Skip to main content
European Commission logo print header

Multilingual Speech Input-Output: Assessment, Methodology and Standardisation

Obiettivo

The objective of the project was to provide a pan-European basis for the assessment of speech technology devices.
The objective of the project was to provide a pan-European basis for the assessment of speech technology devices.

It involves recogniser and synthesiser assessment in each of the participating countries, and is founded on the use and systematic archiving of available databases of words and sounds for general use within the Community. The collation of material coming from the United Kingdom, France, Netherlands, Denmark, Italy and North America was undertaken as a first step. Within this definition phase, tentative uniform protocols were established at the technical level for media and recording conditions. Work was directed towards the establishment of a set of strictly practical tools for everyday use in laboratories and industrial settings in the Community, and also aimed to establish the basis for future techniques, for both speech recognition and speech synthesis. EUROM, a first multilingual database of sounds for language independent assessment of codes, was developed on compact disc read only memory (CD-ROM), and the standard workstation and complete hardware architecture for the speech input output assessment defined.
The completed definition phase was allied to project 64, SPIN. The full project now runs under the Information Processing System and Software (IPSS) area of ESPRIT as SAM (project 2589). It involves recogniser and synthesiser assessment in each of the participating countries, and is founded on the use and systematic archiving of available databases of words and sounds for general use within the Community. The collation of material coming from the UK, France, Holland, Denmark, Italy and North America was undertaken as a first step.
Within this definition phase, tentative uniform protocols were established at the technical level for media and recording conditions. Work was directed towards the establishment of a set of strictly practical tools for everyday use in laboratories and industrial settings in the Community, and also aimed to establish the basis for future development in the field of language-independent assessment techniques, for both speech recognition and speech synthesis.
Exploitation
EUROM, a first multilingual database of sounds for language-independent assessment of codes, was developed on CD-ROM, and the standard workstation and complete hardware architecture for the speech input-output assessment defined.

Argomento(i)

Data not available

Invito a presentare proposte

Data not available

Meccanismo di finanziamento

Data not available

Coordinatore

Birkbeck College, University of London
Contributo UE
Nessun dato
Indirizzo
Malet Street, Bloomsbury
WC1E 7HX London
Regno Unito

Mostra sulla mappa

Costo totale
Nessun dato

Partecipanti (4)