SPANCAT

NAME

spaned - speech editing controlled by annotations

SYNOPSIS

spaned (-w overlapwidth) (-c) control-file outfile

DESCRIPTION

spaned is a program to concatenate a number of speech segments from different files into a single speech signal controlled by times or annotations listed in a control file. The control file consists of one line per speech section to process giving (i) the filename, (ii) the items to copy (separated by '/'), (iii) the annotation item, (iv) the start annotation or time, (v) the end annotation or time. If the item numbers are not given, the last appropriate item in the file is chosen. If no start time or annotation is given then it defaults to the start of the item. If no end time or annotation is given it defaults to the end of the annotated region if a start annotation is supplied, or otherwise to the end of the item. Items should be separated by commas. To prevent clicks signals are slightly overlapped and weighted by a raised cosine window, Speech items must match in sampling rate.

OPTIONS

-I Identify program name and version number.

-w overlap-width Specify time for overlaps at joins. Default 0.025s.

-c Check mode. Access all the files and annotations but don't do the copying.

INPUT ITEMS

SP Speech items

AN Annotation items

OUTPUT ITEMS

SP Concatenated Speech.

HISTORY

file=control Control file name.

VERSION/AUTHOR

1.2 Mark Huckvale


Fri Jul 09 14:54:43 2004