-------------------------------------------------------------------------------
README: PCRT Corpus
-------------------------------------------------------------------------------

Full name:   Proper name Call Routing Telephone (PCRT) Corpus
Description: Short prompts of proper names and language names collected via the 
             telephone network.
Languages:   South African English

Version:     0.1

Size: A total of 2483 prompts with a duration of 2 hours and 7 minutes.
      The corpus consists of 3 sets of prompts:
      (1) 925 proper name prompts obtained from about 40 speakers
          (names_orig, train & test)
      (2) 555 proper name prompts obtained from about 30 speakers (names)
      (3) 1003 language name prompts obtained from about 50 speakers (langs)

-------------------------------------------------------------------------------

See LICENSE.txt for copyright information

-------------------------------------------------------------------------------

When using this corpus, please cite:
  Neil Kleynhans, Raymond Molapo and Febe de Wet, 
  "Acoustic model optimisation for a call routing system", In Proc. Pattern Recognition
  Association of South Africa annual symposium (PRASA), Pretoria,
  South Africa, Nov 2012, pp 165-172.

bibtex:
@inproceedings{kleynhans12callrouting,
  author = {Neil Kleynhans, Raymond Molapo and Febe de Wet},
  title = {Acoustic model optimisation for a call routing system},
  booktitle = {Proc. Pattern Recognition Association of South Africa annual
               symposium (PRASA)},
  pages = {165--172},
  address = {Pretoria, South Africa},
  year = {2012},
  month = {Nov.}
}

-------------------------------------------------------------------------------

See the documentation directory for additional information with regard to the 
collection, transcription and format of this corpus.

------------------------------------------------------------------------------- 
