Abstract: Softcomputing (fuzzy logic and artificial neural networks) have been widely applied in several fields, above all control and pattern matching. With the fast development and the huge availability of extremely pervasive communication technologies such as Internet, new challenges are prompted. Due to large availability of multimedia data (audio, video, images, etc.), searching information is becoming an increasingly complex task because most of information is not available in text format. Audio information is widely spread in multimedia information and it is strictly related to video and image information.
Audio classification is the first step in the developing of complete process that leads to upgrading current text-based search engine with signal-based information such as audio (sounds, music, and speech). Fuzzy logic and artificial neural networks fit optimally the classification problem of the audio information, due to the fuzzy and the neural nature of recognition of specific audio pattern in complex audio contexts (broadcast news, video, TV programs, advertising, etc.).
Non linear nature of audio perception, audio pattern recognition, and audio information extraction from a mix of unknown sources (unmixing) have a perfect matching with fuzzy logic inference and with neural classification.
Both fuzzy logic and neural networks will be discussed in three main audio processing areas, word spotting, speaker recognition, and music pattern recognition Audio features extraction algorithms are firstly explained, then the modelling of a fuzzy inference engine from feature distribution and the training of an artificial neural network for pattern classification are discussed.

Brief Biography of the Speaker:
M. Malcangi graduated in Computer Engineering from the Politecnico di Milano in 1981. His research is in the areas of speech processing and digital audio processing. He teaches Digital Signal Processing and Digital Audio Processing at the Universita degli Studi di Milano. He has published several papers on topics in digital audio and speech processing. His current research efforts focus primarily on applying soft-computing methodologies (neural networks and fuzzy logic) to speech synthesis, speech recognition, and speaker identification, where deeply embedded systems are the platform that supports the application processing.