Soft Light/Diffusion

[Robert Mojica]

I was looking at images of The Dark Knight and I found this picture. It gave me a better look at how the scene looked off camera. It looks like one source of light which looks like it has some type of diffusion in front of it. Does anyone know what type of diffusion is used for a scene like this and where I can get it? I've used softboxes, but it still does not look as soft as this. Can anyone help?

[Adam Frisch FSF]

You have to first understand what makes something soft. It's got very little to do with diffusion. Here's two experiments that you can play with in one single exercise:

 

Take a golf ball and stick it on top of a pencil and hold it up just in front of the lens of a 1K, 2K, 5K fresnel. Is it soft or hard? You'll find that the light on the golf ball is softer the closer you are to the lens and it will be at its softest right next to the lens. The shadow of the golf ball cast on a nearby wall will also be diffused or even non existent when the ball is that close to the lens. Now move the golf ball away from the lens and you'll notice that it becomes harder and harder (less wrap-around-y, more defined borders between light and dark) and that also the shadow on the wall of the ball has become sharper. Really close to the wall (when it's furthest away from the source of light) you'll see the most defined shadows on the wall and on the ball.

 

Once this relationship has been fully understood, you will see that hardness, or softness, is all about the relative size of the source to the object. This is key to understanding soft lighting. So the only two variables that define softness is placement of the source and the size of it. Any source can be a hard light, and any source can be a soft light.

 

So why use diffusion at all? Well, by enlarging the source, you're in effect making it slightly softer at the same distance. Also, on direct light sources with no lenses, like a redhead or a blonde, the filament itself emits a smaller more concentrated light within the reflected light, which makes it look harder. By diffusing it you're blending the direct light from the filament with the reflected light from the silver reflector, and this evens out the light (making it bigger) so that it more easily can be played as a soft light at the same distance. Move the light further away, and any diffusion you might have added is now getting progressively more negated. Make sense?

 

So make no mistake - put diffusion on any light, any size, any source and move it far enough back from the object you're lighting, and the effects will be a hard light. It's all about relative size. So in the example you had, the white side light on the Joker in Batman is relatively soft. That's probably a 12x12 butterfly soft box right to the left of frame with an HMI in it. The orange backlight up high in the background might be diffused (or it could just be how it flares in the lens), but at that distance and size, it can never be a soft light no matter how much you diffuse it. The only two reasons you're not seeing a hard shadow in front of the Joker from that source, is that either the soft sidelight is overpowering it (most likely), or the orange backlight is not even aimed at the Joker but rather into the lens.

 

Last truism (advanced bonus :) ) - whenever a source is smaller than the object it's supposed to light, at any distance, it will be impossible to achieve softness (depending on how you define soft). For this gag to work, the source must be bigger than its object (like in the golf ball experiment), at any distance.

[Robert Mojica]

You have to first understand what makes something soft. It's got very little to do with diffusion. Here's two experiments that you can play with in one single exercise:

 

Take a golf ball and stick it on top of a pencil and hold it up just in front of the lens of a 1K, 2K, 5K fresnel. Is it soft or hard? You'll find that the light on the golf ball is softer the closer you are to the lens and it will be at its softest right next to the lens. The shadow of the golf ball cast on a nearby wall will also be diffused or even non existent when the ball is that close to the lens. Now move the golf ball away from the lens and you'll notice that it becomes harder and harder (less wrap-around-y, more defined borders between light and dark) and that also the shadow on the wall of the ball has become sharper. Really close to the wall (when it's furthest away from the source of light) you'll see the most defined shadows on the wall and on the ball.

 

Once this relationship has been fully understood, you will see that hardness, or softness, is all about the relative size of the source to the object. This is key to understanding soft lighting. So the only two variables that define softness is placement of the source and the size of it. Any source can be a hard light, and any source can be a soft light.

 

So why use diffusion at all? Well, by enlarging the source, you're in effect making it slightly softer at the same distance. Also, on direct light sources with no lenses, like a redhead or a blonde, the filament itself emits a smaller more concentrated light within the reflected light, which makes it look harder. By diffusing it you're blending the direct light from the filament with the reflected light from the silver reflector, and this evens out the light (making it bigger) so that it more easily can be played as a soft light at the same distance. Move the light further away, and any diffusion you might have added is now getting progressively more negated. Make sense?

 

So make no mistake - put diffusion on any light, any size, any source and move it far enough back from the object you're lighting, and the effects will be a hard light. It's all about relative size. So in the example you had, the white side light on the Joker in Batman is relatively soft. That's probably a 12x12 butterfly soft box right to the left of frame with an HMI in it. The orange backlight up high in the background might be diffused (or it could just be how it flares in the lens), but at that distance and size, it can never be a soft light no matter how much you diffuse it. The only two reasons you're not seeing a hard shadow in front of the Joker from that source, is that either the soft sidelight is overpowering it (most likely), or the orange backlight is not even aimed at the Joker but rather into the lens.

 

Last truism (advanced bonus :) ) - whenever a source is smaller than the object it's supposed to light, at any distance, it will be impossible to achieve softness (depending on how you define soft). For this gag to work, the source must be bigger than its object (like in the golf ball experiment), at any distance.

.

Wow, that makes a lot of sense. Thank you for helping me understand it better. For some reason I thought that you had to take the light source further away to achieve the "Softer" look. Do you know where I can purchase the type of diffusion you mentioned?

[David Mullen ASC]

The softness of the projected shadows created by the object in the light are determined by the size of the source relative to the subject, as Adam says.

 

In other words, if you light someone with a light through a 4'x4' frame of diffusion and then switched to a 20'x20' frame of diffusion but farther away, if from the perspective of the subject, the 20'x20' frame fills their vision as much as the closer 4'x4' frame of diffusion did, then the amount of softness would be the same because their relative sizes compared to the subject were the same.

 

The difference between the two then is not softness but fall-off. If the diffusion frame is much farther away, then the subject moving a foot closer or farther from the frame is not going to cause much of a change in brightness -- but if the soft light is very close to the subject, then leaning a few inches closer or father away can cause a visible change in brightness. In fact, if you side-light someone with a diffusion frame very close to their face, then the cheek closest to the light is going to be visibly hotter than the opposite cheek.

[John Holland]

Its called the " Inverse Square Law "

[David Mullen ASC]

Well, technically that applies to a point source so I don't use the term to describe the rate of fall-off from a soft source, though the effect is more or less the same.