An Analysis of the Short Utterance Problem for Speaker Characterization
Source abstract
Abstract
Speaker characterization has always been conditioned by the length of the evaluated utterances. Despite performing well with large amounts of audio, significant degradations in performance are obtained when short utterances are considered. In this work we present an analysis of the short utterance problem providing an alternative point of view. From our perspective the performance in the evaluation of short utterances is highly influenced by the phonetic similarity between enrollment and test utterances. Both enrollment and test should contain similar phonemes to properly discriminate, being degraded otherwise. In this study we also interpret short utterances as incomplete long utterances where some acoustic units are either unbalanced or just missing. These missing units are responsible for the speaker representations to be unreliable. These unreliable representations are biased with respect to the reference counterparts, obtained from long utterances. These undesired shifts increase the intra-speaker variability, causing a significant loss of performance. According to our experiments, short utterances (3–60 s) can perform as accurate as if long utterances were involved by just reassuring the phonetic distributions. This analysis is determined by the current embedding extraction approach, based on the accumulation of local short-time information. Thus it is applicable to most of the state-of-the-art embeddings, including traditional i-vectors and Deep Neural Network (DNN) xvectors.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Speech Recognition and Synthesis · Speech and Audio Processing · Music and Audio Processing
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
ViVoLab, Aragón Institute for Engineering Research (I3A), University of Zaragoza, 50018 Zaragoza, Spain
Funding
Ministerio de Economia, Industria y Competitividad · Gobierno de Aragón
References (37)
Show source-provided references
1. Fifty years of progress in speech and speaker recognition10.1121/1.4784967 ↗
2. Evaluation of a vector quantization talker recognition system in text independent and text dependent modes10.1016/0885-2308(87)90005-2 ↗
3. Robust text-independent speaker identification using Gaussian mixture speaker models10.1109/89.365379 ↗
4. Kenny, P. (2005). Joint Factor Analysis of Speaker and Session Variability: Theory and Algorithms, CRIM. CRIM-06/08-13.
5. Front-End Factor Analysis for Speaker Verification10.1109/TASL.2010.2064307 ↗
6. Lei, Y., Scheffer, N., Ferrer, L., and McLaren, M. (2014, January 4–9). A Novel Scheme for Speaker Recognition Using a Phonetically-aware Deep Neural Network. Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Florence, Italy.10.21236/ADA613971 ↗
7. McLaren, M., Lei, Y., and Ferrer, L. (2015, January 19–24). Advances in Deep Neural Network Approaches to Speaker Recognition. Proceedings of the 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Brisbane, Australia.10.1109/ICASSP.2015.7178885 ↗
8. Snyder, D., Ghahremani, P., Povey, D., Garcia-Romero, D., Carmiel, Y., and Khudanpur, S. (2016, January 13–16). Deep Neural Network-based Speaker Embeddings for End-to-end Speaker Verification. Proceedings of the IEEE Spoken Language Technology Workshop (SLT), San Diego, CA, USA.10.1109/SLT.2016.7846260 ↗
9. Prince, S.J.D., and Elder, J.H. (2007, January 14–21). Probabilistic Linear Discriminant Analysis for Inferences About Identity. Proceedings of the IEEE International Conference on Computer Vision, Rio de Janeiro, Brazil.10.1109/ICCV.2007.4409052 ↗
10. Pairwise discriminative speaker verification in the I-vector space10.1109/TASL.2013.2245655 ↗
11. Speaker verification with short utterances: A review of challenges, trends and opportunities10.1049/iet-bmt.2017.0065 ↗
12. Mandasari, M.I., Mclaren, M., and Van Leeuwen, D.A. (2011, January 27–31). Evaluation of i-vector Speaker Recognition Systems for Forensic Applcation. Proceedings of the 12th Annual Conference of the International Speech Communication Association, Florence, Italy.10.21437/Interspeech.2011-6 ↗
13. Kanagasundaram, A., Dean, D., Sridharan, S., and Fookes, C. (2011). Domain adaptation based Speaker Recognition on Short Utterances. arXiv.10.21437/Interspeech.2011-58 ↗
