*****************************************************
Speech Signal Processing Toolkit Ver 1.0
July 31, 1998
*****************************************************
The Speech Signal Processing Toolkit (SPTK) was developed and has been
used in the research group of Prof. Satoshi Imai (currently he is with
Chiba Institute of Technology) and Prof. Takao Kobayashi (currently he
is with Department of Information Processing, Tokyo Institute of
Technology) at P&I laboratory, Tokyo Institute of Technology. This
time some tools are chosen and arranged for distribution by Associate
Prof. Keiichi Tokuda (currently he is with Department of Computer
Science, Nagoya Institute of Technology) as an organizer in
cooperation with Takashi Masuko (he is a Research Associate at P&I
laboratory, Tokyo Institute of Technology), Dr. Kazuhito Koishida
(currently he is a Research Fellow of JSPS), and some graduate
students.
The original source codes have been written by many people who took
part in research of the research group. The most original source
codes of this distribution were written by Takao Kobayashi (graph,
data processing, FFT, etc.), Keiichi Tokuda (speech analysis, speech
synthesis, etc.), and Kazuhito Koishida (LSP, vector quantization,
etc.).
This version is accompanied by a Reference Manual (a total of 130
pages written in Japanese). The User's Manual will be attached to the
future release. We are also planing to translate the manuals into
English.
********************
Distribution
********************
The Speech Signal Processing Toolkit is free provided that it is only
for academic use and cannot be used commercially. When you distribute
the Speech Signal Processing Toolkit, distribute all files including
this file without any modification. Modifying the programs is allowed
for personal use. There is no warranty for the Speech Signal
Processing Toolkit.
*******************
Environment
*******************
We expect that all programs can be compiled and can work on most UNIX
workstations. Although we have confirmed that all programs can be
compiled in the following environments:
SunOS, Solaris, HP-UX, Free-BSD, OSF1, Linux
we have not confirmed that all programs can work properly in all these
environments since we are debugging them mainly on SunOS and HP-UX.
Note that some commands need C-shell (/bin/csh) since they are
implemented by C-shell scripts.
*******************
Installation
*******************
0) Enter 'src' directory.
1) Read the comments in "Makefile" and make any necessary changes.
2) To build and install all tools, type:
% make
% make install
The X-window library is not required for compilation of all tools
except a command 'xgr' since only 'xgr' uses the X-window library.
**************
Notice
**************
As the version advances, specifications for the Speech Signal
Processing Toolkit will be changed without notifications.
The distributed tools are a sub-set of the whole toolkit. Therefore,
some commands which appear at EXAMPLE and SEE ALSO in the Reference
Manual may not be included in this version.
******************
Bug Report
******************
Bug reports, comments, questions for the Speech Signal Processing
Toolkit are very welcome. Please send them to
dsp-cmnd@ip.titech.ac.jp
by email. We will reply to any questions as much as possible but we
cannot assure that we will reply to all messages.
Please see
http://kt-lab.ics.nitech.ac.jp/~tokuda/SPTK/
for more information about the Speech Signal Processing Toolkit,
(e.g., Examples for using Speech Signal Processing Toolkit).
*************
Notes
*************
Generic properties of each command are summarized as follows:
i) Data has float-type format, i.e., single-precision floating point
format. This can be changed to double-type format by specifying a
compile option. It is noted that in the current version the "float"
specifications in the help messages do not change by the compile
option.
ii) Data files do not have headers or any structures, i.e., they are
flat row files.
iii) Basically they manipulate data through standard input and
standard output.
iv) To display (error) messages they use standard error output rather
than standard output.
v) They do not request interactive key inputs.
vi) Options are specified on the command line.
***************
Credits
***************
Masatsune Tamura
Toshihiko Kato
Gou Hirabayashi
Naohiro Isshiki
Noboru Miyazaki
