A series of tests have been put into effect in a laboratory to improve the cavitation monitoring techniques. Simultaneous vibrations and dynamic pressures have been measured in a reduced scale Francis turbine model with different types of cavitation: weak and intermittent cavitation at the blade outlet, outlet cavitation combined with intermittent von Karman cavitation, strong bubble cavitation and pulsating inlet cavitation. The structure and the fluid-borne noise induced by cavitation have been analyzed in the time and the frequency domains. Initially, the low and high frequency signal contents have been compared for the various types of cavitation. Then, the main hydrodynamic frequencies that modulate in amplitude high frequency bands have been identified. From the analysis of the results several conclusions can be extracted concerning the most suitable sensor, position and signal processing for detection of each type of cavitation. Besides, a detailed analysis of the processed data permits to infer some of their particular hydrodynamic characteristics that can be extrapolated to the real case for reliable identification.