Insight into vocal fold and lip oscillation mechanisms is important for the understanding of phonation and the sound generation process in brass musical instruments. In general, a simplified analysis of the physical 3D fluid-structure interaction process between the living tissues and the airflow is favoured by most workers. Several simple models (lumped parameter models) have been proposed and these represent the tissues as a distribution of elastic mass(es). The mass-spring-damper system is acted on by a driving force resulting from the pressure exerted by the airstream. The results from these theoretical models have been validated 'in-vitro' using rigid or deformable replicas mounted in a suitable experimental set-up. Previous research by the authors focused on the prediction of the pressure threshold and oscillation frequency of an 'in-vitro' replica, in the absence and presence of acoustical feedback. In the theoretical model a lip or vocal fold is represented as a simple lumped mass system. The model yielded accurate prediction of the oscillation threshold and frequency. In this paper a new 'in-vitro' set-up is presented, which overcomes some of the limitations of the previous study. By the use of a digital camera synchronised with a light source and of pressure sensors, this set-up allows 1) measurement of the area of the replica opening and 2) imposition of independent initial conditions, such as height of the initial opening and internal pressure in the replica. The impact of these findings on physical modelling is discussed. The model yielded accurate prediction of the oscillation threshold and frequency. In this paper a new 'in-vitro' set-up is presented, which overcomes some of the limitations of the previous study. By the use of a digital camera synchronised with a light source and of pressure sensors, this set-up allows 1) measurement of the area of the replica opening and 2) imposition of independent initial conditions, such as height of the initial opening and internal pressure in the replica. The impact of these findings on physical modelling is discussed.