Repository logo
 

Research data supporting "Cambridge university transcription systems for the multi-genre broadcast challenge"


No Thumbnail Available

Type

Dataset

Change log

Authors

Woodland, Philip C.  ORCID logo  https://orcid.org/0000-0001-9069-0225
Liu, Xunying 
Qian, Yanmin 
Zhang, Chao 
Gales, Mark J.F. 

Description

Details of audio data availability. Detailed speech recognition output and scoring results for all the submitted systems and components for the transcription systems for the MGB challenge.

Version

Software / Usage instructions

Text .ctm files are the standard form for speech recognition outputs as used by standard NIST scoring software, The output of the scoring process is given in .sys files.

Keywords

Speech Recognition, Transcription, Broadcast data, MGB challenge

Publisher

University of Cambridge
Sponsorship
This work was supported by the EPSRC [grant number EP/I031022/1], Natural Speech Technology programme grant, Cambridge Commonwealth and the European & International Trust.
Relationships
Supplements: