SOCRATES Thematic Network in Phonetics and Speech Communication

Inventory of Internet Resources

Speech, Signal and Language Processing Tools

The Speech Filing System (SFS) is a large collection of inter-working tools for speech analysis, processing and display. It used both for research and in our teaching laboratory at UCL.

URL: http://www.phon.ucl.ac.uk/resource/sfs.htm

Tools for: Signal analysis (tools and methods). Signal manipulation (tools and methods). Speech analysis. Speech synthesis.

Platform Compatibility: IBM-PC. Unix. Approximate RAM: 8Mb. Approximate disk: 20Mb. Access by: FTP.

Czech Text-To-Speech system for Windows. As user application can read files, selected text in any application (via clipboard) and as MCI driver. Application interface: DDE communication. Text to Speech Engine (W95,NT only) form in progress.

URL: http://www.anet.cz/frog

Language: Czech.

Tools for: Speech synthesis.

Platform Compatibility: IBM-PC. Approximate RAM: 2Mb. Approximate disk: 2Mb. Special requirements: Sound Card. Access by: e-mail, demo on web.

WinSAL is a speech analysis system for Windows 95. It allows to load, segment and display multiple audio files in the wav format. Functions include oscillogram, spectrogram, Energy, Pitch, FFT, LPC, Cepstrum, Autocotrrelation, AMDF. Data can be saved to text files, screen hardcopies are possible. WinSAL-V additionally allows to analyze video files in the AVI format.

URL: http://www.media-enterprise.de

Language: English or German.

Tools for: Signal analysis (tools and methods).

Platform Compatibility: IBM-PC. Approximate RAM: 8Mb. Approximate disk: 2Mb. Special requirements: sound card recommended. Access by: WWW download.

The Internet is rich in resources for the study of speech sounds. Here are some that I have collected and used. You should be able to use some of them at least online if your terminal can "view" .au sound files (or RIFF .wavs--see below).

URL: http://weber.u.washington.edu/~dillon/PhonResources.html

Tools for: Speech analysis. Speech synthesis. Speech recognition.

Platform Compatibility: IBM-PC. Apple Mac. Unix. Approximate RAM: 16Mb. Approximate disk: 16Mb. Access by: Remote execution. FTP.

Software for speech signal analysis, syntactic analysis, formalisms, generation, knowledge representation, applications.

URL: http://cl-www.dfki.uni-sb.de/cl/registry/draft.html

Tools for: Linguistic parsing/interpretation. Speech recognition.

Platform Compatibility: IBM-PC. Apple Mac. Unix. Access by: FTP.

On-line and downloadable verion of a large lexicon.

URL: http://www.cogsci.princeton.edu/~wn/

Tools for: Lexicon.

Platform Compatibility: Unix. Approximate RAM: 0Mb. Access by: Remote execution. FTP. Browser or downloadable versions.

The intelligent speech analyser is a software tool running on Macintosh computers. It allows a large number of types of display of speech signals including waveform, FFT, cepstrum, LPC, Auditory spectrum, long-term average spectrum, F0, Jitter, Shimmer, etc. It is unclear as to the cost from the home page.

URL: http://www.sci.fi/~pitchsys

Tools for: Signal analysis (tools and methods). Signal manipulation (tools and methods).

Platform Compatibility: Apple Mac. Access by: Contact author.

This is a list of speech tools produced during the ESPRIT 2589 "SAM" Project. It consists mainly of software tools running on a standard SESAM workstation (PC-DOS based) Labelling , Data collection , Signal edition , Processing , DBMS , Testing - Scoring , Recognition

URL: http://www.icp.grenet.fr/Relator/tools/spbacktools.html

Tools for: Signal analysis (tools and methods). Signal manipulation (tools and methods). Speech recognition. Data collection.

Platform Compatibility: IBM-PC. Access by:

SUMMER INSTITUTE of LINGUISTICS: Computing Resources, including: publications, bibliographies, research projects, training, tools, fonts.

URL: http://www.sil.org/computing/sil_computing.html

Tools for: Signal analysis (tools and methods). Signal manipulation (tools and methods). Speech analysis. Linguistic parsing/interpretation. Data collection.

Platform Compatibility: Access by: FTP.

The site has interactive demos in American English, Mandarin Chinese, and German, and the capability to adjust voice parameters on the fly (English only). Pre-synthesized demos for Mandarin Chinese, French, Italian, Russian, Spanish, and Romanian are also provided.

URL: http://www.bell-labs.com/project/tts/

Tools for: Speech synthesis.

Platform Compatibility: IBM-PC. Apple Mac. Unix. Special requirements: none. Access by: Remote execution.

E m a l y s e : Visualisation - Analysis - Segmentation T a i l o r : Editing - Post-processing - Normalisation M u l t i C V : Format conversion - data export The E m a l y s e is well equipped with features to show all information from the recorded data. It makes it easy and secure to places markers and segments with high precision at the desired positions. The modifications on the data are only done by the T a i l o r program. It does the general data post-processing and is also able to read the Emalyse marker- and segment information to do the required data manipulation. The M u l t i C V changes the format from the data but does not change the data itself. It makes the data readable for general data base and calculation programs as well as graphic presentation programs.

URL: http://www.articulograph.de/demo_p.htm

Tools for: Signal analysis (tools and methods). Speech analysis.

Platform Compatibility: IBM-PC. Access by:

An articulatory speech synthesizer has been developed at our phonetics departement. The input of the program is a phonetic symbol string and its output is an acoustic speech signal. The program can be devided into three subsystems. The first subsystem calculates a movement pattern for the model articulators (lips, tongue, velum, glottis) on the basis of the symbol string (broad phonetic transcription). For this calculation a segmental or a gestural production model can be used. The second subsystem is the articulator model. It calculates the geometry of the vocal tract (the area function) for each time instant. The third subsystem ist the aerodynamic-acoustic model. It calculates the acoustic speech signal.

URL: http://www.rrz.uni-koeln.de/phil-fak/phonetik/synthese/index_e.html

Tools for: Speech synthesis.

Platform Compatibility: IBM-PC. Access by: Remote execution.

COMPOST is a complete environment for text-to-speech developpers. It offers: - A rule-compiler where declarative (grammars) and and procedural (external calls to C code) knowledge can be combined. The standard library offers a Morphological Analysis, a N-class Syntactic filter, a TD-PSOLA synthesizer... - COMPOST is built as a client-server model: a compiled scenario for any language may be loaded by the server and used by clients to convert text into speech - Scenarii have been developped for French, Catalan and Castillan

URL: http://www.icp.grenet.fr/cgi-bin/synthese

Language: .

Tools for: Speech synthesis.

Platform Compatibility: IBM-PC. Unix. Approximate RAM: 2Mb. Approximate disk: 2Mb. Special requirements: Sound Card. Access by:

Presentation of a text-to-audiovisual-speech synthesizer whose front-end is a parametric model of the face.

URL: http://ophale.icp.grenet.fr/2.1.html

Tools for: Speech synthesis.

Platform Compatibility: Access by: Remote execution.

Inventory Top Level.

Thematic Network in Phonetics and Speech Communication