Deep Neural Network Based Guided Speech Bandwidth Extension

Schmidt K, Edler B. Deep Neural Network Based Guided Speech Bandwidth Extension. In: AES Convention 147. Audio Engineering Society; 2019. Paper 10254. Available from: https://aes.org/publications/elibrary-page/?id=20627

@inproceedings{Schmidt2019_20627,
  author = {Schmidt, Konstantin and Edler, Bernd},
  title = {{Deep Neural Network Based Guided Speech Bandwidth Extension}},
  booktitle = {AES Convention 147},
  note = {Paper 10254},
  year = {2019},
  month = oct,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=20627}
}

TY  - CPAPER
TI  - Deep Neural Network Based Guided Speech Bandwidth Extension
AU  - Schmidt, Konstantin
AU  - Edler, Bernd
T2  - AES Convention 147
M1  - Paper 10254
PY  - 2019
DA  - 2019/10/06
UR  - https://aes.org/publications/elibrary-page/?id=20627
PB  - Audio Engineering Society
LA  - en
AB  - Up to today telephone speech is still limited to the range of 200 to 3400 Hz since the predominant codecs in public switched telephone networks are AMR-NB, G.711, and G.722 [1, 2, 3]. Blind bandwidth extension (blind BWE, BBWE) can improve the perceived quality as well as the intelligibility of coded speech without changing the transmission network or the speech codec. The BBWE used in this work is based on deep neural networks (DNNs) and has already shown good performance [4]. Although this BBWE enhances the speech without producing too many artifacts it sometimes fails to enhance prominent fricatives that can result in muffled speech. In order to better synthesize prominent fricatives the BBWE is extended by sending a single bit of side information—here referred to as guided BWE. This bit may be transmitted, e.g., by watermarking so that no changes to the transmission network or the speech codec have to be done. Different DNN con?gurations (including convolutional (Conv.) layers as well as long short-term memory layers (LSTM)) making use of this bit have been evaluated. The BBWE has a low computational complexity and an algorithmic delay of 12 ms only and can be applied in state-of-the-art speech and audio codecs.
ER  -