CLICK ON THE PICTURES TO ZOOM!
Light ON/OFF, hi-resolution orthographic panorama made with Hugin.
For me, there’s not much difference between such a work (the prototype, made of laserprints on paper) and a “traditionnal painting.”
I’m really starting to think it’s just the same thing, like Richter includes amateur photography in the field of “painting.”
The prototype is finally finished!
So, to explain in a few words: 1-the amount of light projected by high-power spots on the wall has been measured. 2-an algorithm has been created to absorb the light using black over white laser-prints until the wall’s surface emits the same amount of light to the eyes of an observer situated in front of it. 3-The central part of this print has been cut and the same process remade a second time within the limits of this cut.
The impression produced is impossible to reproduce in photography, so the pictures there are just “documentation.” The impression is a bit like if the light was actually coming from “behind” the wall, like if the rectangle was a “window,” opening the wall, like in a dark corridor…
I’m really happy with it, I think it would be great to remake this much bigger the next time (and after developping the tools for it, I could make it much faster and more precisely…) I’m thinking of a “virtual corridor.” I’ve always been attracted by the idea of “remaking” the corridors of italian museums, where you walk slowly and stop at every painting on your sides. I could now make the painting of a corridor of paintings… :)
PLEASE CLICK ON THE PICTURES TO ZOOM AND SEE BETTER!
NORMAL LIGHT
TRANSFORMED (EQUALIZED) LIGHT
The central part being equalized.
Something is slightly wrong – the values are increasing where the light is intense -, but it’s still very close to what I wanted (an equalized luminance everywhere in the rectangle).
CLICK ON THE PICTURE TO ZOOM IT!
The “pool” in the middle is due to the limit of absorbance of the black (the contrast is too high between the dakest and the brightest part of the wall.)
NOTE: this picture is 10 times smaller than the original (the blog doesn’t support extra large pictures…)
to download the real Hi-res. picture, click here: http://adrienlucca.files.wordpress.com/2012/07/overview1.png
The bad news: the laser printer cannot handle large black prints, some unit of the printer makes a random glossy layer on it.
the random gloss
the (very) good news id that the luminance correction is not very angle-dependant, means that except for that random gloss, the correction works at a wide variety of viewpoints, that’s really nice.
from the left side, the gloss dissappears and the flat is still “flat.”
below: zooms on different parts from the left side (same shutter speed)
CLICK ON THE PIC TO ZOOM!
Here it is: a “flattened” wall, made of a white wall, 2 halogen lights & 1 spotlight, white paper and black toner ink.
The numbers, L* values in photoshop color picker, are approximate and dont take into account vignetting of the camera, but basically it works! And the random-pixel pattern has no border effect, nice :)
CLICK ON THE PIC TO ZOOM!
CLICK ON THE PIC TO ZOOM!
Above is the calculated density for the black relative to the local luminance of the wall.
come back soon! ;)
Integrating sphere at the Laboratorium voor Lichttechnologie, KAHO Sint-Lieven, Gent, BE
Thanks to Richard Vanraes from KAHO, I was able to measure the distortion of my digital camera in a lab. We used the integrating sphere to get measurements of the response of the entire system CCD sensor-lens with full zoom and fixed 1:5.6 aperture.
I double checked one picture of inside the sphere with a picture made of several pictures, each fragment measuring a part of the CCD response to the light coming from the window of the integrating sphere, that was illuminated with an halogen lamp. We had to wait until the couple lamp-sphere becomes stable: the temperature of the lamp was warming up the sphere and the amount of light passing by the window was slowly growing… for approx. 30 min
Now I have to figure out how to use this to correct my measurements… but that’s a good start :)
multi-samples CCD response measurement
picture of the vignetting at max zoom & 1:5.6 aperture
The vignetting in the 1st horizontal line of the picture (in blue the raw data and in red after a gaussian blur)
Adrien Lucca & Robert Ochshorn, Aalst, 30th June 2012.
The 3 squares on the left were exctracted from the picture. The light is almost perfectly equalized. My Minolta Luminance-Meter gives values 15-17 cd/m² all over the covered surface of the wall.
Robert has to go to Nepal, I’ll continue alone next week.
Thanks Rob’
CLICK ON THIS PICTURE TO ZOOM IT!
Thanks to the measurement work and to a script made by Robert O., we were able to generate this first approximation of the wall luminance correction. This is a laserprint (printed in ABC Europe, Brussels) @ 400 DPI, 90 x 230 cm.
3 kinds of patterns were generated: lines, grids (crossed white lines,) and random squares.
We’ll check tomorrow if that works with the proper lighting.
:)
Simulation of the light coming from a punctual source (imagine a microscopic star) and diffused by a surface that is a “perfect diffusor.” The lamp is at a distance of 1 unit of the center of the surface that measures 6 x 6 units. The luminance for black pixels is set at 5%, and at 95% for white ones. The local luminance decreases as the distance increases following the inverse square root law.
Made under python thanks to Robert Ochshorn.
