TDF Format for LDC Transcripts

The TDF (tab delimited format) used by LDC is a simple file format in which
data is represented as a set of records, which are in turn a set of fields
separated by tab characters. Records may have different types as with the
RTTM format [1]. This means two things:
  1) two different records may have different number of fields, and
  2) two fields, each belonging to a different record, may have different
  meanings, e.g date vs. name.

The TDF format for LDC transcripts is a set of 13-field records plus
some meta-information. The 13-field record is also called segment,
and the fields are listed below.

         1 file             file name or ID     unicode
         2 channel          audio channel       int
         3 start            start time          float
         4 end              end time            float
         5 speaker          speaker name or ID  unicode
         6 speakerType      speaker type        unicode
         7 speakerDialect   speaker dialect     unicode
         8 transcript       transcript          unicode
         9 section          section ID          int
        10 turn             turn ID             int
        11 segment          segment ID          int
        12 sectionType      section type        unicode
        13 suType           SU type             unicode

where SU stands for "sentence unit". Depending on the genre of data, some
fields may not contain useful information and may even be blank. In addition
to the body of segments, there are a few lines of meta-information.  The
first line of the format declares the above field specification for segments
in the following form.

        file;unicode    channel;int     start;float     ...

The second and third lines specifies where the "real" section boundaries are
and what types they are. For example,

        ;;MM sectionTypes       [u'report', u'nontrans', None]
        ;;MM sectionBoundaries  [0.0, 425.3, 586.3, 9999999.0]

means that the first section starts at 0.0 seconds and its type is "report"
and the second section starts at 425.3 seconds and its type is "nontrans"
and this is the last section (last ones are ignored).

These two lines are optional. In fact, some transcripts such as telephone
speech transcripts do not include this information.

Besides these, comments are allowed anywhere in the file except for the
first line. A comment line starts with ";;" and does not affect the file's
data integrity although lines starting with ";;MM" might have meaningful
information as shown in the example above.

References:

RTTM Format, http://www.nist.gov/speech/tests/rt/rt2003/fall/docs/RTTM-
format-v13.pdf


