Research context: As the physical approach of the influence of the facade morphology on the acoustical diffusion in an urban street is an extensive subject for acoustics papers those last ten years, one can note that the perceptual aspect of this interaction has quasi-never been developed. This is why both Cerma and Grecau laboratories intend to proceed into this research topic within the “Bruit et Nuisances Sonores” Research Proposal Challenge, initiated by the Ecology & Sustainable Development Ministry, in collaboration with CSTB and LCPC laboratories. Parametering urban diffusion: This research aims to evaluate the influence of the urban facades morphology on the sound sources perception. Additionally, if this influence remains obvious, the experiment following will allow to define the exact nature of this effect. Therefore, we developed different evaluation procedures through spatial restitution of urban sonic objects, such as the tramway bells signals. This will define the physical characterisation of the propagation space, with identifying the facade morphology as the varying parameter. To do this, the Impulse Response (I.R.) of the street has been acquired into the three dimensions of the geometrical space (xyz), on 6 points distributed into the length of the Vital Carles street in Bordeaux, France. Both recorded through a binaural technique and convoluted by a 4-channel B-Format Impulse Response (I.R.), those bell signals are then auralised through a restitution system that preserves their spatial components.Sound recording and investigation methodology: The auralisation conditions depend on the recording techniques: headphone listening for the bells binaural recording, and immersive conditions in B-format 6.2 (6 channels and 2 subbasses) for the convoluted bell signal. In a very first time, the I.R. physical signal has been auralised by experts (acousticians, musicians, sound engineers), in order to discriminate them, through a free categorisation procedure. This procedure consists into a classification of the sonic corpus elements, without taking into account the number of categories nor the number of elements included in each category. A second task consists into a categorial construction through a free argumentation, leading to a class qualification. The resulting descriptors will then being validated by a second group of listeners, hearing to the same sonic corpus. In a second time, this procedure will be applied to both binaural and convoluted bell signals, through a non expert investigation (mixed students). First results: First results concerning the expert group investigation lead us to conclude on a good sample discrimination, through three emergent categorisation criteria: sound level, spatial feeling/ resonance, source distance. Further experiments will finalise those results with confirming the diffuse reflection effect on urban sound sources perception, through sharpening the previous criteria.