Abstract

The crisis affecting film production models and its paradigm shift towards digital seems to also imply an esthetic revolution. In this quantitative research we aim to explore the interrelation between exposure as an esthetic value and the implementation of digital internegative and capture in order to determine quantitative criteria relative to the adoption of these technologies. In order to do this, one frame every fourteen seconds was extracted out of 264 feature films produced in Hollywood, and their average light levels and low-light range levels were quantitatively analyzed. With this information, we aimed to relate different image-capture and image-treatment methods to their implementation in the Hollywood film-making industry and consider low-light reproduction a crucial factor.

Abstract

It is well known from the literature that road lighting has significant safety benefits. The NZTA Economic Evaluation Manual (EEM) quotes a 35% reduction in crashes as the effect of upgrading or improving lighting where lighting is poor.However, no well-established dose–response relationship to lighting parameters exists from which one can deduce benchmark levels of lighting for safety.This study looked at a sample of street lighting installations spread over the urban areas of nine territorial local authorities. Standard street lighting parameters were measured in the field using a variety of instruments including illuminance meter, luminance meter and digital camera. Field measurements were related to the ratio of night-time to day time crashes as a measure of night time safety vis-a-vis daytime safety.A statistically significant dose–response relationship was found between average road luminance and safety across all traffic volume groups, with an indication that the relationship may be stronger where more serious crashes are involved.Threshold increment was also a significant variable but not so longitudinal uniformity or overall uniformity.The results related to luminance will allow practitioners to better estimate the safety benefits of different levels of lighting resulting in better targeting of expenditure. Document type: Article

Abstract

Dynamic area and volume lights are a common phenomenon in our world. Examples include illuminated advertising, video screens, and lava lamps. Yet rendering such light sources in real-time in interactive scenes remains challenging. In this paper we present a method to render area or volume light sources in a fully dynamic scene, i.e. the geometry, materials, camera, and lights (spatial and spectral composition) can change continuously. The key idea of our approach is to decouple lighting from shading complexity by using a dynamically generated hierarchy of irradiance volumes as an intermediate approximation of incident lighting. The main contribution of this paper is a new computation scheme that updates irradiance volumes directly on the GPU and enables real-time performance. The algorithm is very fast, runs on DX9-level hardware, and is easy to integrate into existing rendering pipelines, which makes it well-suited for games and virtual-reality applications. Current limitations of our algorithm are that it cannot handle shadows and provides only a coarse approximation for glossy reflection.