14. NIST (2008). The NIST Year 2008 Speaker Recognition Evaluation Plan, Technical Report.
15. NIST (2010). The NIST 2010 Speaker Recognition Evaluation, Technical Report.
16. Multi-resolution time frequency feature and complementary combination for short utterance speaker recognition10.1007/s11042-013-1705-4 ↗
17. Alam, M.J., Kenny, P., and Stafylakis, T. (2015, January 6–10). Combining amplitude and phase-based features for speaker verification with short duration utterances. Proceedings of the Annual Conference of the International Speech Communication Association, Dresden, Germany.10.21437/Interspeech.2015-94 ↗
18. Vogt, R., Baker, B., and Sridharan, S. (2008, January 22–26). Factor analysis subspace estimation for speaker verification with short utterances. Proceedings of the Annual Conference of the International Speech Communication Association, Brisbane, Australia.10.21437/Interspeech.2008-274 ↗
19. Kanagasundaram, A., Dean, D., Gonzalez-Dominguez, J., Sridharan, S., Ramos, D., and Gonzalez-Rodriguez, J. (2013, January 29). Improving short utterance based I-vector speaker recognition using source and utterance-duration normalization techniques. Proceedings of the Annual Conference of the International Speech Communication Association, Lyon, France.10.21437/Interspeech.2013-411 ↗
20. Improving short utterance i-vector speaker verification using utterance variance modelling and compensation techniques10.1016/j.specom.2014.01.004 ↗
21. Sarkar, A.K., Matrouf, D., Bousquet, P.M., and Bonastre, J.F. (2012, January 9–13). Study of the Effect of I-vector Modeling on Short and Mismatch Utterance Duration for Speaker Verification. Proceedings of the Interspeech 2012, Portland, OR, USA.10.21437/Interspeech.2012-347 ↗
22. Hautamäki, V., Cheng, Y.C., Rajan, P., and Lee, C.H. (2013, January 25–29). Minimax i-vector extractor for short duration speaker verification. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, Lyon, France.10.21437/Interspeech.2013-696 ↗
23. Guo, J., Xu, N., Li, L.J., and Alwan, A. (2017, January 20–24). Attention based CLDNNs for short-duration acoustic scene classification. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, Stockholm, Sweden.10.21437/Interspeech.2017-440 ↗
24. Cumani, S., Plchot, O., and Laface, P. (2013, January 26–31). Probabilistic linear discriminant analysis of i-vector posterior distributions. Proceedings of the ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing, Vancouver, BC, Canada.10.1109/ICASSP.2013.6639150 ↗
25. PLDA for Speaker Verification with Utterances of Arbitrary Duration
26. Hasan, T., Saeidi, R., Hansen, J.H.L., and Van Leeuwen, D.A. (2013, January 26–31). Duration mismatch compensation for i-vector based speaker recognition systems. Proceedings of the ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing, Vancouver, BC, Canada.10.1109/ICASSP.2013.6639154 ↗
27. Quality measure functions for calibration of speaker recognition systems in various duration conditions10.1109/TASL.2013.2279332 ↗
28. Viñals, I., Ortega, A., Miguel, A., and Lleida, E. (2018). Phonetic Variability Influence on Short Utterances in Speaker Verification. Proc. Iberspeech, 6–9.10.21437/IberSPEECH.2018-2 ↗
29. Decisions With Minimal Speech
30. Ajili, M., Jean-François, B., Waad, B.K., Solange, R., and Juliette, K. (2016, January 13–16). Phonetic content impact on Forensic Voice Comparison. Proceedings of the 2016 IEEE Workshop on Spoken Language Technology, SLT, San Diego, CA, USA.
31. Learning Representations by Back-propagationg Errors10.1038/323533a0 ↗
32. Pelecanos, J., and Sridharan, S. (2001, January 18–22). Feature Warping for Robust Speaker Verification. Proceedings of the ODYSSEY-2001—The Speaker Recognition Workshop, Crete, Greece.
33. Garcia-Romero, D., and Espy-Wilson, C.Y. (2011, January 27–31). Analysis of I-vector Length Normalization in Speaker Recognition Systems. Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH, Florence, Italy.10.21437/Interspeech.2011-53 ↗
34. Siegler, M.A., Jain, U., Raj, B., and Stern, R.M. (1997, January 2–5). Automatic Segmentation, Classification and Clustering of Broadcast News Audio. Proceedings of the DARPA Speech Recognition Workshop, Chantilly, VA, USA.
35. Viñals, I., Ribas, D., Mingote, V., Llombart, J., Gimeno, P., Miguel, A., Ortega, A., and Lleida, E. (2019). Phonetically-aware embeddings, Wide Residual Networks with Time-Delay Neural Networks and Self Attention models for the 2018 NIST Speaker Recognition Evaluation. Interspeech, submitted.10.21437/Interspeech.2019-2417 ↗
36. Zagoruyko, S., and Komodakis, N. (2016). Wide Residual Networks. arXiv.10.5244/C.30.87 ↗
37. Garofolo, J.S., Lamel, L.F., Fisher, W.M., Fiscus, J.G., Pallett, D.S.D., and Dahlgren, N.L. (1993). TIMIT Acoustic-Phonetic Continuous Speech Corpus LDC93S1, Linguistic Data Consortium.