A light source has an apparent size and a direction. Brightness alone does not determine shadow character.
By the end: Predict how changing a source or adding fill changes form and shadow transitions.
A point-like source casts a relatively abrupt shadow boundary. An extended source illuminates an edge from several directions, creating a partial-shadow region called a penumbra. Its apparent angular size from the subject matters: a large source far away can behave more like a small one.
Cloud cover and a nearby broad window can produce soft transitions. Soft does not mean dim, and hard does not mean bright. A small intense lamp and a small weak lamp can have similarly abrupt shadow edges. Diffusion can enlarge the effective source, but only if the illuminated diffuser is large enough as seen from the subject.
Draw a flat luminous strip, an opaque object, and a receiving surface behind it. Every point on the strip can illuminate the surface unless the object blocks that path. Where the whole strip is visible, illumination is unblocked. Where none is visible, the strip contributes no direct light. Between those regions, only part of the strip is visible. That partial visibility makes the penumbra. Room light can brighten even the full shadow.
Use a side-view construction with a 40 cm tall strip, an object 100 cm beyond the source plane, and a screen 20 cm beyond the object. Assume parallel source and screen planes, straight rays, and an opaque edge long enough sideways that its ends do not matter. Rays from the strip's top and bottom just grazing the same object edge separate on the screen. The source height is 40 cm across a 100 cm run. Across another 20 cm, the separation becomes 40 times 20 divided by 100, or 8 cm.
That 8 cm is the transition width for one straight edge in this two-dimensional model. Move the screen to just 5 cm behind the object while keeping the source and object fixed. The width becomes 40 times 5 divided by 100, or 2 cm. A broad source can therefore cast a relatively crisp contact shadow where an object nearly touches a surface, with a softer shadow farther away. Source size alone is insufficient unless the distances between source, object, and receiving surface are also specified.
Now restore the 20 cm gap and replace the strip with a 10 cm source at the same position. The transition becomes 10 times 20 divided by 100, or 2 cm again. This second way of making a narrow transition changes the illumination directions rather than the object-to-screen gap. Increasing the lamp's power without enlarging its emitting area scales the illumination but leaves these boundary rays unchanged. A brighter small source can make a shadow more visible without making its geometric transition wider or narrower.
A diffuser helps only through the area that actually sends light toward the subject. If a tiny bright spot illuminates the middle of a large translucent sheet and the rest contributes almost nothing, the effective source remains concentrated. If the sheet glows across a broad area, the subject receives rays from a wider range of directions. The edge brightness distribution also matters. A uniformly luminous strip and one brightest in the center can share outer boundary rays while producing different gradual changes within the transition region.
Read the drawing as a cross-section, not a photographic prediction of shadow tone. It omits source texture, multiple reflections, curved surfaces, and wavelength effects. A real circular source changes visible area differently from a uniform strip, so equal steps through the penumbra do not necessarily produce equal brightness steps. The reusable mechanism is visibility of the emitting area. To diagnose a shadow, ask how much source each surface point can see, then ask what other illumination reaches it beyond that direct source.
Side light can create gradients and shadows that reveal texture. Light near the camera axis can reduce visible side shadows. Backlighting can separate edges or make a silhouette. These are geometric consequences, not a ranking of universally good and bad lighting.
Fill adds illumination to areas receiving little from the main light. A reflector redirects existing light; it does not create energy. A fill source can reduce scene contrast so that both highlights and shadows fit a capture. Changing exposure alone scales both regions and does not have the same effect.
A matte surface sends light toward the camera according to both the illumination and its orientation. Imagine a small flat patch facing a distant directional source. Tilt it until the same bundle of light spreads over twice the surface area. Each unit of area then receives half as much illumination. This is the geometric reason a rounded matte object can show a gradual change in tone even before it casts a visible shadow. Faces also have varying reflectance and mutual shadowing.
For a numerical example, assume identical matte patches facing the camera, with one oriented squarely toward the source and another at 60 degrees to that direction. The tilted patch receives half the direct illumination in the ideal cosine model, since the cosine of 60 degrees is one half. If the first patch receives 200 units, the second receives 100. At 90 degrees the ideal direct contribution is zero. Room reflections can still illuminate that patch. These are illumination amounts, not display brightness.
Direction changes which surface shapes receive these amounts. Light beside the camera can make many visible front surfaces similarly lit and hide their cast shadows behind them. Moving the light sideways makes orientation differences and relief more apparent. A low bump on paper can show a small lit side and a neighboring shadow under grazing light. Editing contrast cannot create physical illumination evidence absent from the capture. Glossy surfaces need separate care because their reflections depend strongly on viewing direction.
