High Dynamic Range Imaging

A Short History of HDRI

Let’s move on to HDRI itself. As with any photographic technique or technical advancement HDRI has a history and an interesting one at that.

The Beginnings of HDRI

According to Wikipedia the first recognized attempt to overcome the issue of low DR in photography was done by Gustave Le Gray in the 1850’s. LeGray developed a technique we now call ‘combination printing’ in which he used separate negatives of the same scene taken with different exposures to produce a positive that contained more dynamic range than a single exposure could produce. Other photographers picked up and used this technique.

Probably the most well-known of the photographers using this technique was Henry Peach Robinson. He produced several top selling images using combination printing, two of which are below. In the first he shot six separate negatives to capture the dynamic range of the scene. In the second he shot and combined five negatives.

HDRI in The 20th Century

In the first half of the 20th century (1939-1940) Ansel Adams and Fred Archer developed the ‘Zone System’ which quickly became an important advancement in photographic techniques. “The Zone System is a photographic technique for determining optimal film exposure and development [based on a scale of shades ranging from 0 (black) to IX (white)].” Wikipedia

The first thing you need to know is that exposure meters are dumb, really dumb. They are clueless about what they’re photographing. They assume that all scenes have the same average tonal value –middle gray– roughly Zone 5 in the chart below. If a scene is different, for example a snow scene, it will be exposed incorrectly. It will come out middle gray: underexposed. If you follow the meter’s reading, every scene will have the same average middle gray density. You’ll get lots of bad exposures, especially if you photograph in difficult light. Ansel Adams developed the zone system to cope with this situation. His technique was to carefully study a scene, visualize the final print, then determine the correspondence between portions of the scene and tones in the print. He would then meter, expose and develop the negative accordingly. His basic rule was, ‘Expose for the shadows; develop for the highlights’.” The Luminous Landscape

A little over a decade after the Zone System appeared a very different situation requiring high dynamic range in an image arose: accurately recording what happens inside of a nuclear detonation in its first few seconds. In the late 1940’s and early 1950’s high speed photography was available but the massive DR of a nuclear fireball overwhelmed every photographic film of the time. A solution was needed, and quickly.

In the early 1950’s Charles Wyckoff developed a new type of photographic film that consisted of three layers of photosensitive emulsion each having a different level of sensitivity to light (aka. film speed), he called it ‘XR Film’. With it the very high dynamic range of a nuclear explosion could accurately be captured via high speed photography.

Below is a scan of a two page layout that appeared in Life magazine on December 23, 1966 showing a sequence of images of a nuclear explosion captured on Wyckoff’s XR film.

(Because of the quality of the scan I have redone the numbers for each image and have transcribed the text of the short article/caption that appeared below the image on the printed page. Also, the scan does not contain the colors that the author refers to in the text, but they can be inferred by the contrast in each frame.)

“These pictures of a nuclear blast were produced by an unusual type of film; nuclear explosions vary enormously in brightness, from the radiance of a thousand suns to the dimness of a luminous watch dial, a range no conventional film can record. Changes in the brightness indicate changes in temperature. To get accurate data on brightness and thus temperature during nuclear tests, Charles Wyckoff, of a research company called EG&G, Inc. devised a special film called “XR” for extended range. XR Film is comprised of three panchromatic layers: an extremely slow layer (color-coded cyan blue) to record the brightest phenomena; a medium-speed layer (magenta) for intermediate brightness; and a high-speed layer (yellow) for faint phenomena. These photographs show an explosion conducted by the AEC before the 1963 test ban treaty. The searingly hot 300,000 C fireball, recorded on the film’s slow layer at the start of the sequence, takes on-in the false-color code-a deep-bluish cast against a blood-red sky (1). As the fire-ball expands, it first (2) dims to red as it is obscured by opaque gases and then darkens again. In 5, expanding shock waves encounter moisture-laden layers of atmosphere creating rings of condensation which spread, rise and multiply. By 11, they are several miles in diameter. When the rising fireball reaches a cooler altitude, moisture freezes (14), capping the fireball with a cascade of ice. Finally, about a minute after detonation, nothing registers on the XR film but a yellow wisp (20) of radioactive cloud.”

HDRI in The Age Of Digital Images

XR Film remained the only form or HDRI medium, as primitive as it was, until the 1980’s when computing power began to grow exponentially. I began my life with computer hardware in 1987 and have witnessed this explosion of CPU power ever since.

In 1985 Gregory Ward created the first digital HDR image format which he called “Radiance RGBE image format” (Wikipedia and the “Radiance Synthetic Imaging System” site). Later his image file format was designated with the .hdr file extension. From this point HDRI grew quickly beginning with more and more techniques for ‘tone-mapping’ an HDR image.

HDRI as we commonly know it today in still photography consists of capturing multiple digital exposures of a scene each at a different shutter speed (thus each has a different overall exposure) and then combining them into a single image with much higher DR than any of them alone could contain. This process was first proposed by Steve Mann and Rosalind Picard in a 1995 paper. Mann went on to patent the process in 1998.

The history of HDR as we use it today starts in the digital age. Completely unnoticed by most of us, it had its first boom around 1997. In academic circles HDR was whispered about from door to door, traded as the new cool thing to write hot algorithms for. It was a revolutionary new way to deal with image data, and it was said to be your best chance to show your old professor that all he told you about digital images was bogus. In best renaissance tradition, students started discussing the nature of light again and how to represent it digitally. Along the way, this research had a dramatic impact on simulating light behavior and expressing it in render algorithms. Christian Bloch – “The HDRI Handbook 2.0”