.TH POCKETSPHINX_BATCH 1 "2007-08-27"
.SH NAME
pocketsphinx_batch \- Run speech recognition in batch mode
.SH SYNOPSIS
.B pocketsphinx_batch
.RI \fB\-hmm\fR
\fIhmmdir\fR
\fB\-dict\fR
\fIdictfile\fR
[\fI options \fR]...
.SH DESCRIPTION
.PP
Run speech recognition over a list of utterances in batchmode.  A list
of arguments follows:
.TP
.B \-adcdev
name for audio input (platform-specific)
.TP
.B \-adchdr
Size of audio file header in bytes (headers are ignored)
.TP
.B \-adcin
Input is raw audio data
.TP
.B \-agc
Automatic gain control for c0 ('max', 'emax', 'noise', or 'none')
.TP
.B \-agcthresh
Initial threshold for automatic gain control
.TP
.B \-allphone
Do phoneme recognition
.TP
.B \-alpha
Preemphasis parameter
.TP
.B \-backtrace
Print back trace of recognition results
.TP
.B \-beam
Beam width applied to every frame in Viterbi search (smaller values mean wider beam)
.TP
.B \-bestpath
Run bestpath (Dijkstra) search over word lattice (3rd pass)
.TP
.B \-bestpathlw
Language model probability weight for bestpath search
.TP
.B \-cachesen
Cache senone scores from first pass search
.TP
.B \-cep2spec
Input is cepstral files, output is log spectral files
.TP
.B \-cepdir
files directory (prefixed to filespecs in control file)
.TP
.B \-cepext
Input files extension (prefixed to filespecs in control file)
.TP
.B \-ceplen
Number of components in the input feature vector
.TP
.B \-cmn
Cepstral mean normalization scheme ('current', 'prior', or 'none')
.TP
.B \-cmninit
Initial values (comma-separated) for cepstral mean when 'prior' is used
.TP
.B \-compallsen
Compute all senone scores in every frame (can be faster when there are many senones)
.TP
.B \-ctl
file listing utterances to be processed
.TP
.B \-ctlcount
No. of utterances to be processed (after skipping \fB\-ctloffset\fR entries)
.TP
.B \-ctlincr
Do every Nth line in the control file
.TP
.B \-ctloffset
No. of utterances at the beginning of \fB\-ctl\fR file to be skipped
.TP
.B \-dict
pronunciation dictionary (lexicon) input file
.TP
.B \-dither
Add 1/2-bit noise
.TP
.B \-doublebw
Use double bandwidth filters (same center freq)
.TP
.B \-dsratio
Frame GMM computation downsampling ratio
.TP
.B \-fbtype
FB Type of mel_scale or log_linear
.TP
.B \-fdict
word pronunciation dictionary input file
.TP
.B \-feat
Feature stream type, depends on the acoustic model
.TP
.B \-fillpen
Filler word transition penalty
.TP
.B \-frate
Frame rate
.TP
.B \-fsg
state grammar
.TP
.B \-fsgbfs
Force backtrace from FSG final state
.TP
.B \-fsgctlfn
finite state grammar control file
.TP
.B \-fsgusealtpron
Use alternative pronunciations for FSG
.TP
.B \-fsgusefiller
(FSG Mode (Mode 2) only) Insert filler words at each state.
.TP
.B \-fwd3g
Use trigrams in first pass search
.TP
.B \-fwdflat
Run forward flat-lexicon search over word lattice (2nd pass)
.TP
.B \-fwdflatbeam
Beam width applied to every frame in second-pass flat search
.TP
.B \-fwdflatefwid
Minimum number of end frames for a word to be searched in fwdflat search
.TP
.B \-fwdflatlw
Language model probability weight for flat lexicon (2nd pass) decoding
.TP
.B \-fwdflatsfwin
Window of frames in lattice to search for successor words in fwdflat search 
.TP
.B \-fwdflatwbeam
Beam width applied to word exits in second-pass flat search
.TP
.B \-fwdtree
Run forward lexicon-tree search (1st pass)
.TP
.B \-hmm
containing acoustic model files.
.TP
.B \-hyp
output file name
.TP
.B \-hypseg
output with segmentation file name
.TP
.B \-input_endian
Endianness of input data, big or little, ignored if NIST or MS Wav
.TP
.B \-kdmaxbbi
Maximum number of Gaussians per leaf node in kd-Trees
.TP
.B \-kdmaxdepth
Maximum depth of kd-Trees to use
.TP
.B \-kdtree
file for Gaussian selection
.TP
.B \-latsize
Lattice size
.TP
.B \-lifter
Length of sin-curve for liftering, or 0 for no liftering.
.TP
.B \-live
Get input from audio hardware
.TP
.B \-lm
trigram language model input file
.TP
.B \-lmctl
a set of language model
.PP
The
.B \-hmm
and
.B \-dict
arguments are always required.  Either
.B \-lm
or
.B \-fsg
is required, depending on whether you are using a statistical language
model or a finite-state grammar.  To do batchmode recognition, you
will need to specify a control file, using
.B \-ctl
This is a simple text file containing one entry per line.  Each entry
is the name of an input file relative to the
.B \-cepdir
directory, and without the filename extension (which is given in the
.B \-cepext
argument).
.PP
If you are using acoustic feature files as input (see
.BR sphinx_fe (1)
for information on how to generate these), you can also specify a subpart
of a file, using the following format:
.PP
.RS
.B FILENAME START\-FRAME END\-FRAME UTTERANCE-ID
.RE
.SH AUTHOR
Written by numerous people at CMU from 1994 onwards.  This manual page
by David Huggins-Daines <dhuggins@cs.cmu.edu>
.SH COPYRIGHT
Copyright \(co 1994-2007 Carnegie Mellon University.  See the file
\fICOPYING\fR included with this package for more information.
.br
.SH "SEE ALSO"
.BR pocketsphinx_continuous (1),
.BR sphinx_fe (1).
.br