Fill changes the contrast calculation. Suppose a main light produces 200 illumination units on a lit patch and 20 on a shadow patch. Their ratio is 10 to 1. Add a broad fill contributing 30 units to each. The totals become 230 and 50, so the ratio is 4.6 to 1. Fill need not reach only the shadow to reduce contrast. Its equal addition is a much larger proportional change for the initially dark patch, which rises by two and a half times rather than just 15 percent.
Compare that addition with doubling capture time. With unchanged illumination, doubling time scales the original captured signals from 200 and 20 to 400 and 40. Their ratio stays 10 to 1. Raising ISO alone may produce a brighter rendering, but does not add illumination or collected photons. A reflector that sends more existing source light onto the dark side can reduce scene contrast before capture. It may also alter the lit side, color, and reflections. The exact change depends on placement and the surface it presents to the subject.
Do not reduce every shadow as far as possible. If the purpose is to show shallow relief, excessive fill can make orientation differences difficult to see. If the purpose is a readable expression, an unlit eye socket may matter more than strong modeling of the cheek. Compare the same object under side light with and without a pale card nearby. Describe what becomes visible and what becomes less distinct. Automatic rendering may alter the apparent tonal difference.
| Case | Lit patch | Shadow patch | Ratio |
|---|---|---|---|
| Main illumination alone | 200 | 20 | 10 : 1 |
| Add 30 illumination units to each | 230 | 50 | 4.6 : 1 |
| Instead double time, no fill | 400 signal units | 40 signal units | 10 : 1 |
| Change | Likely effect under stated conditions |
|---|---|
| Larger apparent source | Broader shadow transition |
| Add fill to shadow side | Lower subject contrast |
| Double ideal point-source distance | Quarter illumination |
The source-angle drawing compares flat emitters with radii of 12 and 50 at a distance of 200 in the same units. Radius is half the width, so their widths are 24 and 100. Seen from the subject, they span about 6.9 and 28.1 degrees. You need not calculate angles to compare them. Their width-to-distance ratios are 24 divided by 200, or 0.12, and 100 divided by 200, or 0.5. The second sends light across a much wider spread of directions.
An angle describes geometry, not the source's power. A large dim panel and a small powerful lamp could deliver the same total illumination to a chosen patch while making different shadow transitions. Conversely, dimming a uniformly glowing panel without changing its emitting area can reduce illumination while preserving its geometric angular extent. Real dimming may change color or brightness uniformity, so check the actual source rather than assuming it. The model separates size and amount so that the word soft does not become a substitute for either low intensity or low contrast.
Distance changes both quantities when the same source moves. Consider a 40 cm flat source centered toward the subject. At 100 cm its width-to-distance ratio is 0.4. At 200 cm it is 0.2, so its apparent width is smaller. Its angular widths are about 22.6 and 11.4 degrees, respectively. A shadow transition can narrow as the source moves away. Do not automatically assign a precise one-quarter illumination to this nearby extended source, because its different emitting points have different distances and directions to the surface.
There is a separate regime where inverse-square reasoning is useful. Assume an ideal point source, no beam shaping or reflections, and a receiving surface held square to the light. At 1 m let illumination be 100 units. At 2 m it is 25, and at 3 m it is about 11.1, because distance squared is 1, 4, and 9. To preserve collection with fixed aperture, time must grow by factors of 1, 4, and 9. ISO changes alone would not restore the lost collection.
Distance also changes illumination differences across a scene. Put a subject 1 m from that point source and a background 2 m from it, with both surfaces facing the source. Their illumination ratio is 4 to 1. Move the source back so the subject is 3 m away and the background 4 m away. Their illumination ratio becomes 16 divided by 9, about 1.78 to 1. Both become dimmer, but the relative falloff becomes less steep because the additional meter is a smaller fraction of the total distance.
Increasing power ninefold would restore the subject's original illumination at 3 m in this ideal example. The background would then receive 9 times 100 divided by 16, or 56.25 units, rather than its original 25. Equal subject brightness therefore does not imply equal background brightness. This is a useful contrast with changing exposure, which scales both recorded regions together. Windows, panels, room reflections, and shaped beams need different models.
For an ideal point source radiating into free space, illumination on a surface facing it decreases with distance squared. Doubling distance quarters illumination. Large nearby sources, beam-forming optics, reflections, and changing surface angle can violate this simple approximation. Do not use it as an exact rule for every window or softbox.
