DOI QR코드

DOI QR Code

주파수 영역 심층 신경망 기반 음성 향상을 위한 실수 네트워크와 복소 네트워크 성능 비교 평가

Performance comparison evaluation of real and complex networks for deep neural network-based speech enhancement in the frequency domain

  • 황서림 (연세대학교 지능형신호처리연구실) ;
  • 박성욱 (강릉원주대학교 전자공학과) ;
  • 박영철 (연세대학교 지능형신호처리연구실)
  • 투고 : 2021.12.03
  • 심사 : 2022.01.04
  • 발행 : 2022.01.31

초록

본 논문은 주파수 영역에서 심층 신경망 기반 음성 향상 모델 학습을 위하여 학습 대상과 네트워크 구조에 따라 두 가지 관점에서 성능을 비교 평가한다. 이때, 학습 대상으로는 스펙트럼 매핑과 Time-Frequency(T-F) 마스킹 기법을 사용하였고 네트워크 구조는 실수 네트워크와 복소 네트워크를 사용하였다. 음성 향상 모델의 성능은 데이터 셋 규모에 따라 Perceptual Evaluation of Speech Quality(PESQ)와 Short-Time Objective Intelligibility(STOI) 두 가지 객관적 평가지표를 통해 평가하였다. 실험 결과, 네트워크의 종류와 데이터 셋 종류에 따라 적정한 훈련 데이터의 크기가 다르다는 것을 확인하였다. 또한, 데이터의 크기와 학습 대상에 따라 복소 네트워크보다 실수 네트워크가 비교적 높은 성능을 보이기 때문에 총 파라미터의 수를 고려한다면 경우에 따라 실수 네트워크를 사용하는 것이 보다 현실적인 해결책일 수 있다는 것을 확인하였다.

This paper compares and evaluates model performance from two perspectives according to the learning target and network structure for training Deep Neural Network (DNN)-based speech enhancement models in the frequency domain. In this case, spectrum mapping and Time-Frequency (T-F) masking techniques were used as learning targets, and a real network and a complex network were used for the network structure. The performance of the speech enhancement model was evaluated through two objective evaluation metrics: Perceptual Evaluation of Speech Quality (PESQ) and Short-Time Objective Intelligibility (STOI) depending on the scale of the dataset. Test results show the appropriate size of the training data differs depending on the type of networks and the type of dataset. In addition, they show that, in some cases, using a real network may be a more realistic solution if the number of total parameters is considered because the real network shows relatively higher performance than the complex network depending on the size of the data and the learning target.

키워드

참고문헌

  1. A. Narayanan and D. Wang, "Ideal ratio mask estimation using deep neural networks for robust speech recognition," Proc. IEEE ICASSP. 7092-7096 (2013).
  2. T. Gerkmann, M. Krawczyk-Becker, and J. Le Roux, "Phase processing for single-channel speech enhancement: History and recent advances," IEEE Signal Process. Mag. 32, 55-66 (2015).
  3. H.-S. Choi, J-H Kim, J. Huh, A. Kim, J.-W. Ha, and K. Lee,"Phase-aware speech enhancement with deep complexu-net," Proc. ICLR. 2019.
  4. S. A. Nossier, J. Wall, M. Moniri, C. Glackin, and N. Cannings, "Mapping and masking targets comparison using different deep learning based speech enhancement architectures," Proc. IJCNN. 1-8 (2020).
  5. K. Paliwal, K. Wojcicki, and B. Shannon, "The importance of phase in speech enhancement," Speech Commun. 53, 465-494 (2011). https://doi.org/10.1016/j.specom.2010.12.003
  6. K. Tan and D. Wang, "Complex spectral mapping with a convolutional recurrent network for monaural speech enhancement," Proc. IEEE ICASSP. 6865-6869 (2019).
  7. Y. Hu, Y. Liu, S. Lv, M. Xing, S. Zhang, Y. Fu, J. Wu, B. Zhang, and L. Xie, "Dccrn: Deep complex convolution recurrent network for phase-aware speech enhancement," Proc. Interspeech, 2472-2476 (2020).
  8. S. Santurkar, D. Tsipras, A. Ilyas, and A. Madry, "How does batch normalization help optimization?," Proc. NeurIPS. 1-11 (2018).
  9. C. K. Reddy, V. Gopal, R. Cutler, E. Beyrami, R. Cheng, H. Dybey, S. Matusevych, R. Aichner, A. Aazami, S. Braun, and J. Gehrke, "The interspeech 2020 deep noise suppression challenge: Dataset, subjective testing framework, and challenge results," arXiv preprint arXiv:2005.13981 (2020).
  10. J. S. Garofolo, L. F. Lamel, W. M. Fisher, J. G. Fiscus, D. S. Pallett, N. L. Dahlgren, and V. Zue, "Timit acoustic phonetic continuous speech corpus," Linguistic Data Consortium (1993).
  11. A. Varga and H. J. M. Steeneken, "Assessment for automatic speech recognition: Ii. noisex-92: A database and an experiment to study the effect of additive noise on speech recognition systems," Speech commun. 12, 247-251 (1993). https://doi.org/10.1016/0167-6393(93)90095-3
  12. E. Vincent, J. Barker, S. Watanabe, J. Le Roux, F. Nesta, and M. Matassoni, "The second 'chime'speech separation and recog-nition challenge: Datasets, tasks and baselines," Proc. IEEE ICASSP. 126-130 (2013).
  13. J. Barker, R. Marxer, E. Vincent, and S. Watanabe, "The third 'chime' speech separation and recognition challenge: Dataset, taskand baselines," Proc. ISRU. 504-511 (2015).
  14. A. W. Rix, J. G. Beerends, M. P. Hollier, and A. P. Hekstra, "Perceptual evaluation of speech quality (PESQ)-A new method for speech quality assessment of telephone networks and coders," Proc. IEEE ICASSP. 749-752 (2001).
  15. C. H. Taal, R. C. Hendriks, R. Heusdens, and J. Jensen, "An algorithm for intelligibility prediction of time-frequency weighted noisy speech," IEEE Trans. on Audio, Speech, and Lang. Process. 19, 2125-2136 (2011). https://doi.org/10.1109/TASL.2011.2114881