The short answer
A fair comparison needs three conditions matched: the camera angle, the scale of each item inside the frame, and the lighting direction and colour. When all three match, a visible difference is worth noting. When one of them does not match, that difference explains most of what you are seeing, and the honest conclusion is “condition mismatch, not comparable yet”.
This guide is part of an independent series that supports a Lovegobuy spreadsheet workflow. It talks about image conditions and measurement habits, and it deliberately avoids claiming that any item is genuine or better. If you are still collecting views, read the guide to blurry or cropped photos first, and the QC notes guide afterwards for how to write the result down.
Lock angle, scale and lighting in that order
Angle comes first because it changes proportion more than anything else. A garment photographed from 20 degrees above looks shorter than the same garment photographed level. Once the angle is matched, scale becomes the next problem: an item that fills 90% of the frame will look larger than one filling 45%, even if they are identical size. Only after those two are settled does lighting matter, and it matters mostly for colour and surface texture rather than shape.
In practice you rarely get perfect matches, so the goal is not perfection. The goal is to know which conditions match, so you can say precisely which comparison is supported. Writing “angle matched, lighting different, scale unknown” is far more useful than writing “looks similar”.
The table scrolls inside its own frame on narrow screens.
| Condition | Matched | Mismatched | What stays valid |
|---|---|---|---|
| Angle | Both side-on, camera level | One level, one slightly from above | Colour and hardware finish only |
| Scale | Item fills a similar share of frame | One wide shot, one tight crop | Stitching style, if both are sharp |
| Lighting | Same direction and colour temperature | Soft box versus warm room light | Shape and silhouette only |
| Resolution | Both shareable originals | One forwarded compressed copy | Layout and coarse colour blocks |
Two pairs, each with exactly one mismatch
Deliberate mismatch makes the effect easier to see than a perfect pair does. Consider pair one: two images of the same category, both shot side-on with similar framing, but one under a neutral soft box and one under a warm ceiling lamp. The warm image will read slightly beige. If you list “tone difference” as a fact without the lighting note, you have turned a lighting artefact into an item claim.
Now consider pair two: both images under the same light, both sharp, but one shot from level and one from a slightly high angle. Here the tone reads consistently, so colour comparison is supported. Proportion is not: the high-angle image compresses height, so any length comparison is invalid. Two pairs, two different valid readings — and in both cases the useful output is a statement about which comparison is allowed.
This is why a single “similarity score” is not very useful for photo work. A number cannot carry the sentence “colour comparable, proportion not”. Keep the conditions in words and use numbers only where a real measurement exists.
The order is: match conditions, check readability, then compare. Doing it in any other order just produces confident sentences that are hard to defend later.
Scaling without lying to yourself
If you resize one image to match the other, write the factor down. Resizing does not create new detail; it only changes how large the pixels appear. A 760-pixel image enlarged to match a 1600-pixel one will look softer, and that softness can be mistaken for a material difference.
A cleaner approach is to use a physical reference already in the frame, such as a tape measure or a flat card of known width, and compare ratios rather than raw pixel sizes. When no reference is present, treat any size comparison as approximate and say so. This is also the reason the guides here show inches with centimeters in brackets: two units make it obvious when a figure is a rounded display value rather than an exact measurement.
For colour, keep the same rule. Note the light source, then compare. If one image has a visible colour cast across the whole frame — including the background — that cast is almost certainly the lighting, not the fabric.
Worked example: a 3.4% length gap
Here is a small example you can recalculate. Image A shows a flat-measured length of 29 cm. Image B shows 28 cm. The gap is 1 cm, and 1 ÷ 29 ≈ 3.4%. That is the entire arithmetic, and it is teaching values only: no real listing was measured.
Now apply the conditions: if both readings started at a clearly marked seam and the same tape was used, a 3.4% gap is small but worth a second reading. If one reading started at a folded edge and the other at an open edge, the 1 cm is entirely explained by where the tape began, and the correct note is “measurement origin differs, not comparable”.
A second, independent reading from the same origin is the cheapest way to settle it. Two readings that agree within a few millimeters are much stronger evidence than one clean-looking number.
Writing the conclusion
Finish every comparison with one sentence that names the supported claim and one that names the open question. For example: “Colour and hardware finish match under similar light; length is 1 cm apart and not yet settled because the measurement origins differ.” That sentence survives a second look, and it tells the next person exactly what to re-check.
Once written, drop it into the coverage checklist on the home page as a follow-up, or read the QC notes guide to structure the same sentence into observation, inference and request. To start a new row, go back to the category list.
Source note: this page relies on general photo-comparison practice and simple percentage arithmetic. Any statement about how an item is photographed, measured or described in an official Lovegobuy process must be confirmed on the official help pages, since conditions vary by country and item type. All figures here are teaching values.
Where to go next
Use the QC notes guide to record what the comparison supports, and the blur and crop guide if one of the images turns out to be unreadable.
Questions readers ask about this
Can I compare a listing photo with a warehouse photo?
You can, but only for details both images show under similar conditions. Different cameras, distances and light sources make tone and proportion unreliable, so mark those points as unmatched instead of reading a difference into them.
How much length difference is meaningful?
A small rule of thumb is that a gap under about 2% of the total length is usually inside flat-measurement noise. A gap above roughly 5% is worth a second measurement from the same starting point before you treat it as real.
Should I brighten one image to match the other?
Adjusting brightness changes the pixels you are comparing, so note the adjustment if you do it. A safer habit is to compare the same detail under each image’s own light and record the light difference separately.