1. Data format

The database mainly contains speech data and impulse response data. The speech data is raw data measured by a microphone and A/ D converter and consists of a 16-bit integer format. Conversely, the impulse response data is yielded by calculation and consists of a 32-bit real number format. Both are raw format without any header and sampled using 48 kHz.

File	Format	Sample
Sound	16 bit integer Raw format (48kHz)	metalic impulsive sounds pan.dat (90kB)
Impulse response	32 bit float Raw format(48kHz)	Impulse response ir383.dat (92kB)

2. Software for displaying and manipulating waveforms

There are various software programs for displaying speech signal data as a waveform on the screen, performing simple editing work, or converting a file format to WAVE format. Among these are some useful software programs, introduced below, for processing speech data in the database by speech and acoustics researchers.

• SPWAVE : http://www.itakura.nuee.nagoya-u.ac.jp/people/banno/spLibs/spwave/
  Function: Display, replay, edit, and transform of sound file
  Distribution: Free software
  Author: BANNO Hideki, Nagoya Univ.
  OS: UNIX, Windows, Macintosh

• SPTK: http://kt-lab.ics.nitech.ac.jp/~tokuda/SPTK/
  Function: Speech Signal Processing Toolkit
  -- speech analysis, speech synthesis, display and edit of sound file
  Distribution: Free software
  Author: Kobayashi Lab. Nagoya Institute of Technology
  OS: UNIX

• GoldWave http://www.goldwave.com/
  Function: Display, replay, edit, and transform of sound file
  Distribution: Shareware
  Author: Chris S. Craig
  OS: Windows

• Cool Edit 2000: http://www.syntrillium.com/cooledit/
  Function: Display, replay, edit, and transform of sound file
  Distribution: Commercial
  Author: Syntrillium Software Corp.
  OS: Windows

• DAP: http://www.cee.hw.ac.uk/~richardk/
  Function: Display, replay, edit, and transform of sound file
  Distribution: Free software
  Author: Richard Kent
  OS: UNIX

3. SPWAVE - software for displaying and manipulating speech data file

SPWAVE software for displaying and manipulating speech data file is in the DVD-ROM of the database, courtesy of BANNO Hideki, the author. Binary versions of Windows 95 and 98, Mac OS and Linux are supplied. However, it is necessary to upadte SPWAVE. It is therefore recommended that you transfer a latest version from SPWAVE web site.

• Microsoft Windows95/98
  • spwave-0.6.0-1.win.zip

• Mac OS (Power PC)
  • spwave-0.6.0-1.sea.hqx

• Linux RPM version (for libc6 system)
  • spwave-0.6.0-1.i386.rpm (GTK version)
  • spwave-xm-0.6.0-1.i386.rpm (Motif version)

    The program works properly on Vine 2.0 and RedHat 7. A GTK version requires a shared library of GTK+ version 1.2.X. In addition, to work this binary file, the following plug-in package has to be installed.

  • spPlugin-0.8.1-3.i386.rpm

Usage

Only the basic operation method of SPWAVE is described herewith. For details, refer to the User's Guide presented on the SPWAVE web site.

• Start-up
  In Windows and Mac OS, the program is launched by double clicking on the icon. Start-up by command line is carried out as follows.

        spwave [options...] [filename...]
        

  For displaying the raw format of the 16-bit integer (speech) and 32-bit real number (impulse response) processed in the database, the program can be started by attaching the following options.

        $ spwave -b 16 -f 48000 filename ……… 48kHz, 16bit, integer
        $ spwave -b 33 -f 48000 filename ……… 48kHz, 32bit, float
        

• Reading the file
  Alternating to specifying the file by command line, the file can be read by clicking "Open" on the "File" menu.
• Exiting
  The file can be closed by selecting "Exit" on the "File" menu.
• Display range
  Display range can be changed by selecting "Enlarge" or "Reduce" on the "Display" menu or the scroll bar located at the bottom of the screen. In addition, after specifying the range on the screen by dragging, only the selected range can be displayed in an enlarged size by "Enlarge specified range" on the "Display" menu.

Screen shot of SPWAVE