Alwin de Rooij | Portfolio

Artificial neural networks are a valuable and practical tool for digital analysis and classification of sound. They are used in a variety of subfields in digital signal processing such as speech recognition, music information retrieval, nonlinear filters, sound synthesis, sound localization and more. The main advantage of using neural networks is that they can provide real world sub optimal solutions for problems that are otherwise hard to solve.

Audio Neurons is a software environment that enables the user to experiment with sound and multilayered feedforward neural networks and is supplemented with the backpropagation algorithm as well as several statistical and visualization functions. The environment proves to be especially powerful in analysis and classification. Examples are pitch recognition, instrument classification, voice classification and experimental other usage.