Why your before and after photos do not match

Two photos, twelve weeks apart, and the second one somehow looks worse. Or identical. Or so different that it feels dishonest.

Almost always the body did what it did and the photograph moved. There are four variables involved, they are not equally important, and only one of them can be meaningfully repaired after the fact.

1. Camera distance, the one nobody notices

This is the biggest offender and the hardest to spot, because a photo taken from a metre away and a photo taken from two metres away look like ordinary photos. Nothing in either one announces that the distance changed.

What changes is apparent proportion. A lens close to a subject exaggerates whatever is nearest to it and shrinks everything behind. Move a phone in by half a metre and the part of you closest to the camera, usually the torso or the nose, grows relative to everything else. Move it back and the same body reads as narrower and more evenly proportioned.

Which means: a progress photo taken slightly closer than the first one can erase twelve weeks of real work, and one taken slightly further back can invent progress that never happened.

It also creates a trap specific to fat loss. As people get leaner they often, unconsciously, step slightly closer, because the photo feels like it needs to show more. That single habit works directly against the thing they are documenting.

The fix is a floor mark. A strip of tape where your heels go and a fixed place for the phone. Not “about two metres”, an actual mark. This is a one-off decision that removes the largest variable in the entire process.

2. Camera height, which quietly rearranges you

Height does something different from distance: it changes which parts of the body the lens is looking down at or up at.

A phone at waist height looks up, which lengthens the legs and shortens the torso. A phone at eye height looks slightly down, which does the reverse. Neither is wrong, but they are not comparable, and a phone propped against whatever object happened to be free that week wanders between them constantly.

For full body photos, chest height and level is the standard, because it minimises how much of the body is being viewed at an angle. For faces, eye level. For rooms, level with the camera not tilted at all, because tilting turns straight walls into converging ones and the amount of convergence changes with every degree of tilt.

The fix is a fixed surface. A shelf, a windowsill, a chest of drawers, a stack of books that lives in that spot. The specific height matters far less than it being the same one every time.

3. Angle, the one that compounds

Rotation is the sneakiest of the four because it accumulates.

A body turned five degrees from the last photo is imperceptible in isolation. Five degrees a session, over twelve sessions, is sixty degrees, and by photo ten you are comparing a front view to a three-quarter view without ever having decided to. The same thing happens to a phone resting against a wall that gets nudged slightly every week.

Angle drift is what makes people say the middle of their photo set is unusable. It rarely breaks a set suddenly. It erodes it.

The fix is a reference in frame. Something rigid that does not change: a door frame, a light switch, a picture rail, a window edge. Anything you can check the new photo against. This is exactly what a ghost overlay automates, but a light switch and some attention does the same job.

4. Light, the only one you can partly fix later

Light does not change your shape. It changes how much of your shape is visible.

Overhead light, the kind in most bathrooms, throws shadow into every dip and texture. On a body that reads as muscle definition. On skin it reads as texture and pores. On a hairline it reads as thinning. Flat frontal light, the kind you get facing a window, removes almost all of it.

The two are so different that the same person photographed under both, minutes apart, can look like two stages of a transformation. This is the mechanism behind the entire genre of misleading fitness before and afters, and it does not require any dishonesty to happen accidentally.

Light is also the only variable in this list with a partial after-the-fact remedy. Exposure and contrast can be nudged toward matching, at least enough to compare. You cannot undo the direction the light came from, but you can stop one photo being twice as bright as the other.

The fix is one window and one hour. Facing a window, ceiling light off, at roughly the same time. Overcast days are the most repeatable light there is.

The four variables, and which survive an editor

Camera distance

Changes apparent proportions. Nothing in either photo announces that it moved, which is what makes it the worst offender.

Cannot fix later

Camera height

Decides which parts of you the lens looks down at. Waist height lengthens legs, eye height does the reverse.

Cannot fix later

Angle

Five degrees a session is invisible once and obvious by photo ten. It erodes a set rather than breaking it.

Cropping only hides it

Light direction

Does not change your shape, changes how much of it you can see. Overhead light carves definition that is not there.

Partly fixable

How much does each one actually matter?

Variable	What it changes	Fixable afterwards?	Fix at capture
Distance	Apparent proportions	No	Floor mark, fixed phone spot
Height	Which parts are foreshortened	No	One fixed surface
Angle	Which view you are comparing	No, cropping only hides it	A rigid reference in frame
Light	Visible definition and texture	Partly, exposure and contrast	One window, ceiling light off, same hour

Three of the four have to be decided before the photo exists. That is the whole argument for fixing framing at capture time rather than in an editor.

Why editing afterwards makes it worse, not better

There is a standing temptation to solve this in software: load both photos, rotate, scale and warp until they overlap, and export.

It does produce two images that line up. It also does two things that quietly damage the record.

Cropping throws away real information. Forcing two differently framed photos into a matching crop means discarding the edges of both. On a full body shot that often means losing exactly the outline you were trying to compare.

Warping invents information. Any rotation or perspective correction resamples the image, which means every pixel in the output is a guess assembled from neighbouring pixels in the input. For a holiday photo that is fine. For a record of a body, a wound, a treatment or a build, it means the thing you are examining has been partly synthesised.

Neither operation recovers the angle you did not shoot from. They make the mismatch less visible while making the photo less true, and those are opposite goals.

The shortest possible version

Mark the floor. Fix the phone’s spot and height. Put something rigid in frame. Face a window with the ceiling light off. Shoot at the same time of day.

Then the only thing left in the photograph that can change is the thing you are photographing, which was the point.