=========================================================================
ASR and Closed Captioning from LDC

LDC produced ASR using the BBN Audio Indexer System, version 1.3RC7.
We provide the files as-is. The output for 3 of the Arabic files is
missing.

  tv6.asr.arb.tar.gz
  tv6.asr.chn.tar.gz
  tv6.asr.eng.tar.gz

LDC has also provided closed captioning for the English sources

  tv6.cc.eng.tar.gz

=========================================================================
Arabic and Chinese ASR/MT

  tv6.asr.mt.arabic.gz
  tv6.asr.mt.chinese.gz

Each file contains the sentence-level translations of the Arabic or
Chinese video data - one sentence per line. Each line has the
following data (in order from left to right):

1) file name
2) start time offset from start of file in seconds
3) stop time offset from start of file in seconds

BBN indicates the outputs "reflect the current status of the automatic
speech recognition and machine translations."

TRECVID is very greatful to Bing Xiang along with Ralph Weischedel and
John Makhoul at BBN for contributing this system output.

In presentation and papers you can cite the following publication as
background on the system which did the ASR and MT:

	TO BE DETERMINED AND ANNOUNCED