Try a matte object near a window and farther into the room. Observe shadow-edge width, direction, and background brightness separately. Use safe low-heat lights; keep diffusers away from hot lamps. Never look at the sun through an optical viewfinder or magnifying optics. A lighting exercise does not require dangerous direct-sun experiments.
Use a matte object with a clear raised edge, such as a folded piece of paper, near a broad window. Choose a position where the light comes mainly from the side. Before making any photograph, inspect where the object blocks the window, where its shadow meets the table, and where the shadow extends farther away. Predict a narrower transition near contact and a broader transition where the object stands higher. This prediction concerns source visibility, not how dark the camera will choose to make the picture.
For a second comparison, bring a pale card toward the shadow side without moving the object or blocking the window. Look for a change in shadow brightness and in the object's dark surfaces. The card redirects existing light, so its orientation matters. A card facing away from the window may send little useful light back. Avoid treating white card size alone as a guarantee of fill. A small well-positioned surface can produce more relevant reflected light than a large surface that receives little illumination or faces the wrong direction.
If a camera is available, include the object and enough surroundings to identify the same shadow region in both photographs. Keep viewpoint and framing steady. Use available exposure or white-balance lock. Otherwise report the visible result as a combined lighting-and-processing comparison. An automatic camera might darken the whole second image when the card brightens the scene. That global change can obscure an increase in physical shadow illumination. Comparing the shadow with the lit side within each frame is more informative than comparing shadow brightness alone.
Suppose a simplified linear capture gives lit and shadow values of 200 and 20 before fill, then 230 and 50 after fill. If automatic exposure halves the second capture, its readings become 115 and 25. The shadow now looks only slightly higher numerically, while the lit side becomes much lower. The within-frame ratios still change from 10 to 1 to 4.6 to 1. This cancellation works for common linear scaling. Tone curves and local processing can change those ratios, so ordinary displayed JPEG numbers are not direct light-meter readings.
When you move farther from the window, several variables can change together. Its apparent size usually decreases, its direction can shift, and nearby walls may contribute a different fraction of the total illumination. Record a concrete observation, such as a sharper shadow boundary beside the raised paper edge, rather than asserting that every window follows inverse-square falloff. A confounded comparison calls for a simpler next arrangement.
Judge the result at the intended output size as well as in a close view. A tiny shadow transition may be invisible after reduction even though it reveals texture in a larger reproduction. Acceptable sharpness depends on that output, while changing light direction does not repair defocus or motion blur. Keep improvised diffusers away from hot lamps. Without any camera, sketch the observed contact shadow and fill change, then use the numerical cases to explain which changes affect geometry, illumination, or rendering.
| Linear model capture | Lit value | Shadow value | Lit / shadow |
|---|---|---|---|
| Before fill | 200 | 20 | 10 |
| After fill, same exposure | 230 | 50 | 4.6 |
| After fill, exposure halved | 115 | 25 | 4.6 |
An ideal point source moves from 1 m to 2 m while surface orientation stays fixed. What exposure-time factor compensates for lost illumination?
Use inverse distance squared.
| Distance | Illumination | Compensating time |
|---|---|---|
| 1 m | 1 | 1 |
| 2 m | 1/4 | 4 |
Illumination becomes one quarter, so time must become four times as long if aperture and scene behavior stay fixed. This also increases motion integration. An extended nearby source may not follow that exact factor.
You add a reflector that mainly brightens a face’s shadow side without changing the lit side much. How does this differ from raising exposure?
Ask whether the scene ratio changes.
| Operation | Lit side | Shadow side | Ratio |
|---|---|---|---|
| Exposure ×2 | ×2 | ×2 | Same |
| Targeted fill | About unchanged | Higher | Smaller |
Fill reduces the ratio between the two sides. Raising exposure alone multiplies both and leaves that ratio unchanged, potentially clipping the lit side. The reflector can therefore make more of the subject fit the capture range without requiring highlight loss.
The lessons and answers stand alone. These optional secondary references explain the optics and imaging concepts. Numerical examples are constructed, not measured camera specifications. Visual judgments depend on the intended image; the course does not claim a universal aesthetic score.
Qualitative overview of apparent source size and shadow transitions. Some distance wording is overgeneralized; use the separate inverse-square reference and the stated source geometry for quantitative calculations. Real fixtures and nearby extended sources need more than a point-source approximation.
Geometric dilution for an ideal point source. Beam-forming optics, surface angle, and nearby extended emitters limit direct application.