High Dynamic Range (HDR) Photography is the technique of taking photos with higher dynamic range than normal cameras can provide. HDR Tone Mapping is the process of transforming HDR photos into normal dynamic range photos suitable for display and print.
Essential HDR™ (code name Wukong) is a revolutionary new HDR Tone Mapping software. It includes two proprietary tone mapping operators: Fast Tone Balancer™, a "global" tone mapping algorithm, and Detail Revealer™, a "local" tone mapping algorithm. Both tone mapping algorithms provide superior results and lightening fast performance.
The name "Wukong" (meaning "aware of emptiness") comes from the playful monkey king in ancient Chinese mythology. He is extremely intelligent, powerful, versatile, and is known for breaking all the rules in the Heavenly Kingdom.
Imaging Luminary's Fast Tone Balancer tone mapping technology is one of the fastest and most effective global tone mapping operators. It compresses the overall contrast using the most mathematically optimal "curve" and provides immediate visualization of HDR photos.
Imaging Luminary's revolutionary tone mapping technology Detail Revealer™ is based on how the Human Visual System copes with high dynamic range scenes. A sophisticated mathematical approach was introduced to simulate how human perceive the real world. As a result, Detail Revealer successfully reveals local details in the same way as how human eyes observe local details and lighting variations.
Detail Revealer starts with explicit and correct definition of the human visual system when coping with high dynamic range scenes. This is very different from other tone mapping algorithms such as Fattal02[1]. The result is greatly improved image quality and speed in comparison with other approaches that are heuristic and intuitive in nature.
This definition is formulated as a mathematical optimization problem. This mathematical optimization problem is solved through advanced algorithm innovation. In comparison with other popular detail tone mapping technologies used by most mainstream HDRI software in today's market, our Detail Revealer technology has the following advantages:
Detail Revealer's preview result and full size results are almost exactly the same
The preview image is usually a smaller sized version of the full image. The size reduction will cause changes in local detail data. As a result, many local tonemappers will suffer. If you use the same set of parameter values, the results from the preview image and the full sized image are usually dramatically different. This widely reported issue caused a tremendous amount of frustration among photographers. Often times the users spend a lengthy amount of time working on the preview image to fine-tune it to perfection, only to find out the full image tonemapped using the same exact set of parameters appear very different and totally unacceptable. The superior algorithm of Detail Revealer successfully avoided this problem. Detail Revealer is able to maintain a very high level of consistency between preview image result and the full image tonemapped result.
Detail Revealer does not produce "halo" or "glowing edge" artifacts.
Detail Revealer is based on the most natural vision system, your eyes. It first detects and understands the edge of the objects and the background. In this example, sky is treated completely differently than the buildings.
Lighting consistency is maintained without sacrificing details. The result is clean and clear pictures.
In other local tonemappers, to reduce "halo" or other artifacts, you have to reduce the level of local details. The result is usually described as "haze".
Detail Revealer is a brand new algorithm from a totally different approach. The result does not have this issues at all.
Detail Revealer does not render random brightness variations.
Just like the human vision system, Detail Revealer is based on a sophisticated Artificial Intelligence (A.I.) engine that can regoconize the pattern of large bright areas. In this example, the sky is recogonized as one entity; so it is treated consistently regardless how the tree branches separates different areas of the sky.
Fast, and scales well with dual core CPUs. Speed is well above 2 million pixels per second for large images using today's mainstream PCs. You don't have to wait for minutes any more!
References:
[1] Raanan Fattal, Dani Lischinski, Michael Werman. "Gradient Domain High Dynamic Range Compression", ACM SIGGRAPH 2002
