus.

Rapid advances in machine-learning based generative modeling of speech make its use in speech coding attractive. However, the current performance of such models drops rapidly with noise contamination of the input, preventing use in practical applications. We present a new speech-coding scheme that is based on features that are robust to the distortions occurring in speech-coder input signals. To this purpose, we encourage the feature encoder to provide the same independent features for each of a set of linguistically equivalent signals, obtained by adding various noises to a common clean signal. The independent features, subjected to scalar quantization, are used as a conditioning vector sequence for WaveNet. Our experiments show that a 1.8 kb/s implementation of the resulting coder provides state-of-the-art performance for clean signals, and is additionally robust to noisy input.

Subfields: Artificial Intelligence, Speech & Audio, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Speech & Audio

In Signal Processing

Citations in the corpus

High Fidelity Neural Audio Compression (2022, background): “Neural based audio codecs have been recently proposed and demonstrated promising results (Kleijn et al., 2018; Valin & Skoglund, 2019b; Lim et al., 2020; Kleijn et al., 2021; Zeghidour et al., 2021; Omran et al., 2022; Lin et al., 2022; Jayashankar et al., 2022; Li et al.; Jiang et al., 2022),…”
Speech Resynthesis from Discrete Disentangled Self-Supervised Representations (2021, background): “Neural speech codecs have been recently proposed and demonstrated promising results [30, 31].”
Enhancing into the Codec: Noise Robust Speech Coding with Vector-Quantized Autoencoders (2021, methodology): “In [12], this issue was addressed by training a noise-robust feature extractor based on the Siamese learning paradigm, and then training a WaveNet model conditioned on those features.”
Neural Feature Predictor and Discriminative Residual Coding for Low-Bitrate Speech Coding (2022, methodology): “Various other methods have been developed to improve the quality of the features, regarding robustness [18, 19], scalability [20] and the variability issues [21].”