What side, front and top light each do

Light from the side reveals texture and edges that light from straight ahead flattens out, and a rating tool only ever scores what it can see. Change the direction the light is coming from and the same subject can hand the tool a visibly different image, without one thing about the subject itself changing.

What direction actually does to an image

Front-on light, the kind that comes from behind or beside the camera and hits the subject squarely, minimises shadow and flattens surface detail. Side light casts shadow along every contour, which is what lets a viewer, human or model, read shape and texture at all - it is the difference between a relief map and a flat photograph of the same terrain. Top light behaves like a weaker, more diffuse version of side light, softer shadows in different places. Perception research shows how much shape rides on this: Johnston, Hill and Carman (1992) note that when shading is ambiguous the visual system assumes light from above, and found that lighting faces from below disrupted how viewers represented their shape. None of these is more truthful than another; they are three different renderings of one subject, and a tool has no way to know which rendering it is looking at unless the direction is held constant.

Why this moves axes that read detail

Rubrics that score anything texture- or edge-adjacent are sensitive to how much relief the light gave the camera to work with. Flat front light can under-represent detail that is genuinely there, because the shadow that would have shown it never formed. Harsh side light can over-represent it, throwing shadow into creases that a softer angle would smooth over. Both are readings of the same subject under different light, and a tool comparing two such readings is not comparing the subject - it is comparing the lighting.

Why consistency beats brightness

The instinct is to chase more light, on the reasoning that more light means more visible detail, and that instinct is only half right. Whether an image is bright is a different question from whether its light direction repeats, and the second one is what a comparison actually depends on. A dim photo lit from a consistent direction is more useful for tracking change over time than a bright one lit from wherever the room's light happened to be that day. Direction is the variable to fix; exposure only has to be adequate and then held the same way.

Front, side and top, compared directly

Front light - roughly from the camera's position - minimises the shadow a viewer or a model can use to read shape, because the light source and the viewing angle are nearly the same, leaving almost nowhere for a shadow to fall. It is the most forgiving light to set up and the least informative to score from, which is a real trade-off, not a flaw in the subject. Side light, roughly ninety degrees off the camera axis, produces the most shadow and therefore the most legible relief, at the cost of being the least forgiving to get consistent - a source moved a few degrees changes the shadow pattern noticeably. Top light sits between the two: softer than hard side light, more revealing than flat front light, and it is what most indoor ambient lighting already approximates, which is partly why it is a common accidental default rather than a deliberate choice. None of these three is the "correct" one; the requirement is picking one and returning to it, not identifying which is objectively best.

A quick check for whether direction actually moved

If a score shifts between two sessions and distance, angle and crop all check out the same, direction is worth a look before anything else, because it is one of the least visually obvious of the variables to have drifted. The tell is where the shadows fall in the frame relative to the subject's own contours - if a shadow that used to fall on one side now falls on the other, or has mostly disappeared, the light moved, and any score change attributed to the subject should be treated with suspicion until that is ruled out.

Where this fits in a protocol

Light direction belongs on the same short list as distance, angle and crop - the handful of things worth pinning down before a second result means anything against a first. Standardising a submission is the general version of this argument; light direction is one line item inside it. The specific choice between two light sources that both hold direction constant - flash against window light - is a narrower question with its own answer.

What a model is actually reading

A tool is not evaluating "good lighting" as a separate judgement; it is reading whatever detail the light happened to expose and folding that into whatever axis depends on detail. What the underlying model does with a given frame goes further into that mechanism than this note needs to; the practical point is that the model has no way to discount for lighting it was never told about.

The gear that is actually involved

None of this calls for studio equipment. The relevant gear question is a single lamp or a window, not a kit, and the discipline is choosing one source and one direction, then keeping it. The same discipline pays off if the photo is headed to a person rather than a model - a human judge is reading the same shadows the tool would, and a subject lit consistently is easier to assess honestly either way. Rate Cock's own photo guide specifies a light direction for its own submissions, which is the correct instinct wherever the file ends up.