What is neural network processing?
Artificial Neural Networks are structures simulated on a computer that work much like a network of nerve cells in nature, like in the human brain. Instead of using one single processor to do an enormous amount of number crunching in a very short time, neural networks take their advantage from their massively parallel structure. The way they are connected can store information, and makes it possible to detect patterns in a seemingly chaotic stream of data due to their robustness to changes and small deviations in these patterns. One could say, just as you are able to recognize a person from different viewing perspectives, neural networks can be used to recognize acoustic instruments in all their many incarnations, relatively independently of how they are actually played or recorded.
