mask.1

speech signal process tools

.\" Copyright (c) 1996 Entropic Research Laboratory, Inc.; All rights reserved
.\" @(#)mask.1	1.3 1/26/96 ERL
.ds ]W (c) 1996 Entropic Research Laboratory, Inc.
.TH  MASK 1\-ESPS 1/26/96
.SH NAME

.nf
mask \- masks/fades out or mixes segments of sampled data file
.fi
.SH SYNOPSIS
.B
mask
[
.BI \-f " mask_field"
][
.BI \-s " subfile"
][
.BI \-x " debug_level"
]
.I maskfile
.I infile
.I outfile
.SH DESCRIPTION
\fIMask\fR is useful for fading or masking out selected segments of an
utterance.  Both the input and output files, \fIinfile\fR and
\fIoutfile\fR are FEA_SD files of SHORT data type.  The file
\fImaskfile\fR is a FEA file with the field \fImask_field\fR that
contains a masker to fade or mask out segments in \fIinfile\fR.  Data
in \fIoutfile\fR is a result of the masker multiplied with the data in
\fIinfile\fR.  The file \fIsubfile\fR is an optional FEA_SD file of
SHORT data type with sampling frequency the same as \fIinfile\fR.  If
\fIsubfile\fR is specified, the "masker" signal should be bounded in
the range of 0 and 1, since it will be used to fade between the infile
and subfile.  Each sampled data point in \fIoutfile\fR is the sum of
the masker multiplied with the data point in \fIinfile\fR and (1 -
masker) multiplied with the data point in \fIsubfile\fR.
.PP
\fIMask\fR is particularly useful for gating out the 
unvoiced segments of an utterance or to substitute unvoiced regions from
one version of an utterance into another utterance.  Use \fIget_f0(1-ESPS)\fR
to obtain voicing estimate in the \fIprob_voice\fR field of its output
file as decisions to mask out the unvoiced regions in \fIinfile\fR.
.PP
If the record frequency in \fIinfile\fR is larger than that in 
\fImaskfile\fR, the masker in \fImaskfile\fR is interpolated
linearly.  
.PP
If the starting times, as indicated by the generic header 
item \fIstart_time\fR,
and ending times are different for \fImaskfile\fR and \fIinfile\fR,
\fIoutfile\fR starts with the later starting time and ends
with the earlier ending time.  The substitution
signal in \fIsubfile\fR must start no later than the the later staring time
of \fIinfile\fR and \fImaskfile\fR, and end no earlier than the earlier 
ending time of these files.
.SH OPTIONS
.PP
The following options are supported:
.TP
.BI \-f " mask_field \fR["prob_voice"]\fP"
The field name in the input FEA file \fImaskfile\fR that contains masker
signal
for masking/fading data segments in \fIinfile\fR or mixing segments in 
\fIinfile\fR
and \fIsubfile\fR.  The default name is \fI"prob_voice"\fR
as the field name for probablity of voicing for file produced by 
\fIget_f0 (1-ESPS)\fR.  This field
is a scalar field of FLOAT data type.  If the \fB-s\fR option is used,
this field should be bounded between 0 and 1; otherwise, a warning message
is sent to \fIstderr\fR, and an undefined result is produced.
.TP
.BI \-s " subfile"
The file name of a FEA_SD file which will be mixed with that of 
\fIinfile\fR in the proportion to the masker specified by the 
\fImask_field\fR field in \fImaskfile\fR.
.TP
.BI \-x " debug_level \fR[0]\fP"
If 
.I debug_level
is positive,
.I mask
prints debugging messages and other information on the standard error
output.  The messages proliferate as the  
.I debug_level
increases.  If \fIdebug_level\fP is 0 (the default), no messages are
printed.  
.SH ESPS PARAMETERS
No ESPS parameter processing is supported
.PP
.SH ESPS COMMON
No ESPS common processing is supported
.PP
.SH ESPS HEADERS
\fIMask\fR writes the usual generic header items \fIrecord_freq\fR and
\fIstart_time\fR in the output file header.
.PP
.SH FUTURE CHANGES
.PP
.SH EXAMPLES
.PP
.SH ERRORS AND DIAGNOSTICS
.PP
.SH BUGS
.PP
None known.
.SH REFERENCES
.PP
.SH "SEE ALSO"
mergefea(1-ESPS)
.PP
.SH AUTHORS
David Talkin, Derek Lin
.PP