Industrial audio fingerprinting distributed system with CORBA and Web Services

With digital technologies, music content providers face serious challenges to protect their rights. Due to the widespread nature of music sources, it is very difficult to centralize the audio management. Audio fingerprinting allows the identification of audio content regardless of the audio format and without the need of additional metadata. Monitoring the audio being broadcasted by the TV and radio stations of a country requires the design and implementation of a scalable, robust and modular framework. We have chosen CORBA as distributed environment. The whole functionality needs to be decoupled from clients. To do so, Web services have been deployed. The audio identification core uses a Hidden Markov Model-based audio fingerprinting technology. The paper discusses the design and implementation issues of a complete distributing system that automatically monitors audio content, specifically music and commercials.Today, a working prototype of such a system already exists, and is ...

• Type: Conference
• Year: 2004
• Where: ISMIR
• Authors: Jose Pedro, Vadim Tarasov, Eloi Batlle, Enric Guaus, Jaume Masip
• Keywords: Audio Content, ISMIR 2004, Markov Model-based Audio, Music Content Providers