shorten.1

语音处理平台 可以分析语音能量 第一振峰频率等数据

.TH SHORTEN 1 "22 Dec 1995"
.SH NAME
shorten \- fast compression for waveform files
.SH SYNOPSIS
.B shorten
[-hlu] [-a #bytes] [-b #samples] [-c #channels] [-d #bytes] [-m #blocks]
[-n #dB] [-p #order] [-q #bits] [-r #bits] [-t filetype] [-v #version]
[waveform-file [shortened-file]]
.LP
.B shorten -x
[-hl] [ -a #bytes] [-d #bytes]  [shortened-file [waveform-file]]
.SH DESCRIPTION
.IX  shorten ""  "\fLshorten\fP \(em compress waveform files"
.IX  files  "shorten command"  files  "\fLshorten\fP \(em compress waveform files"
.LP
.B shorten 
reduces the size of waveform files (such as audio) using Huffman coding
of prediction residuals and optional additional quantisation.  In
lossless mode the amount of compression obtained depends on the nature
of the waveform.  Those composing of low frequencies and low
amplitudes give the best compression, which may be 2:1 or better.
Lossy compression operates by specifying a minimum acceptable
segmental signal to noise ratio or a maximum bit rate.   Lossy
compression operates by zeroing the lower order bits of the waveform,
so retaining waveform shape.
.LP
If both file names are specified then these are used as the input and
output files.  The first file name can be replaced by "-" to read from
standard input and likewise the second filename can be replaced by "-"
to write to standard output.  Under UNIX, if only one file name is
specified, then that name is used for input and the output file name
is generated by adding the suffix ".shn" on compression and removing
the ".shn" suffix on decompression.  In these cases the input file is
removed on completion.  The use of automatic file name generation is
not currently supported under DOS.  If no file names are specified,
shorten reads from standard input and writes to standard output.
Whenever possible, the output file inherits the permissions, owner,
group, access and modification times of the input file.
.SH OPTIONS
.TP
.BI \-a " align bytes"
Specify the number of bytes to be copied verbatim before compression
begins.  This option can be used to preserve fixed length ASCII
headers on waveform files, and may be necessary if the header length
is an odd number of bytes.
.TP
.BI \-b " block size"
Specify the number of samples to be grouped into a block for processing.
Within a block the signal elements are expected to have the same
spectral characteristics.  The default option works well for a large
range of audio files.
.TP
.BI \-c " channels"
Specify the number of independent interwoven channels.  For two signals,
a(t) and b(t) the original data format is assumed to be
a(0),b(0),a(1),b(1)...
.TP
.BI \-d " discard bytes"
Specify the number of bytes to be discarded before compression or
decompression.  This may be used to delete header information from a
file.  Refer to the -a option for storing the header information in the
compressed file.
.TP
.BI \-h
Give a short message specifying usage options.
.TP
.BI \-l
Prints the software license specifying the conditions for the
distribution and usage of this software.
.TP
.BI \-m " blocks"
Specify the number of past blocks to be used to estimate the mean and
power of the signal.  The value of zero disables this prediction and
the mean is assumed to lie in the middle of the range of the relevant
data type (i.e. at zero for signed quantities).   The default value is
non-zero for format versions 2.0 and above.
.TP
.BI \-n " noise level"
Specify the minimum acceptable segmental signal to noise ratio in dB.
The signal power is taken as the variance of the samples in the current
block.  The noise power is the quantisation noise incurred by coding the
current block assuming that samples are uniformally distributed over the
quantisation interval.  The bit rate is dynamically changed to maintain
the desired signal to noise ratio.  The default value represents
lossless coding.
.TP
.BI \-p " prediction order"
Specify the maximum order of the linear predictive filter.  The
default value of zero disables the use of linear prediction and a
polynomial interpolation method is used instead.  The use of the
linear predictive filter generally results in a small improvement in
compression ratio at the expense of execution time.   This is the only
option to use a significant amount of floating point processing during
compression.  Decompression still uses a minimal number of floating
point operations.

Decompression time is normally about twice that of the default
polynomial interpolation.  For version 0 and 1, compression time is
linear in the specified maximum order as all lower values are searched
for the greatest expected compression (the number of bits required to
transmit the prediction residual is monotonically decreasing with
prediction order, but transmitting each filter coefficient requires
about 7 bits).   For version 2 and above, the search is started at
zero order and terminated when the last two prediction orders give a
larger expected bit rate than the minimum found to date.   This is a
reasonable strategy for many real world signals - you may revert back
to the exhaustive algorithm by setting -v1 to check that this works
for your signal type.
.TP
.BI \-q " quantisation level"
Specify the number of low order bits in each sample which can be
discarded (set to zero).  This is useful if these bits carry no
information, for example when the signal is corrupted by noise.
.TP
.BI \-r " bit rate"
Specify the expected maximum number of bits per sample.  The upper bound
on the bit rate is achieved by setting the low order bits of the sample
to zero, hence maximising the segmental signal to noise ratio.
.TP
.BI \-t " file type"
Gives the type of the sound sample file as one of
{ulaw,alaw,s8,u8,s16,u16,s16x,u16x,s16hl,u16hl,s16lh,u16lh}.  ulaw is
the natural file type of ulaw encoded files (such as the default sun .au
files) and alaw is a similar byte-packed scheme.  All the other types
have initial s or u for signed or unsigned data, followed by 8 or 16 as
the number of bits per sample.  No further extension means the data is
in the natural byte order, a trailing x specifies byte swapped data, hl
explicitly states the byte order as high byte followed by low byte and
lh the converse.  The default is s16, meaning signed 16 bit integers in
the natural byte order.

