Links to third-party software sites

Please note two things in using this page: 1) internet addresses (URL's) and their directory structures frequently change, so there is a risk that some of the links provided below may become out of date. 2) Listing any particular site or piece of software does not constitute a recommendation or endorsement. Except that I do think that these links are worth a mention.

1. Compilers

1.1. DJGPP: a C compiler for MS-DOS

The latest version (v2) of this C compiler, DJGPP, can be found at D. J. Delorie’s very informative web site ( , including a form-based guide to help you determine what zip files to download. The collection of DJGPP files on the CD-ROM was a combination of (1.5 Megabytes, unzips to 4MB), (1.9 MB) and (2.5 MB, unzips to 5.5 MB). (As future modifications to the software are made, higher numbers may be used in these file names instead of 203, 2952 and 2951, in future. In any case, downloading djdev*.zip, gcc*, and bnu*, where * indicates some numbers should be OK.) The documentation files in are also recommended. README.1ST is a short guide to installation. My even shorter guide is as follows:

  1. Create a directory called C:\DJGPP.
  2. Download the zipfiles to this directory.
  3. Unzip the zip files from that directory. (In Windows 95, use an unzip program which supports long file names.) N.B. Preserve the directory structure within each zipfile, but unload the contents of the three zipfiles all together in the DJGPP directory, not into three separate folders (bnu2951b, djdev203, and gcc2952b).
  4. Use an editor to add the following two lines to the file C:\autoexec.bat:
      set PATH=C:\DJGPP\BIN;%PATH%
  5. Restart your computer.
The first time I attempted to run gcc, it gave an error message saying that it could not find one of the files it needed. The problem was because a file had been corrupted either in the archive or on downloading and could not be unzipped properly. It was easily rectified by downloading the relevant zip file again.

1.2. SWI-Prolog

This is a reliable Edinburgh-standard version of Prolog developed at the Psychology Department at the University of Amsterdam, and downloadable from The SWI-Prolog home page provides a good deal of information about its many virtues. For use under Windows, shift-click on the “Win32 binary” link. The downloadable code for version 3.4.4 was about 1MB long. When the file called w32pl344 has downloaded, clicking on its icon will guide you step-by-step through a self-installation process which will create various subdirectories in C:\Program Files\pl and set up file properties so that clicking on files with the suffix “.pl” (for “Prolog”) will execute them using the SWI-Prolog interpreter.
2. Sound file editors and analysis software

2.1. Cool Edit

Cool Edit is a sound file editor originally produced by Syntrillium, which also has some functions for analysing and manipulating signals. The evaluation version of Cool Edit 2000 is one of a suite of different versions. Adobe Systems Incorporated acquired the technology assets of Syntrillium Software in May 2003. On August 18th, 2003 Adobe released a rebranded version of Cool Edit Pro 2.1 as Adobe Audition™ software. See Adobe's Cool Edit site,, for further details.

2.2. WaveSurfer

WaveSurfer is an open source tool for sound visualization and manipulation. WaveSurfer is being developed at the Centre for Speech Technology (CTT) at KTH in Stockholm, Sweden, one of the world's leading centres of speech research. As well as KTH's excellent software, WaveSurfer also contains some of the excellent speech analysis tools that were previously sold by Entropic (ESPS/waves+) before it was acquired by Microsoft, including a pitch tracking algorithm based on the ESPS tool get_f0, and a formant tracking function based on the ESPS tool formant, which are among the best pitch- and formant-tracking tools available. 

2.3. Praat

Praat, like WaveSurfer, is another excellent and versatile package of tools for speech analysis.

3. Finite-state toolkits

3.1. PC-KIMMO: A Two-Level Processor for Morphological Analysis.

Information and downloads are available from SIL.

3.2. Xerox Research Centre Europe's finite-state toolkits and demonstrations.

3.3. AT&T Labs - Research finite-state machine library: it remains to be seen how long these will be available.

4. Other Links

4.1. The comp.speech Frequently Asked Questions WWW site provides a range of information on speech technology, including speech synthesis, speech recognition, speech coding, and related material. There are 250 comp.speech WWW pages and they include over 500 hyperlinks to speech technology web sites, ftp servers, mailing lists, and newsgroups.

4.2. The Hidden Markov Model Toolkit HTK is a portable toolkit for building and manipulating hidden Markov models. HTK is primarily used for speech recognition research, and is in use at hundreds of sites worldwide. The HTK release contains extensive documentation and examples. HTK was originally developed at Cambridge University Engineering Department, but was later sold and developed by Entropic until 1999 when Microsoft bought Entropic. Microsoft has now licensed HTK back to Cambridge University, which still redistributes HTK. In short, while Microsoft retains the copyright to the existing HTK code, you can get access to it from the Cambridge HTK website.