A diffuser softens light when it effectively turns into a larger apparent light source (relative to your subject) and/or evens out “hot spots.” It just takes away light when it mainly acts like a dim filter because its apparent size (from the subject’s point of view) doesn’t meaningfully increase.

When diffusion actually softens light

Softness is about the shadow edge: how quickly tones transition from light to shadow. That transition gets smoother when light arrives from a wider range of angles—meaning the subject “sees” a larger source. A diffuser can help by becoming the new emitting surface.

1) The diffuser must matter more than the bare source

If the diffuser is large enough and positioned so the subject “sees” it as the primary source, the diffuser surface becomes the effective light source. This is why a big sheet of diffusion close to someone’s face can look creamy: the face is being lit by a broad, nearby surface rather than a small point.

Practical cue: look at catchlights (reflections in eyes, glossy objects). If the catchlight shape/size matches the diffusion panel/softbox front, the diffuser is behaving like the source, and softness is real.

2) Distance is part of softness (not just size)

A large diffuser far away can still look small to the subject. Conversely, a modest diffuser very close can look huge. This “apparent size” concept is the core reason you can’t judge softness by modifier diameter alone. (strobist.blogspot.com)

Practical cue: If you double the distance from diffuser to subject, the diffuser appears about half as large, and shadows get crisper—even if nothing else changed.

3) Diffusion can improve uniformity even when softness changes little

Not all diffusion is about shadow edges. Sometimes the goal is to reduce a bright center (“hot spot”) and produce a more even field across the front of a softbox or panel. An inner baffle in a softbox is a classic example: it can make the front face more uniform, which often reads as smoother light on skin because highlights are less peaky and transitions are less patchy.

Practical cue: When your subject has a shiny forehead, glasses, or reflective product surfaces, a more uniform emitting surface reduces ugly bright patches—even if the shadow edge doesn’t dramatically change.

When diffusion mostly just takes away light

Diffusion always costs output. The question is whether you get softness in exchange, or mostly just a dimmer scene.

1) The diffuser is too close to the light and too far from the subject

If you place a small diffuser right on the face of a small light (or right over an LED panel) and the subject is several feet away, the subject still “sees” roughly the same apparent source size—maybe only slightly larger. In that case, you’ve mostly inserted a scattering filter that reduces intensity and spreads the beam a bit, but doesn’t create much softer shadow geometry.

Rule of thumb: Diffusion placed near the lamp behaves more like beam control; diffusion placed near the subject behaves more like softness control.

2) The diffuser isn’t larger than what you already have

If your light is already effectively large (for example: a big softbox close to the subject), adding an extra diffusion layer may not meaningfully increase apparent size. At that point, you’re mainly:

• losing a fraction to a couple stops of light,
• reducing peak highlights slightly,
• perhaps smoothing minor unevenness.

This is the “diminishing returns” point: the first step (turning a small source into a big one) is dramatic; additional layers can be subtle.

3) You’re using diffusion to fix a problem caused by distance

A common trap is keeping the light far away (for convenience or framing) and adding thicker diffusion to “make it soft.” Thick diffusion can’t fully compensate for a small apparent source. You may end up with the same crisp shadow edge—just at a higher ISO or wider aperture.

If softness is the goal, the more reliable move is usually: bring the (diffused) source closer or make the emitting surface larger, then adjust exposure accordingly.

4) The diffusion is so dense (or stacked) that output loss dominates

Denser materials and multiple layers can quickly eat output. Stacking is sometimes useful to make the surface more even, but after a point you’re mainly throwing away photons without changing geometry. (Photography Stack Exchange)

Practical cue: If adding a second layer forces you into uncomfortable settings (high ISO noise, slow shutter blur, maxed-out flash recycle) and the shadow edge barely changes, you’ve crossed into “mostly light loss.”

A simple way to predict which side you’re on

Use this quick mental checklist before you set anything up:

1. From the subject, how big does the diffuser look?
   If it looks big, it’ll usually look soft. If it looks small, it won’t—no matter how “milky” the material is.
2. Is the diffuser the thing the subject is actually seeing?
   If the bare lamp is still visibly “the source” (small bright area), diffusion is acting more like a filter than a new light surface.
3. What problem are you trying to solve: shadow edge or highlight texture?

• Shadow edge softness → size + proximity.
• Highlight texture/uniformity → diffusion can help even if size doesn’t change much.

Practical setups that clarify the difference

Setup A: “Softening” that actually softens (diffuser close to subject)

• Put a 3×3 ft diffusion panel (or the front of a softbox) 1–2 ft from the subject.
• Put the lamp 1–3 ft behind the diffusion.
  Result: the subject is lit by a broad, nearby surface—soft shadow edges, gentle transitions.

Setup B: “Softening” that mostly dims (diffuser close to lamp)

• Put a small diffuser cap or thin diffusion sheet right on the lamp.
• Place the lamp 6–10 ft from the subject.
  Result: output drops, beam spreads a bit, but the source still appears small; shadow edges remain relatively hard.

Setup C: The “hot spot fix” (inner diffusion/baffle)

• Keep your softbox front diffusion on.
• Add/remove the inner baffle.
  Result: often subtle change in edge softness, but noticeable improvement in evenness and specular highlight behavior—especially on reflective surfaces.

How to choose diffusion strength without guessing

Instead of thinking “more diffusion = softer,” think in two dials:

Dial 1: Apparent size (main softness dial)

• Increase the emitting surface size, or
• move it closer to the subject.

Dial 2: Scatter/texture (secondary dial)

• Add diffusion to reduce harsh highlight texture and hot spots,
• but stop when exposure cost becomes painful.

If you need a measurable approach: take two quick test shots with and without the extra diffusion layer, same framing and distance. Compare:

• the edge transition of a shadow (jawline or nose shadow),
• the shape and intensity of specular highlights (forehead, glasses, product reflections),
• and whether you had to change exposure settings in a way you dislike.

Why does this matter

Because “diffusion” is often used as a cure-all; knowing when it changes light geometry versus when it just costs exposure headroom helps you get the look you want with fewer compromises.

Sources

• Lee Filters – Diffusion Index (LDI): https://leefilters.com/lighting/diffusion-index-ldi/ (LEE Filters)
• Rosco – Diffusion material in film & video: https://emea.rosco.com/en/diffusion-material-film-video-applications (Rosco)
• Strobist – Apparent light size (softness vs distance): https://strobist.blogspot.com/2007/07/lighting-102-unit-21-apparent-light.html (strobist.blogspot.com)