Robust automatic transcription of English speech corpora

Kabir, Ahsanul; Giurgiu, Mircea; Barker, Jon

doi:10.1109/iccomm.2010.5509116

Cited by 4 publications

(3 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The DARPA Robust Automatic Transcription of Speech (RATS) corpus is comprised of a large amount of primarily telephony audio from various languages degraded by transmitting the audio through 8 different radio transmission conditions. Further details about the corpus and channel conditions used can be found in Kabir et al (0000) . The corpus contains speech in Arabic, Farsi, Dari, Urdu, and Pashtu.…”

Section: Corpora Evaluation and Benchmarkingmentioning

confidence: 99%

Nonlinear waveform distortion: Assessment and detection of clipping on speech data and systems

Hansen

Stauffer

Xia

2021

Speech Communication

View full text Add to dashboard Cite

Section: Corpora Evaluation and Benchmarkingmentioning

confidence: 99%

Nonlinear waveform distortion: Assessment and detection of clipping on speech data and systems

Hansen

Stauffer

Xia

2021

Speech Communication

View full text Add to dashboard Cite

“…Labels were taken from the clean sets when possible to ensure they convey the spoken message; for Aurora 4 using the given clean close-microphone condition; for CHiME 3, as there are no available clean labels, we used the ones produced by the force-alignment from the best performing setup; the CHiME 2 clean labels were used from forced alignment system detailed in (Kabir et al, 2010).…”

Section: Discriminability Criteriamentioning

confidence: 99%

On the relevance of auditory-based Gabor features for deep learning in robust speech recognition

Martínez

Mallidi

Meyer

2017

Computer Speech & Language

View full text Add to dashboard Cite

Previous studies support the idea of merging auditory-based Gabor features with deep learning architectures to achieve robust automatic speech recognition, however, the cause behind the gain of such combination is still unknown. We believe these representations provide the deep learning decoder with more discriminable cues. Our aim with this paper is to validate this hypothesis by performing experiments with three different recognition tasks (Aurora 4, CHiME 2 and CHiME 3) and assess the discriminability of the information encoded by Gabor filterbank features. Additionally, to identify the contribution of low, medium and high temporal modulation frequencies subsets of the Gabor filterbank were used as features (dubbed LTM, MTM and HTM respectively). With temporal modulation frequencies between 16 and 25 Hz, HTM consistently outperformed the remaining ones in every condition, highlighting the robustness of these representations against channel distortions, low signal-to-noise ratios and acoustically challenging real-life scenarios with relative improvements from 11 to 56% against a Mel-filterbank-DNN baseline. To explain the results, a measure of similarity between phoneme classes from DNN activations is proposed and linked to their acoustic properties. We find this measure to be consistent with the observed error rates and highlight specific differences on phoneme level to pinpoint the benefit of the proposed features.

show abstract

“…"bin blue at f two now"). Phonetic transcriptions were automatically generated using a Viterbi forced alignment procedure [12].…”

Section: Experimental Conditionmentioning

confidence: 99%

Continuous visual speech recognition for audio speech enhancement

Benhaim

Sahbi

Vittey

2015

2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

We introduce in this paper a novel non-blind speech enhancement procedure based on visual speech recognition (VSR). The latter is based on a generative process that analyzes sequences of talking faces and classifies them into visual speech units known as visemes. We use an effective graphical model able to segment and label a given sequence of talking faces into a sequence of visemes. Our model captures unary potential as well as pairwise interaction; the former models visual appearance of speech units while the latter models their interactions using boundary and visual language model activations. Experiments conducted on a standard challenging dataset, show that when feeding the results of VSR to the speech enhancement procedure, it clearly outperforms baseline blind methods as well as related work.

show abstract

Robust automatic transcription of English speech corpora

Cited by 4 publications

References 5 publications

Nonlinear waveform distortion: Assessment and detection of clipping on speech data and systems

Nonlinear waveform distortion: Assessment and detection of clipping on speech data and systems

On the relevance of auditory-based Gabor features for deep learning in robust speech recognition

Continuous visual speech recognition for audio speech enhancement

Contact Info

Product

Resources

About