Classification of Music Using Neural Net
A neural network has been used to classify program material into groups according to the room acoustics that best suit the recording. The network uses Kohonens feature map followed by learning vector quantization and a real-time system for preprocessing and classification. Although an accuracy of better than 80% was achieved, classification can vary greatly from one case to another.
This paper costs $20 for non-members, $5 for AES members and is free for E-Library subscribers.