Specific optimisations are applied to ulaw and alaw files.  If lossless
compression is specified with ulaw files then a check is made that the
whole dynamic range is used (useful for files recorded on a SparcStation
with the volume set too high).  Lossless coding of both file types uses
an internal format with a monotonic mapping to linear.  If lossy
compression is specified then the data is internally converted to
linear.  The lossy option "-r4" has been observed to give little
degradation.
.TP
.BI \-u
The ulaw standard (ITU G711) has two codes which both map onto the zero
value on a linear scale.   The "-u" flag maps the negative zero onto the
positive zero and so yields marginally better compression for format
version 2 (the gain is significant for older format versions).
.TP
.BI \-v " version"
Specify the binary format version number of compressed files.   Legal
values are currently 1 and 2, higher numbers generally giving better
compression.   Detection of format version on decode is automatic.
.TP
.BI \-x " extract"
Reconstruct the original file.  All other command line
options except -a and -d are ignored.

.SH METHODOLOGY

shorten works by blocking the signal, making a model of each block in order
to remove temporal redundancy, then Huffman coding the quantised prediction
residual.

.SS Blocking
The signal is read in a block of about 128 or 256 samples, and
converted to integers with expected mean of zero.  Sample-wise-interleaved
data is converted to separate channels, which are assumed independent.

.SS Decorrelation

Four functions are computed, corresponding to the
signal, difference signal, second and third order differences.  The
one with the lowest variance is coded.  The variance is measured by
summing absolute values for speed and to avoid overflow.

.SS Compression

It is assumed the signal has the Laplacian probability
density function of exp(-abs(x)).  There is a computationally
efficient way of mapping this density to Huffman codes, The code is in
two parts, a run of zeros, a bounding one and a fixed number of bits
mantissa.  The number of leading zeros gives the offset from zero.
Signed numbers are stored by calling the function for unsigned numbers
with the sign in the lowest bit.  Some examples for a 2 bit mantissa:
.LP
.RS
.ft B
.nf
100        0
101        1
110        2
111        3
0100       4
0111       7
00100      8
0000100    16
.fi
.ft R
.RE

This Huffman code was first used by Robert Rice, for more details see
the technical report CUED/F-INFENG/TR.156 included with the shorten
distribution as files tr154.tex and tr154.ps.

.SH SEE ALSO
.BR compress (1), pack (1).
.LP
.SH DIAGNOSTICS
.LP
Exit status is normally 0.  A warning is issued if the file is not properly
aligned, i.e. a whole number of records could not be read at the end
of the file.
.SH BUGS
Large values of '-c' or '-b' cause MS-DOS to throw a wobbly.
Presumably this is a (lack of) memory management problem.
.LP
An easy way to test shorten for your system is
to use "make test", if this fails, for whatever reason, please report it.
.LP
No check is made for increasing file size, but valid waveform files
generally achieve some compression.  Even compressing a file of random
bytes (which represents the worst case waveform file) only results in a
small increase in the file length (about 6% for 8 bit data and 3% for 16
bit data).  There is one condition that is know to be problematic, that
is the lossy compression of unsigned data without mean estimation -
large file sizes may result if the mean is far from the middle range
value.  For these files the value of the -m switch should be non-zero,
as it is by default in format version 2.
.LP
There is no provision for different channels containing different data types.
Normally, this is not a restriction, but it does mean that if lossy coding
is selected for the ulaw type, then all channels use lossy coding.
.LP
It would be possible for all options to be channel specific as in the -r
option.   I could do this if anyone has a really good need for it.
.LP
See the file "change.log" for a history of bug fixes.
.LP
Please mail me immediately at the address below if you do find a bug.

.SH AVAILABILITY
The latest version can be obtained by anonymous FTP from
svr-ftp.eng.cam.ac.uk, in directory comp.speech/coding.  The sources
are available for UNIX machines in files shorten.tar.Z and
shorten.tar.gz and for DOS machines as file shorten.zip.  All
distributions contain a DOS executable.

.SH AUTHOR
Copyright (C) 1992-1995 by Tony Robinson and SoftSound Ltd
([e-mail])
.LP
Shorten is available for non-commercial use without fee.   See the LICENSE
file for the formal copying and usage restrictions.