Monophonic sound source separation with an unsupervised network of spiking neurones