Robust Recognition of Conversational Telephone Speech via Multi-condition Training and Data Augmentation
In this paper, we focus on automatic recognition of telephone conversational speech in scenario, when no amount of genuine telephone recordings is available for training. The training set contains only data from a significantly different domain, such as recording of broadcast news. Significant mismatch arises between training and test conditions, which leads to deteriorated performance of the resulting recognition system. We aim to diminish this mismatch using the data augmentation.
Speech compression and narrow-band spectrum are significant features of the telephone speech. We apply these effects to the training dataset artificially, in order to make it more similar to the desired test conditions. Using such augmented dataset, we subsequently train an acoustic model. Our experiments show that the augmented models achieve accuracy close to the results of a model trained on genuine telephone data. Moreover, when the augmentation is applied to the real-world telephone data, further accuracy gains are achieved.
KeywordsCompression Data augmentation Multi-conditional training Conversational speech
This work was supported by the Technology Agency of the Czech Republic (Project No. TH03010018).
- 1.Amodei, D., et al.: Deep speech 2: End-to-end speech recognition in English and Mandarin. In: International Conference on Machine Learning, pp. 173–182 (2016)Google Scholar
- 6.FFmpeg team: Ffmpeg - cross-platform solution to record, convert and stream audio and video. Software version: 20170525–b946bd8. https://www.ffmpeg.org/
- 7.Fraga-Silva, T., et al.: Active learning based data selection for limited resource STT and KWS. In: Sixteenth Annual Conference of the International Speech Communication Association (2015)Google Scholar
- 8.Fraga-Silva, T., et al.: Improving data selection for low-resource STT and KWS. In: 2015 IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU), pp. 153–159. IEEE (2015)Google Scholar
- 9.Garofolo, J.S., et al.: TIMIT acoustic-phonetic continuous speech corpus. Linguist. Data Consortium, 10(5) (1993)Google Scholar
- 10.Jaitly, N., Hinton, G.E.: Vocal tract length perturbation (VTLP) improves speech recognition. In: Proceeding of the ICML Workshop on Deep Learning for Audio, Speech and Language, pp. 625–660 (2013)Google Scholar
- 11.Kanda, N., Takeda, R., Obuchi, Y.: Elastic spectral distortion for low resource speech recognition with deep neural networks. In: 2013 IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU), pp. 309–314. IEEE (2013)Google Scholar
- 12.Kemp, T., Waibel, A.: Unsupervised training of a speech recognizer: recent experiments. In: Eurospeech (1999)Google Scholar
- 14.Kneser, R., Ney, H.: Improved backing-off for m-gram language modeling. In: 1995 International Conference on Acoustics, Speech, and Signal Processing 1995, ICASSP 1995, vol. 1, pp. 181–184. IEEE (1995)Google Scholar
- 15.Ko, T., Peddinti, V., Povey, D., Khudanpur, S.: Audio augmentation for speech recognition. In: INTERSPEECH, pp. 3586–3589 (2015)Google Scholar
- 17.Ma, J., Schwartz, R.: Unsupervised versus supervised training of acoustic models. In: Ninth Annual Conference of the International Speech Communication Association (2008)Google Scholar
- 19.Polacky, J., Jarina, R., Chmulik, M.: Assessment of automatic speaker verification on lossy transcoded speech. In: 2016 4th International Workshop on Biometrics and Forensics (IWBF), pp. 1–6. IEEE (2016)Google Scholar
- 20.Raghavan, S., et al.: A comparative study on the effect of different codecs on speech recognition accuracy using various acoustic modeling techniques. In: 2017 Twenty-third National Conference on Communications (NCC), pp. 1–6. IEEE (2017)Google Scholar
- 21.Seltzer, M.L., Yu, D., Wang, Y.: An investigation of deep neural networks for noise robust speech recognition. In: 2013 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 7398–7402. IEEE (2013)Google Scholar
- 22.Siegert, I., Lotz, A.F., Maruschke, M., Jokisch, O., Wendemuth, A.: Emotion intelligibility within codec-compressed and reduced bandwidth speech. In: ITG Symposium, Proceedings of Speech Communication, vol. 12, pp. 1–5. VDE (2016)Google Scholar
- 23.Torch team: Torch - a scientific computing framework for luajit. http://torch.ch
- 25.Xiong, W., et al.: The Microsoft 2016 conversational speech recognition system. In: 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 5255–5259. IEEE (2017)Google Scholar
- 26.Young, S., Young, S.: The HTK hidden Markov model toolkit: design and philosophy. Entrop. Cambridge Res. Lab. Ltd. 2, 2–44 (1994)Google Scholar