man.diff

speech signal process tools

.\" Copyright (c) 1998 Entropic Research Laboratory, Inc. All rights reserved.
.\" @(#)barkspec.1	1.1 7/1/98 ERL
.ds ]W (c) 1998 Entropic Research Laboratory, Inc.
.if n .ds - ---
.if t .ds - \-\h'-0.2m'\-
.TH BARKSPEC 1\-ESPS 7/1/98
.SH NAME

barkspec \- Bark scaled critical band spectrum

.SH SYNOPSIS
.B
barkspec
[
.BI \-a " add_const"
] [
.BI \-m " mult_const"
] [
.BI \-n " num_freqs"
] [
.BI \-r " range"
] [
.BI \-x " debug_level"
] [
.BI \-B " bark_range"
] [
.BI \-H " freq_range"
] [
.BI \-P " param_file"
] [
.BI \-S " spec_type"
] [
.BI \-X
]
.I input.spec
.I output.spec
.SH DESCRIPTION
.PP
This program reads an ESPS spectrum (FEA_SPEC) file containing
power spectra on a linear frequency scale.
To each input spectrum it applies a bank of critical-band filters
with uniform spacing on the Bark scale.
It writes the resulting Bark spectra to an output FEA_SPEC file.
.PP
If
.I input.spec
is ``\-'', standard input is read.
If
.I output.spec
is ``\-'', results are written to standard output.
The input and output should not be the same file;
however, it is okay to run the program as a filter by specifying
``\-'' for both input and output.
.PP
For the input file,
.I freq_format
must be SYM_EDGE (see
.IR FEA_SPEC (5\-ESPS)).
This is the normal output format used by
.IR fft (1\-ESPS)
(for real spectra) and by
.IR me_spec (1\-ESPS).
The output file is in ARB_FIXED format,
meaning that the header contains an explicit list of the frequencies
corresponding to the spectral values in the records.
The output values may be written
either in units of power or log power (dB)\*-see
.B \-S
under Options.
(That is, the output
.I spec_type
may be either PWR or DB\*-see
.IR FEA_SPEC (5\-ESPS).)
In either case a further arbitrary linear scaling of the output values
may be specified\*-see options
.B \-a
and
.BR \-m .
.PP
The computation of the Bark spectrum follows
that presented in ref. [1].
The frequency
.I f
in hertz corresponding to a Bark-scale value
.I b
is given by:
.IP
.if n \{\
f = 600 sinh (b/6)
\}
.if t \{\
.IR f " = 600 sinh (" b "/6)"
\}
.LP
The critical band filtering function
.I F
is determined by the following equation ([1, 2] with a small alteration):
.IP
.if n \{\
10 log\x'1'\d\d10\u\u F(b)  =  7 \- 7.5 (b \- 0.210)
.br
			\- 17.5 [0.196 + (b \- 0.210)\x'-1'\u\u2\d\d] \x'-1'\u\u1/2\d\d
\}
.if t \{\
.EQ
.\" 10~log sub 10~F(b)~=~7~-~7.5^(b~-~0.210)
.\" ~-~17.5^ {[0.196~+~(b~-~0.210) sup 2 ]} sup 1/2
.nr 99 \n(.s
.nr 98 \n(.f
.ps 10
.ft 2
.ds 11 "\f11\fP\f10\fP
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "\f1log\fP
.ds 13 "\f11\fP\f10\fP
.as 12 \v'0.3m'\s-3\*(13\s+3\v'-0.3m'
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "F\|\f1(\fPb\|\f1)\fP
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "\(eq
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "\f17\fP
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "\(mi
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "\f17\fP\f1.\fP\f15\fP
.as 11 "\*(12
.ds 12 "\|
.as 11 "\*(12
.ds 12 "\f1(\fPb
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\|\*(12
.ds 12 "\(mi
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "\f10\fP\f1.\fP\f12\fP\f11\fP\f10\fP\f1)\fP
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "\(mi
.as 11 "\*(12
.ds 12 "\ 
.as 11 "\*(12
.ds 12 "\f11\fP\f17\fP\f1.\fP\f15\fP
.as 11 "\*(12
.ds 12 "\|
.as 11 "\*(12
.ds 12 "\f1[\fP\f10\fP\f1.\fP\f11\fP\f19\fP\f16\fP
.ds 13 "\ 
.as 12 "\*(13
.ds 13 "\(pl
.as 12 "\*(13
.ds 13 "\ 
.as 12 "\*(13
.ds 13 "\f1(\fPb
.as 12 "\*(13
.ds 13 "\ 
.as 12 "\|\*(13
.ds 13 "\(mi
.as 12 "\*(13
.ds 13 "\ 
.as 12 "\*(13
.ds 13 "\f10\fP\f1.\fP\f12\fP\f11\fP\f10\fP\f1)\fP
.ds 14 "\f12\fP
.as 13 \v'-0.65m'\s-3\*(14\s+3\v'0.65m'
.as 12 "\*(13
.ds 13 "\f1]\fP
.as 12 "\*(13
.ds 13 "\f11\fP\(sl\f12\fP
.as 12 \v'-0.7m'\s-3\*(13\s+3\v'0.7m'
.as 11 "\*(12
.ds 11 \x'0'\f2\s10\*(11\s\n(99\f\n(98
.nr 11 \w'\*(11'
.nr MK 0
.if 132>\n(.v .ne 132u
.rn 11 10
\*(10
.ps \n(99
.ft \n(98
.EN
\}
.LP
The function was designed so that
the following three properties hold to good approximation [2]:
(1) The 3 dB bandwith is 1 bark.
(2) The asymptotic slopes are +10 dB/bark below the center frequency
and \-25 dB/bark above.
(3) The peak value is 0 dB and occurs when
.IR b " = 0 bark."
Both instances of the constant 0.210 above
were shown as 0.215 in the original references.
That value and the constant 0.196 were the result
of a numerical search with limited precision [2].
The value 0.210 used here results from an exact calculation,
given the other coefficients.
The \-3 dB points for
.I F
are
.IR b " = \-0.59 bark"
and
.IR b " = 0.41 bark."
.PP
In principle the Bark spectrum is the result of transforming the
original power spectrum from linear frequencies into the Bark domain,
convolving with
.I F,
and sampling the result
