Searching for audio samples within a library can be a tedious and time-consuming task. In this paper, we report on the design of a pilot automatic classification system that utilises timbral properties to automatically classify audio samples. At this stage of the study, we have decided to work only with orchestral audio samples. In addition, we conducted a perceptual experiment to evaluate the performance of the system across five timbral attributes: breathiness, brightness, dullness, roughness and warmth. Promising classification results indicate that this approach may be suitable for further work that could also benefit some music production tasks.

Publication Date


Publication Title

Proceedings of the 2nd AES Workshop on Intelligent Music Production

Organisational Unit

School of Art, Design and Architecture