InSE-NET: A Perceptually Coded Audio Quality Model based on CNN

Jiang G, Biswas A, Bergler C, Maier A. InSE-NET: A Perceptually Coded Audio Quality Model based on CNN. In: AES Convention 151. Audio Engineering Society; 2021. Paper 10514. Available from: https://aes.org/publications/elibrary-page/?id=21478

@inproceedings{Jiang2021_21478,
  author = {Jiang, Guanxin and Biswas, Arijit and Bergler, Christian and Maier, Andreas},
  title = {{InSE-NET: A Perceptually Coded Audio Quality Model based on CNN}},
  booktitle = {AES Convention 151},
  note = {Paper 10514},
  year = {2021},
  month = oct,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=21478}
}

TY  - CPAPER
TI  - InSE-NET: A Perceptually Coded Audio Quality Model based on CNN
AU  - Jiang, Guanxin
AU  - Biswas, Arijit
AU  - Bergler, Christian
AU  - Maier, Andreas
T2  - AES Convention 151
M1  - Paper 10514
PY  - 2021
DA  - 2021/10/06
UR  - https://aes.org/publications/elibrary-page/?id=21478
PB  - Audio Engineering Society
LA  - en
AB  - Automatic coded audio quality assessment is an important task whose progress is hampered by the scarcity of human annotations, poor generalization to unseen codecs, bitrates, content-types, and a lack of flexibility of existing approaches. One of the typical human-perception-related metrics, ViSQOL v3 (ViV3), has been proven to provide a high correlation to the quality scores rated by humans. In this study, we take steps to tackle problems of predicting coded audio quality by completely utilizing programmatically generated data that is informed with expert domain knowledge. We propose a learnable neural network, entitled InSE-NET, with a backbone of Inception and Squeeze-and-Excitation modules to assess the perceived quality of coded audio at a 48 kHz sample rate. We demonstrate that synthetic data augmentation is capable of enhancing the prediction. Our proposed method is intrusive, i.e. it requires Gammatone spectrograms of unencoded reference signals. Besides a comparable performance to ViV3, our approach provides a more robust prediction towards higher bitrates.
ER  -