Home of Professional Cinematography since 1996
Soft Sources & Light Fall Off
>I'm doing a very large prelight without the advantage of having my gear package yet and only limited time to do tests, and I'm trying to figure out how softening sources affects falloff. Empirically, I know that soft sources do not throw as far, units like lekos, dedos, and spotty pars throw farther than fresnels; but does anyone know the relationship between diffusion and the inverse square law? Does any one actually know the mathematics of this? Also, any opinions of the most output efficient diffusions? I am a big fan of 1/2 soft frost for this purpose; any opinions?
>The inverse square law is the same for all light sources. The only exception to this that I know of is a softbox from the Swiss manufacturer, Broncolor, called the Hazylight. The Hazylight is a modified parabolic shape that focuses the lightsource behind its diffuser in such a way that the the light loss over distance is a bit shy of the inverse square of the distance ... note: this is not a big deviation.
>The inverse square law is only effective when dealing with a point source.
>Without previous testing you cannot predict light falloff from a diffuse source by taking one reading and doing the math. I've done much testing, specifically with sources shining through various diffusion materials. I don't have the numbers with me but I'll try to quanitfy the differences for you when I can. I can tell you that the more diffuse, the farther from the inverse square law you get.
>Adam "love my seconic" Sternin DP NY
>On the suject of soft sources and falloff:
>Although it is true that there is no way to "do the math" since all combinations of difusion material, size of the diffusion surface relative to subject, and source of the light are variable factors, one fact remains true:
>Once a diffusion material is introduced into the path of a light source, the surface of the material BECOMES the light source, and the inverse square law is effective from the diffusion surface to the subject.
>Meaning that :
>If you move the diffusion surface closer to the subject you will increase the light intensity by the same proportion as if you moved a direct source closer to the subject.
>This is NOT TRUE if you move the light closer to the diffusion material. You must keep the distance between the light source and the diffusion the same as you move the screen, but moving "the rig" then becomes measurable in terms of the inverse square law.
>Also, keep in mind that, no matter how diffuse a light source is, the further it is from the subject, the more specular it becomes. This is why clipping 216 to the barndoors of a light is much less effective than the same 216 attached to a 4X4 frame six feet away from the light source. Likewise, that 4x4 frame will seem to be a much softer source when it is six feet away from the subject. Move it to ten feet away, and the contrast will increase on the subject until it looks the same as any direct light.
Joe Di Gennaro, SOC
Director of Photography/Camera Operator
Former New Yorker, Recent Californian
>Sorry, folks, diving in after an absence due to travel. As a lapsed physics major, I have to take exception to the inverse square discussions. Point one: Inverse square law describes the fall off of radiation from a point source with respect to distance. The fall off from a line source (a single fluorescent tube, for instance) of infinite length is inversely proportional to the distance - that is to say, if you rent an infinitely long kino tube (or make a row of them across the stage) then the light output as you move away from it will fall off by half every time you double the distance.
>The fall off from a planar source is...well, actually, if the plane is infinite in size, there is NO fall off, because the further away you get from it, the more of it you see (and the more of it sees you, and radiates on you.)
>So who cares?
>If you build a wall of light - shower curtain, muslin, whatever - as long as your subject is relatively close to it - as long as it looks like a BIG plane of light to the subject, the stop will be the same irrespective of where the subject moves. Diffusion frames start off acting like planar sources until you get a distance away from them... and when they get very far away, they act like point sources. Example: An orange that is 1 foot away from a 4x4 of muslin can be moved around without the stop changing at all, whereas the same orange being lit by a 4x4 of muslin that is 40 feet away will see the muslin as being much more like a point source, and the inverse square law will more nearly apply. Parabolic reflectors and fresnel lenses both "cheat" the inverse square law by re-directing light - concentrating it. They have the effect of making a given source "look like" a larger source farther away . By this I mean that the rays being more parallel to each other, there is a bit less fall off than you would have from a point source at the same distance. An extreme case of a bright source that is very far away is the sun. It is so far away, that the distance from one side of the set to the other , or one side of town to the other, is such a small precentage of the overall distance (93 million miles) as to not affect brightness. That is what is so much fun with reflectors - they don't fall off with the square of the distance FROM THE REFLECTOR...they fall off with the square of the distance from the sun plus the distance to the reflector.
>Different types of diffusion act differently. 1000H, Rolux, 216, pretty much occlude the original source and their surface becomes the source. 250, 251, the hampshires, opal, and china silks, lie in different places on the continuum between "source at the distance of the light to subject" and "the diffusion is the source"
>General caveats - all this assumes perfect lenses, perfect reflectors, perfect vacuum, etc etc. Some settling of contents may occur due to shipping and handling. Your results may vary. Tests items for colorfastness before washing.
