How far away should you stand from a green screen to kill your shadow?

You cannot solve a shadow with distance, because distance makes it worse. A shadow lands on the screen at a lateral offset of <code>d x tan(theta)</code>, where theta is how far your key light sits off the camera axis — so at 45 degrees it moves from 0.30 m off-axis when you stand 0.3 m back to 2.00 m off-axis at 2.0 m, walking out from behind you and across the frame. What distance does buy you is less spill: moving from 1.0 m to 2.0 m cuts the screen light falling on your waist to 45.7% and to 20.3% at 3.0 m. So the advice is half right and half backwards. Deal with the shadow using the angle of your key light, not the distance to the screen, and deal with spill with distance and with flagging. One more thing the numbers show: the spill that lands on the bottom of your subject mostly comes from the green floor section, which contributes 42% of the screen light at your waist — darkening that section does more than any amount of stepping back.

The advice is half right

'Stand further from the screen' is one of the most repeated pieces of green screen advice, and it bundles two separate problems that pull in opposite directions.

Distance genuinely reduces spill. The screen occupies less of the hemisphere around you as you back away, and the light it throws onto you falls off accordingly.

Distance does not hide your shadow. It moves your shadow further from your body, which is the opposite of hiding it. If you have been stepping back to get rid of a shadow, you have been making it more visible.

Below is a model, not a measurement: a 3.0 x 2.5 m Lambertian screen with a 1.5 m floor section, a subject standing in front of it, and scalar irradiance integrated numerically over the screen. All spill figures are given relative to the value at 1.0 m, so nothing here assumes how bright your lights are — you can put your own measured number in.

Spill: what distance actually buys you

Screen light falling on the subject, relative to standing 1.0 m away:

0.3 m — waist 1.434, head 1.562, feet 1.156
0.5 m — waist 1.308, head 1.378, feet 1.111
1.0 m — 1.000, 1.000, 1.000 (reference)
1.5 m — waist 0.704, head 0.719, feet 0.574
2.0 m — waist 0.457, head 0.508, feet 0.168
3.0 m — waist 0.203, head 0.257, feet 0.084
4.0 m — waist 0.110, head 0.145, feet 0.052

The first half metre is the most valuable one: going from 0.5 m to 1.0 m already removes about a quarter of the screen light. Beyond that it flattens out — 2.0 m to 4.0 m only takes you from 0.457 to 0.110, and that costs you a lot of floor space.

One detail worth noticing: the head falls off slowest. At 4.0 m the head is still at 0.145 against 0.052 at the feet, nearly three times as much. If you back up far enough, spill stops being a bottom-of-frame problem and becomes a hair-and-shoulders problem.

The spill on your legs comes from the floor, not the wall

Here is the part that surprised us, and it changes what you should fix.

Measure the same subject with only the vertical screen, no floor section, and the feet come out lowest — 0.646 at 0.5 m, 0.688 at 1.0 m, 0.795 at 2.0 m, against 1.000 at the waist. A tall vertical screen lights the middle of your body most, because that is where the screen is symmetrically above and below you.

Add the green floor section that every real studio lays down, and the ordering inverts:

0.5 m — feet 1.334, waist 1.000, head 0.855
1.0 m — feet 1.569, waist 1.000, head 0.811
2.0 m — feet 0.577, waist 1.000, head 0.902

That floor section contributes 32.1% of the screen light at your waist when you stand 0.5 m back, 42.0% at 1.0 m, and 41.7% at 2.0 m. It is not a minor contributor, it is roughly half of the problem.

So the phenomenon where green appears on the lower half of your subject first is not caused by the wall behind you at all. It is caused by the green floor under you. The fix is correspondingly specific: lay something dark over the floor section, or light it separately and dimmer. That is far cheaper than quadrupling your floor space by stepping back. (The 2.0 m figure dropping to 0.577 is just the model — at that distance you are standing past the end of the 1.5 m floor section.)

Shadows: distance is the wrong lever

A shadow falls on the screen at a lateral offset of d x tan(theta) from your centre line, where theta is the angle between your key light and the camera axis. Once that offset exceeds about half your shoulder width — call it 0.25 m — the shadow is no longer behind you and it becomes visible.

Offset in metres, by key light angle and distance:

15 degrees — 0.08 at 0.3 m, 0.13 at 0.5 m, 0.27 at 1.0 m, 0.54 at 2.0 m
30 degrees — 0.17, 0.29, 0.58, 1.15
45 degrees — 0.30, 0.50, 1.00, 2.00
60 degrees — 0.52, 0.87, 1.73, 3.46

Against the 0.25 m threshold, here is what is hidden and what is not:

15 degrees — hidden at 0.3 m and 0.5 m, visible at 1.0 m and beyond
30 degrees — hidden only at 0.3 m, visible from 0.5 m
45 degrees and 60 degrees — visible at every distance

Every extra metre of distance multiplies the offset by the same factor, no matter how far back you already are. There is no distance at which a 45-degree key light stops throwing a visible shadow.

And the shadow gets bigger and softer as you go

Distance also changes the character of the shadow, which matters because a large soft shadow is harder to deal with than a small hard one. Magnification is 1 + d/D where D is the lamp-to-subject distance, taken as 3.0 m here, and penumbra width grows with source size x d / D for a 60 cm softbox:

0.3 m — magnification 1.10, penumbra 0.060 m, shoulder shadow 0.55 m wide
1.0 m — 1.33, 0.200 m, 0.67 m
2.0 m — 1.67, 0.400 m, 0.83 m
3.0 m — 2.00, 0.600 m, 1.00 m

At 3.0 m the shadow is twice the width of your shoulders and its edge is 60 cm soft. That does not read as a shadow any more, it reads as an unevenly lit screen — which is a harder problem, because now you are fighting the lighting condition that our piece on uneven screens describes.

So how far should you stand?

Split the two problems and the answer stops being a compromise.

For the shadow, use angle, not distance. Bring the key light closer to the camera axis. At 15 degrees the shadow stays hidden up to about 0.5 m, and if you can go nearer to on-axis it stays hidden further out. Nothing about the subject-to-screen distance helps here.

For spill, use the first metre and then flags. The 0.5 m to 1.0 m step removes about a quarter of the screen light for very little floor space; past 2.0 m you are spending metres for diminishing returns. Beyond that, put flags or black side panels between the screen and your subject — that removes screen light without moving anyone.

Then deal with the floor separately. It is around 40% of the screen light at your waist and the direct cause of spill on your legs. Cover it or dim it, in preference to any other change.

And keep the screen evenly lit regardless. Distance does nothing for evenness, and evenness is the condition most often violated before anyone starts adjusting anything.

Where these numbers stop

This is a geometric model, integrated numerically, not a measurement from a studio.

The screen is treated as a perfectly Lambertian, perfectly uniform rectangle. Real screens have hot spots, seams, wrinkles and falloff, all of which change the distribution — though not the direction of any trend here.

The subject is a set of points at three heights, not a body. It does not occlude itself, and it does not cast the shadow that this article is partly about; the shadow figures come from a separate, purely geometric calculation.

The floor section is 1.5 m long and as wide as the screen, which is one common arrangement and not the only one. A shorter or longer floor section moves the foot numbers accordingly.

And the scalar irradiance model deliberately ignores the orientation of the surface receiving the light, which is right for hair and shoulders facing every which way and wrong for a flat panel aimed at the camera.