SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition
Open accessThe DOI matches its Crossref record; 51 of the 51 references it cites were found in a registry or an index.
Abstract
We present SpecAugment, a simple data augmentation method for speech recognition. SpecAugment is applied directly to the feature inputs of a neural network (i.e., filter bank coefficients). The augmentation policy consists of warping the features, masking blocks of frequency channels, and masking blocks of time steps. We apply SpecAugment on Listen, Attend and Spell networks for end-to-end speech recognition tasks. We achieve state-of-the-art performance on the LibriSpeech 960h and Swichboard 300h tasks, outperforming all prior work. On LibriSpeech, we achieve 6.8% WER on test-other without the use of a language model, and 5.8% WER with shallow fusion with a language model. This compares to the previous state-of-the-art hybrid system of 7.5% WER. For Switchboard, we achieve 7.2%/14.6% on the Switchboard/CallHome portion of the Hub5'00 test set without the use of a language model, and 6.8%/14.1% with shallow fusion, which compares to the previous state-of-the-art hybrid system at 8.3%/17.3% WER.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“SpecAugment is a spectrum augmentation method, which is only suitable for the audio deepfake model based on front-end feature extraction, and it is not easy to operate on the audio waveform.”
From Voice Deepfake Detection Using the Self-Supervised Pre-Training Model HuBERT (2023) · cites this paper as [19]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“In addition to balanced sampling, we also apply SpecAugment [23] and random time shift to alleviate the class-imbalance issue.”
From Vocalsound: A Dataset for Improving Human Vocal Sounds Recognition (2022) · cites this paper as [23]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Rethinking the Inception Architecture for Computer Vision
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.308 · Cited by 32,138
Speech recognition with deep recurrent neural networks
Graves, A., Mohamed, A.-R., & Hinton, G. (2013) · 2013 IEEE International Conference on Acoustics, Speech and Signal Processing · doi:10.1109/icassp.2013.6638947 · Cited by 8,893 · cited 1×
“Deep Learning has been applied successfully to Automatic Speech Recognition (ASR) [1], where the main focus of research has been designing better network architectures, for example, DNNs [2], CNNs [3], RNNs [4] and end-to-end models [5, 6, 7].”
Method“Librispeech: an ASR corpus based on public domain audio books,”
Panayotov, V., Chen, G., Povey, D., & Khudanpur, S. (2015) · 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2015.7178964 · Cited by 6,072
Kaldi Speech Recognition Toolkit
Povey, D. (2024) · Infoscience (Ecole Polytechnique Fédérale de Lausanne) · doi:10.57702/jb3fvbn9 · Cited by 4,609
Context-dependent pre-trained deep neural networks for large-vocabulary speech recognition
Dahl, G. E., Yu, D., Deng, L., & Acero, A. (2012) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/tasl.2011.2134090 · Cited by 3,011 · cited 1×
“Deep Learning has been applied successfully to Automatic Speech Recognition (ASR) [1], where the main focus of research has been designing better network architectures, for example, DNNs [2], CNNs [3], RNNs [4] and end-to-end models [5, 6, 7].”
MethodImproved regularization of convolutional neural networks with cutout
DeVries, T., & Taylor, G. W. (2017) · arXiv · doi:10.48550/arxiv.1708.04552 · Cited by 2,776
AutoAugment: Learning Augmentation Strategies From Data
Cubuk, E. D., Zoph, B., Mané, D., Vasudevan, V., & Le, Q. V. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00020 · Cited by 2,733
Self-citationTowards End-To-End Speech Recognition with Recurrent Neural Networks
Graves, A., & Jaitly, N. (2014) · Cited by 2,557
Listen, attend and spell: A neural network for large vocabulary conversational speech recognition
Chan, W., Jaitly, N., Le, Q., & Vinyals, O. (2016) · 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2016.7472621 · Cited by 2,226
Self-citationSWITCHBOARD: telephone speech corpus for research and development
Godfrey, J. J., Holliman, E. C., & McDaniel, J. (1992) · [Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp.1992.225858 · Cited by 2,134 · cited 2×
“On Switchboard 300h (LDC97S62) [22], we obtain 7.2% WER on the Switchboard portion of the Hub5’00 (LDC2002S09, LDC2003T02) test set, and 14.6% on the Call-Home portion, without using an LM.”
BackgroundInfluential
Research works printed but not linked: 2
- [41][41] K. Vesely, A. Ghoshal, L. Burger, and D. Povey, 'Sequencediscriminative training of deep neural networks,' in INTERSPEECH , 2013. -
- [43][43] G. Zweig, C. Yu, J. Droppo, and A. Stolcke, 'Advances in AllNeural Speech Recognition,' in ICASSP , 2017. -
Cited by
Items 1,001–1,013 of 1,013
Shin, H.-S., Heo, J., Kim, J.-H., Lim, C.-Y., Kim, W., & Yu, H.-J. (2024) · ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp48485.2024.10448453 · Cited by 21
“To augment training data, we employed media codec 1 augmentation, speed perturbation set to 0.9 and 1.1, SpecAugment [22] for randomly masking 0 to 20 frequency axes, and colored additive noise augmentation with signal-to-noise ratios randomly selected between 10 and 40.”
