Float space refers to a method for storing and manipulating data whose dynamic range exceeds what can be represented on typical computer monitor.
In a typical graphics system, pixel brightness is measured in the range from 0 to 1, zero being the darkest, 1 the brightest.
In float space, pixels can have brightness values far beyond this range, both negative and positive. While these pixel values will be clipped when displayed–negative pixel values will be clipped to black, while pixel values greater than one will be clipped to white–nevertheless the greater dynamic range is vital when performing color corrections that bring those pixel values back into visible range.
These extreme values are essential when working with specialized CG mattes, such as depth and surface normal passes.
Here, an image of a light bulb has pixels around the filament with values thousands of times brighter than 1. When the image is darkened, those pixels are clearly visible.
Float space is also essential for preserving image data for display or cinema projectors, whose dynamic range is significantly broader than typical computer workstation monitors.
OpenEXR has emerged as a standard file format for float-based data
Subscribe for the latest resources to stay up to speed with the hottest trends