>Some day I will have time to develop the above ramble into a coherent explanation of light and its behavior on set. Until then, I hope this does not just cloud the issue:-)
Mark H. Weingartner
Lighting and VFX for Motion Pictures
>That was an interesting piece on information but it raises a few questions in my mind. Working under the assumption that the inverse square law does not apply when we talk about soft sources, there should be no difference in terms of fall off if we were lighting a face with a soft source (2ft by 2ft) source 1ft away when compared to lighting it with another source ( same "relative" size, i.e. the radiating source is bigger) 10 ft away.
>Assuming...The only source is the light, there is no other surface/source.
>The light reading, in both the cases, is the same at the point where the face is closest to the light.
>So the only variable, theoretically, is the distance.
>I hope the enquiry is clear and look forward to some en"light"enment.
>I see what you are getting at...a 2ft x 2ft source 1 foot away would exhibit some fall-off as you moved back and forth towards it because it is of finite area.
>In theory the only way to have no fall-off would be if it were much bigger than the 2x2 (at that distance) I would agree that the wrap and look of the light would be similar between the 2x2 source 1 ft away and the larger source farther away that has the same apparent size (20x20ft 10 ft away, for instance) but moving a foot closer or farther to the 2x2 source 1 ft away would change the look and stop more than moving a foot closer or farther to the 20 x 20 10 feet away.
>Even though the difference would not be discibed by the inverse square law, none the less, the proportional difference of a foot's change is very different between the two cases. I would venture to say that you would not read any difference or see a change in the wrap of the light in the case of the 20 x20, but in the case of the 2x2 1 foot away, moving one more foot away would change the wrap of the light significantly (you can draw the light angles out in scale on a piece of paper and see what I mean.) and probably show some difference in exposure on an incident meter
>Right - you'd have to move about 10 feet on the larger source to equal the change in moving 1 foot on the smaller one. It is proportional.
Wade K. Ramsey, DP
Dept. of Cinema & Video Production
Bob Jones University Greenville, SC 29614
>The inverse square law is a geometric progression. There's an easy way to keep track if/when stop changes are significant when moving a light source closer or further from the subject. Everyone here knows how aperture numbers progress:
>1.4 2.0 2.8 4.0 5.6 8.0 11 16 22 32 45 64 90 128
>What some of you may not be aware of is that this progression can be used as a distance scale to determine light level changes. If you have a light that is 1.4 feet away from the subject and move it away to 2 feet away, the light changes by -1 stop. If you move it from 1.4 feet to 2.8 feet away, you loose 2 stops.
>If you move this same light from 11 feet to 16 feet away (5 foot change) you loose one stop. If you move 5 feet closer, from 11 feet to 6 feet, you gain almost 2 stops.
>Obviously (I've now learned), your mileage may vary, depending on how large a light source you are using, relative to its distance to subject.
>This knowledge has helped me enormously in saving time in remetering ... I know if I need to take another reading when I move the position of the light only a small amount. If the light is already close to the subject any change in distance needs to be remetered. If it's "far", then that same change of distance results in a stop changes that may be so minor as to be insignificant.
>Dominic pointed out :
>>Though there's probably another 1600 cml'ers saying "do we need this? ..There are >many paths to the kingdom. I know people who have a total scholarly grasp of all >this math stuff and how it can be applied...who can't seem to make pretty pictures.
>I also know people who don't get any of the math - or care - who, nevertheless, make beautiful pictures.
>Some people have an intuitive grasp of this stuff as a result of empirical data - messing around with lights until things look good.
>My gut feeling is that knowing some of the math behind what we do makes it easier to get what you want faster, and I encourage people to get their feet wet in it, since it can save so much time...but whatever works for you is what you should do.
Mark H. Weingartner
Lighting and VFX for Motion Pictures
>The inverse square law is a geometric progression.
> 1.4 2.0 2.8 4.0 5.6 8.0 11 16 22 32 45 64 90 128
>Thank you Cliff and CML for this discussion. There I was last week, teaching a class on sensitometry and exposure. (and logarithms (arrrggghh!)). I explained geometric series. I sensed a feeling of "do we need this?". You have boosted my confidence that it's a good thing to know about geometric series.
>Though there's probably another 1600 cml'ers saying "do we need this? let's get back to digital cinema and Robert Frost and the important things in life."
Group Technology & Services Manager