MethodInfluentialExplainable Deepfake and Spoofing Detection: An Attack Analysis Using SHapley Additive exPlanations
Ge, W., Todisco, M., & Evans, N. (2022) · The Speaker and Language Recognition Workshop (Odyssey 2022) · doi:10.21437/odyssey.2022-10 · Cited by 18
Generalizable Audio Deepfake Detection via Latent Space Refinement and Augmentation
Huang, W., Gu, Y., Wang, Z., Zhu, H., & Qian, Y. (2025) · arXiv · doi:10.48550/arxiv.2501.14240 · Cited by 16
eP-ALM: Efficient Perceptual Augmentation of Language Models
Shukor, M., Dancette, C., & Cord, M. (2023) · 2023 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.02016 · Cited by 16
Zheng, X., Meng, D., Chen, D., Wong, W.-K., To, K.-H., Zhu, L., Wu, J., Liang, Y., Leung, K.-S., Wong, M.-H., et al. (2024) · PLOS Computational Biology · doi:10.1371/journal.pcbi.1012083 · Cited by 14
CoughLIME: Sonified Explanations for the Predictions of COVID-19 Cough Classifiers
Wullenweber, A., Akman, A., & Schuller, B. W. (2022) · 2022 44th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) · doi:10.1109/embc48229.2022.9871291 · Cited by 13
“Employing these rather simple decompositions is motivated by the fact that they have led to excellent results for audio data augmentation [16].”
ESB: A Benchmark For Multi-Domain End-to-End Speech Recognition
Gandhi, S., von Platen, P., & Rush, A. M. (2022) · arXiv · doi:10.48550/arxiv.2210.13352 · Cited by 13
Enhancing the Robustness of AI-Generated Text Detectors: A Survey
Liu, X., Li, Y., & Li, K. (2025) · Mathematics · doi:10.3390/math13132145 · Cited by 10
SpeechNet: A Universal Modularized Model for Speech Processing Tasks
Chen, Y.-C., Chi, P.-H., Yang, S.-W., Chang, K.-W., Lin, J.-H., Huang, S.-F., Liu, D.-R., Liu, C.-L., Lee, C.-K., & Lee, H.-Y. (2021) · arXiv · doi:10.48550/arxiv.2105.03070 · Cited by 6
“Particularly during the training of ASR, we apply frequency and time masking augmentation (Park et al., 2019) on the speech inputs.”
MethodFasoli, A., Chen, C.-Y., Serrano, M., Venkataramani, S., Saon, G., Cui, X., Kingsbury, B., & Gopalakrishnan, K. (2022) · Interspeech 2022 · doi:10.21437/interspeech.2022-413 · Cited by 5
Unified Keyword Spotting and Audio Tagging on Mobile Devices with Transformers
Dinkel, H., Wang, Y., Yan, Z., Zhang, J., & Wang, Y. (2023) · ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp49357.2023.10095534 · Cited by 4
“Having obtained a spectrogram, we augment the data using Specaugment [25].”
MethodOn robustness in multimodal learning
McKinzie, B., Cheng, J., Shankar, V., Yang, Y., Shlens, J., & Toshev, A. (2023) · arXiv · doi:10.48550/arxiv.2304.04385 · Cited by 1
Unsupervised Speech Recognition
Baevski, A., Hsu, W.-N., Conneau, A., & Auli, M. (2021) · arXiv · doi:10.48550/arxiv.2105.11084
Metadata
- Authors
- Park, D. S., Chan, W., Zhang, Y., Chiu, C.-C., Zoph, B., Cubuk, E. D., & Le, Q. V.
- Affiliations
- Google (United States) (US)
- Venue
- Interspeech 2019
- Publisher
- ISCA
- Published
- 2019-09-15
- Pages
- 2613-2617
- ISSN
- 2308-457X (ISSN-L)
- OpenAlex
- W2936774411
- Cited by
- 3,558
- References
- 51
- Citations by year
- 2026: 239 · 2025: 415 · 2024: 450 · 2023: 649 · 2022: 603 · 2021: 738 · 2020: 391 · 2019: 69
- FWCI
- 172.29
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Interspeech (USB) (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Physical Sciences › Computer Science › Signal Processing › Music and Audio Processing
- Keywords
- SpecAugment, data augmentation, speech recognition, frequency masking, feature warping, end-to-end speech recognition, LibriSpeech, switchboards, CALLHOME, word error rate
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.