CHiME2 WSJ0
| Item Name: | CHiME2 WSJ0 |
| Author(s): | Emmanuel Vincent, Jon Barker, Shinji Watanabe, Jonathan Le Roux, Francesco Nesta, Marco Matassoni |
| LDC Catalog No.: | LDC2017S10 |
| ISBN: | 1-58563-801-3 |
| ISLRN: | 071-714-384-459-0 |
| DOI: | https://doi.org/10.35111/cxwc-kb75 |
| Release Date: | June 15, 2017 |
| Member Year(s): | 2017 |
| DCMI Type(s): | Sound |
| Sample Type: | pcm |
| Sample Rate: | 16000 |
| Data Source(s): | microphone speech |
| Project(s): | CHiME |
| Application(s): | speech recognition |
| Language(s): | English |
| Language ID(s): | eng |
| License(s): |
LDC User Agreement for Non-Members |
| Online Documentation: | LDC2017S10 Documents |
| Licensing Instructions: | Subscription & Standard Members, and Non-Members |
| Citation: | Vincent, Emmanuel, et al. CHiME2 WSJ0 LDC2017S10. Web Download. Philadelphia: Linguistic Data Consortium, 2017. |
| Related Works: | View |
Introduction
CHiME2 WSJ0 was developed as part of The 2nd CHiME Speech Separation and Recognition Challenge and contains approximately 166 hours of English speech from a noisy living room environment. The CHiME Challenges focus on distant-microphone automatic speech recognition (ASR) in real-world environments.
CHiME2 WSJ0 reflects the medium vocabulary track of the CHiME2 Challenge. The target utterances were taken from CSR-I (WSJ0) Complete (LDC93S6A), specifically, the 5,000 word subset of read speech from Wall Street Journal news text.
LDC also released CHiME2 Grid (LDC2017S07) and CHiME3 (LDC2017S24).
Data
Data is divided into training, development and test sets. All data is provided as 16 bit WAV files sampled at 16 kHz. The noisy utterances are in isolated form and in embedded form. The latter involves five seconds of background noise before and after the utterance. Seven hours of noise background not part of the training set are also included.
Also included are baseline scoring, decoding and retraining tools based on Cambridge University' s tool, HTK (the Hidden Markov Toolkit) and related recipes. These tools include three baseline speaker-independent recognition systems trained on clean, reverberated and noisy data, respectively, and a number of scripts.
Samples
Please listen to the following samples:
Updates
None at this time.