This paper is centered on the relationship between intuitive and quantitative approaches for the understanding of luminous ambience in order to formalize criteria for the classification of luminous ambiences.
We measured luminance and illuminance levels on glazed and opaque surfaces in interior spaces and built a synthetic scheme of these levels. We then analyzed and compared these measures and their interpretation to impressions felt by several subjects in the spaces, or to intentions expressed by the architect during the design of the project when available.
This comparison has proved itself very meaningful and shows that many relations between the intuitive approach and quantitative measured data may be established. The second part of the work deals with the definition of criteria to classify ambiences. They allowed us to prepare a computer tool, based on neural nets which will be used to help architects store and browse through a large number of images.