The goal of the thesis was to change reduce the noise produced from a small cavity with grazing sub sonic flow, via alterations in geometry. Research was done into how noise was produced in cavities and existing research in the field. While most work had been done in super sonic flow, little had been done for small cavities at slow, automobile level speeds. Despite the small amount of work done, several features of geometry were noted as affecting the production of sound in cavities. Several of these were chosen to be tested. A modular test rig was designed to allow easy change between the various geometries being tested. It had a flat top surface (of which the cavity came off) to keep the experiment generic, while the bottom surface was a combination of two slopes to keep disturbance to the flow at a minimum and reduce the background noise. A new working section, contraction and anechoic termination were designed and produced to add onto the existing wind tunnel in the mechanical engineering wing at the University of Canterbury. The noise production from a cavity is dependent upon the interaction between the vortices travelling in a 'shear layer' over the cavity throat interact with the front and rear edge of the cavity throat. By minimising the impact to the vortices as they travel between cavity edges, significant noise attenuation can occur. The most successful strategies were lowering the rear edge relative to the front edge and putting ramps on the front and rear edges. To test the principle of superposition, these two strategies, along with an 'L' plate (the third best strategy) were put together to make a 'Super' cavity. the noise spectrum from this cavity was almost identical to the background noise. Removal of the 'L' plate improved sound attenuation even more. Work to do in the future includes testing more examples of each geometry modification, as time constraints minimised the variations on each modification tested. Additionally, the super cavity should be tested in a curved, aerofoil/roof rack extrusion, to see if it as effective in a curved environment as a flat one. All experimental work was done at 100km/h, to coincide with the open road speed limit in New Zealand. The tests should be redone at different speeds, so a more complete picture of the cavity noise can be produced.