Automatic language identification is a method to identify a language from a speech signal uttered by an unknown speaker. The Automatic Language Identification System (ALiS) would perform an automated categorization of human utterances on the basis of language. ALiS is based upon speech and signal processing techniques that are used to extract the language dependent features from a clip of speech, forming a language print. This language print is then compared against a library of pre-generated language prints using innovative comparison methodologies to relate the speech utterance to the closest language set.

Interoperability
ALiS makes use of Web services to provide cross-platform integration. This enables it to be integrated with almost any platform that provides support for the HTTP and SOAP protocols. Web services exchange XML based data using the SOAP protocol and then send these packets using HTTP to a HTTP server that can understand SOAP. This makes ALIS a perfect black box that will require a clip of speech (a WAV file or stream of data) and will, after processing it, return the most probable language. The following diagram is one situation where a Linux based system may interact with the ALiS web service.

Requirements for ALiS Server
• Windows platform
• 2GB RAM (to accommodate 10 requests)
• Network card
• P4 3.0Ghz or above
• .Net Framework