at a selected set of uniformly spaced points on the Bark scale,
These points correspond to the peak values of a set of translates of
.I F,
which are used as weighting functions
in computing a set of weighted sums of input spectral values.
In practice [1, 3] it is more convenient to perform equivalent computations
in terms of linear frequencies
.I f.
.PP
The spacing of the output points is often chosen to be 1 bark,
so that the associated 3 dB intervals
exactly cover the region of interest. 
The actual spacing depends on the number of points
and the range to be covered\*-see options
.BR \-B ,
.BR \-H ,
and
.BR \-n .
Note that the Bark-scale values associated with output points are
.I peak
``frequencies'' of weighting functions,
not the centers of the associated 3 dB intervals.
As noted above for
.I F,
the 3 dB intevals extend 0.59 bark to the left of the peak
and 0.41 bark to the right.
.PP
The output field
.I tot_power
(see
.IR FEA_SPEC (5\-ESPS))
is filled in with a copy of the input
.I tot_power.
A special generic header item
.I bark_freqs
is included in the output file header
to record the selected set of uniformly spaced Bark values
.I b.
The equivalent linear frequencies (nonuniformly spaced)
are recorded in the generic header item
.I freqs.
.SH OPTIONS
.PP
The following options are supported.
Default values are shown in brackets.
.TP
.BI \-a " add_const" "\fR [0.0]\fP"
.TP
.BI \-m " mult_const" "\fR [1.0]\fP"
Before being written out, each computed power
.I S
(or log power\*-see option
.BR \-S )
may be transformed into a scaled value
.IR add_const " + " mult_const " * " S .
The default values 0 and 1 for
.I add_const
and
.I mult_const
result in no change in the output values.
Option
.B \-a
overrides any value specified for
.I add_const
in the parameter file.
Option
.B \-m
overrides any value specified for
.I mult_const
in the parameter file.
.TP
.BI \-n " num_freqs" "\fR [(see text)]"
The number of equally spaced Bark values
at which spectral values will be computed.
The default is the width in barks of the specified range (see
.BR \-B ),
rounded to the nearest integer.
Specifying an argument of 0 implies this default value.
This option overrides any value specified for
.I num_freqs
in the parameter file.
.TP
.BI \-r " start" : "last" "\fR [1:(last in file)]"
.TP
.BI \-r " start" :+ "incr"
.TP
.BI \-r " start"
The range of input records to be processed.
In the first form, a pair of unsigned integers gives the numbers
of the first and last records of the range.
(Counting starts with 1 for the first record in the file.)
Either
.I start
or
.I last
may be omitted; then the default value is used:
1 for
.I start
and the last record in the  file for
.I last.
If
.IR last " = " start " + " incr,
the second form (with the plus sign) specifies the same range as the first.
The third form (omitting the colon) specifies a single record.
This option overrides any values of
.I start
and
.I nan
in the parameter file.
The implied value of
.I nan
is 1 +
.IR last " \- " start
(first form), 1 +
.I incr
(second form), or 1 (third form).
.TP
.BI \-x " debug_level" "\fR [0]\fP"
A positive value of
.I debug_level
calls for debugging output,
which is printed on the standard error output.
Larger values result in more output.
For the default value of 0, no messages are printed.
.IP
Output at level 2 includes the same frequency table that option
.B \-X
generates.
.TP
.BI \-B " bark_low" : "bark_high" "\fR [0:(Nyquist equivalent)]\fP"
.TP
.BI \-B " bark_low" :+ "width"
The Bark-scale range to be covered.
The first form specifies the range by giving the upper and lower limits
as a pair of real numbers.
The second form (with the plus sign)
specifies the range by its lower limit and width.
Thus, if
.IR bark_high " = " bark_low " + " width,
the two forms specify the same range.
The uniformly spaced Bark values
(at which spectral values will be computed)
are chosen so that the corresponding 3 dB intervals (1 bark wide)
span the specified range.
The default is to attempt to cover the range of Bark values
corresponding to the entire range of frequencies in the input file.
A specified value of 0 for
.I bark_high
implies use of the default value\*-the Bark-scale value corresponding
to the Nyquist rate.
The uniform interval between Bark values is equal to
.RI ( bark_high " \- " bark_low " \- 1)\c
.RI /( num_freqs " \- 1),"
where
.I num_freqs
is the number of such values (see option
.BR \-n ).
The first such value is
.IR bark_low " + 0.59."
(Recall that 0.59 bark is the distance
from the lower \-3 dB point of the filter to the peak.)
The uniform interval is required to be positive if
.IR num_freqs " > 1,"
and in any case
.IR bark_high " \- " bark_low
must be at least 1 bark.
This option overrides any value of
.I bark_low,
.I band_low,
.I bark_high,
or
.I band_high