Speech intelligibility plays a key role in determining the quality of verbal communication. It depends on the acoustic characteristics of the room and the signal-to-noise ratio (SNR). Background noise strongly shapes the intelligibility, so being aware of its components is fundamental. In fact, HVAC noise, anthropic noise and external activities noise contribute to background noise at different frequencies. In this work it is shown how it is possible to distinguish the various noise sources which contribute to background noise applying statistical techniques to sound level meter measurements taken during lectures. The same techniques are used to characterize the speech level, which is the signal in the SNR. Organizing the data collected with a sound level meter, an asymmetrical distribution is built and studied. Various techniques permit the association of each sound source to the corresponding sound level: percentile levels, Gaussian mixture model based on peak detection, Gaussian mixture model and k-means clustering. In the present study these techniques are applied to several university lessons in order to identify different sound signals, in particular the received speech level and the student activity. Results are compared and discussed.