CHiME2 WSJ0

registration: 2017/06/17 08:25:23, last modified: 2017/06/17 08:47:03
title
  • CHiME2 WSJ0
title.abbreviation
title.alternative
creator
  • researcher Emmanuel Vincent
  • researcher Jon Barker
  • researcher Shinji Watanabe
  • researcher Jonathan Le Roux
  • researcher Francesco Nesta
  • researcher Marco Matassoni
subject
  • speech corpus
subject.linguisticField
  • computational_linguistics
  • phonetics
  • phonology
subject.monoMultilingual
  • monolingual
subject.resourceSubject
  • corpus
description
  • *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.

    *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.
description.price *This metadata is only as a guide
  • $0.00 2017 Member
  • $50.00 Non-Member
  • $0.00 Reduced-License
  • $0.00 Extra Copy
description.language
  • English
description.inputDevice
  • microphone
description.inputEnvironment
  • noisy_condition noisy living room environment
description.speakingStyle
  • read_speech
description.speechMode
  • monologue
description.samplingRate
  • 16 kHz
description.additionalData
publisher
contributor
contributor.attribute.author.motherTongue
contributor.attribute.author.dialect
contributor.attribute.author.level
contributor.attribute.author.age
contributor.attribute.author.gender
contributor.attribute.speaker.motherTongue
contributor.attribute.speaker.dialect
contributor.attribute.speaker.level
contributor.attribute.speaker.age
contributor.attribute.speaker.gender
contributor.attribute.speaker.number
date.created
date.issued
  • 2017-06-15
date.modified
type
  • Sound
type.discourseType
  • report
type.linguisticType
  • primary_text
type.purpose
  • developing_technologies
type.style
  • written
type.form
  • unfixed
type.sentence
  • long
  • short
type.annotation
type.annotationSample
type.sample
format.extent
  • Number of words: 5,000
format.medium
  • Web Download
format.encoding
format.markup
format.functionality
identifier
identifier.doi
identifier.islrn
source
  • Read speech from Wall Street Journal news text (CSR-I (WSJ0) Complete)
language
  • English
language.area
  • America
relation
relation.utilization *This metadata is automatically extracted
coverage.temporal
coverage.spacial
rights