|
..::pseudo hdri::..
With a standard image, we have a colour range of 0 to 255 for all RGB settings. Therefore, when a true white colour is reflected in an object that has less than perfect reflection, these bright white surfaces will, in turn, be reduced as the software is simply classing it as a standard colour. To get these bright highlight areas (such as sunlight through a window, or an interior light) we have to intensify these areas of illumination beyond the default white colour so that when they are reflected they are still very bright, generating the desired reflected highlight.
A high dynamic range image (HDRI) generates the (aesthetically) correct result as we can go beyond the normal white value, but is a bit of a time consuming process; shooting multiple plates of the same environment using a reflective chrome sphere takes set up time and can lead to problems, especially on a busy set with multiple people moving around, resulting in a large amount of painting to get each exposure image right so they all line up. In addition to this, shooting exterior shots is difficult as objects also tend to move, such as clouds, foliage etc. Even though HDRI is one of the latest buzzwords, it is often not necessary to use it as we can just tweak the standard, or “Low Dynamic Range” image (as it has come to be known). As the reduced intensity of the white in the standard reflection needs to be brighter to stand out, we can amend the “LDR” image’s output curves by clamping off the values below the bright white end of the curve and then cranking up the white end, so that the white values are multiplied several times to simulate a bright surface of illumination, resulting in a bright reflection in these parts that stands out where most of the other reflected environment areas are not as visible.
This generates a quick fix solution and is suitable for most reflective uses (especially with environments with contrast), however it is not suitable for images with areas of similar brightness such as a lot of whites and yellows as the output may not be able to differentiate between them (eg someone wearing a white t-shirt next to a light!)
 |
We are using two images in this scene – the main environment background image and an unwrapped chrome sphere image which is placed in the raytrace reflection’s environment map slot. |
 |
To bring out the bright areas of the images, they have been dropped into Output maps (the map’s own outputs could be tweaked if desired), the curve clamped off and the top end value increased by a factor of 20, resulting in a bright highlight reflection. |
 |
The difference between the two versions of the same scene is quite dramatic – the left part is with the untouched reflection & environment, with the right version having the amended output curves. |
 |
Zip file
to accompany. |
..::quick tip::..
As we are using values over RGB255,255,255 white, we can enable exposure to get a more natural result. Also, try overlaying multiple outputs to simulate multiple “exposures” using the Composite map.
Initially
published: 3D
World magazine, Issue 67, August 2005.
Copyright
© Pete
Draper, August 2005. Reproduction without permission
prohibited.
www.xenomorphic.co.uk |