One of the most challenging issues in neuroethology concerns the neural substrates for song production, perception and learning in songbirds. However, electrophysiological studies of the song system of songbirds are often fragmented and centered on a small number of selected neurons. Here, we have developed a new extensive approach to record a great number of single units in the brain of a songbird, the European starling. The aim of this approach is to formulate quantitative assumptions about the electrophysiological characteristics of the brain nucleus investigated: e.g. the proportion of auditory neurons, neuronal selectivity, etc. We applied a mapping method using multisite recording electrodes and online isolation of single-units, without preselecting neurons with a search stimulus. As an example of the application of this technique, we have mapped the responses to a variety of natural and artificial acoustic stimuli recorded systematically throughout the HVC of both awake and anaesthetized male starlings. This method appears to be powerful and allowed us to quantitatively compare responses obtained in awake and anaesthetized birds by recording over 1000 single units. We think that, in future, this will enable us to characterize and compare parts of nuclei or entire nuclei and to better understand how the song system works.