Jewelry product photography has three problems most product shoots do not. Metal reflects its surroundings, so the studio ends up in the frame. The pieces are small, so a buyer cannot judge size from a packshot. Stone and alloy colour has to match the physical item, because it is the attribute buyers claim they were shown. The shot set that answers all three is four images: packshot, macro, scale, worn.
A ring photographs beautifully on its own and still gets returned. The buyer saw a stone filling their screen, then opened a box holding something the width of a grain of rice. Nothing about the photo was retouched. The scale was never communicated.
Jewelry sits at the sharp end of ecommerce photography for that reason. The margin is high, the return cost is high, and the two things that drive returns, size and colour, are exactly the two things a close-up product shot handles worst.
This covers the four-shot set, why metal is difficult to light, how to give a buyer scale they can read, the federal guidance on depicting stones larger than they are, and where generated imagery helps against where it must not go.
Key Takeaways
- Four shots carry a jewelry listing: a clean packshot, a macro detail, a scale reference, and the piece worn. Each answers a different question and none of them substitutes for another.
- Metal photographs its surroundings, not itself. A polished surface is a mirror, so what you see in the shot is the room. Control the reflection and you have controlled the image.
- Scale is the failure buyers punish. Etsy’s Seller Handbook suggests putting accessories on a model where possible, or holding the product in your own hand for comparison (Etsy Seller Handbook, September 2026). If that hand is generated rather than photographed, the rules for generating a hand for scale apply.
- There is federal guidance about making a stone look bigger. Note 3 to §23.1 of the FTC Jewelry Guides says an illustration or depiction of a diamond or other gemstone that portrays it in greater than its actual size may mislead consumers, unless a disclosure is made about the item’s true size (16 CFR 23.1, accessed September 2026).
- Amazon’s main image rule is the strictest constraint on a small product. Amazon asks for pure white at RGB 255, 255, 255, with the product at 85% of the image (Amazon product image guide, accessed September 2026). On a stud earring, hitting 85% means shooting close, which makes the scale shot mandatory somewhere else in the set.
- Generated imagery is safe on the set and unsafe on the stone. Backgrounds, hands, necks and scenes are presentation. Carat, cut, colour and setting are the product.
What makes jewelry harder to photograph
Three properties compound. The pieces are small, which pushes you into macro work where depth of field collapses to millimetres. The surfaces are specular, so they act as mirrors rather than diffusers. And the colour is a claim, because a buyer purchasing 14k yellow gold with a London blue topaz is buying those exact descriptions.
A cotton t-shirt forgives all three. It is large, it scatters light in every direction, and nobody returns it because the grey read slightly cool.
Jewelry forgives none of them, which is why the shot set has to be built deliberately rather than assembled from whatever came out well.
The four-shot set every listing needs
| Shot | What it answers | Where it goes |
|---|---|---|
| Packshot | What is it, and what does the whole piece look like | Main image, marketplace listing |
| Macro detail | What is the quality of the setting, stone and finish | Secondary image, high-consideration buyers |
| Scale reference | How big is it, really | Secondary image, the returns preventer |
| Worn | What does it look like on a person | Secondary image, and usually the ad creative |
The packshot carries the marketplace requirement. The macro carries the price justification. The scale shot prevents the return. The worn shot does the selling. Small leather goods split the same four jobs slightly differently, and the wallet shot list adds a loaded frame because thickness is the claim buyers check.
Many small jewelry brands shoot the first two and skip the second two, which is exactly backwards for the two things that cost money after the sale.
The complete product photography shot list covers the general seven-angle version of this. Jewelry compresses it, because a ring has fewer meaningful angles than a handbag and far more meaningful detail per square millimetre.
Reflections: the surface is a mirror
A polished metal surface does not have a colour of its own in a photograph. It shows whatever is in front of it. Point a bare light at a silver band and you get a hard white streak and black surroundings, because the band is reflecting a small bright source against a dark room. The same physics decides every scene in a watch catalog, where a domed crystal gathers a wide arc of the room into a few millimetres of sapphire, and nineteen watch scenes sort by exactly what ends up in it.
The fix is to give the metal something worth reflecting.
Make the source large and close. A big diffused source relative to a small object wraps around it, so the reflection becomes a broad soft gradient instead of a hot line. On jewelry this is the single highest-impact change, because the object is tiny and almost any softbox is large relative to it.
Build a white surround. A light tent, a cylinder of white paper, or three bounce cards forming a box gives the metal a bright even world to mirror. What reads as “well-lit metal” is usually just metal reflecting a lot of white.
Kill your own reflection. The camera, the tripod and the person operating it all appear in a curved polished surface. A black card with a lens-sized hole cut in it, positioned in front of the lens, removes the camera from the reflection.
Flag for shape. Pure white surround with no dark areas leaves the metal flat and shapeless. Introduce one narrow black card to create a dark band in the reflection, which is what reads as an edge.
Stones behave differently. A faceted gem needs a small directional source to produce sparkle, because sparkle is specular return from individual facets. Soft light kills it. That means jewelry with both polished metal and faceted stones often needs a large soft source for the band and a small hard one for the stone, which is why the category takes longer per shot than almost anything else.
Scale: the shot most listings skip
A buyer cannot judge size from a product photographed against nothing. There is no reference in the frame, and every listing thumbnail is the same size on screen, so a pendant and a dinner plate occupy identical space.
Etsy’s Seller Handbook is direct about this. It suggests putting accessories on a model where possible, or holding the product in your own hand for comparison. Its clothing advice adds a note of the model’s measurements and the item size, so the buyer has a reference for fit (Etsy Seller Handbook, September 2026). A jewelry seller quoted in the handbook describes length with physical reference points such as chin length or just below the collarbone, and puts the exact measurements further down in the item details (Etsy Seller Handbook, September 2026). A wrist or a neck is usually enough, which is the second of the four levels of person a product shot can use. Soft accessories have the same scale problem and less room to solve it, since Amazon asks for clothing accessories flat and without a model in the main image, as scarf product photography sets out.
Three ways to communicate scale, in order of how well they work:
- Worn on a body. A ring on a hand, a necklace at the collarbone, earrings against a jawline. This is the most readable and it doubles as the shot that sells.
- Against a familiar object. A coin, a fingertip, a matchstick. Readable, cheap, and slightly clinical for a luxury listing.
- A dimension callout on the image. Measurements printed over the photo. Precise and least persuasive. Amazon’s main image rules describe the product alone on pure white, so keep the callout off the main image.
Use one and two together. The worn shot sells and the measured detail confirms. Wrist-worn pieces need a fifth frame on top of the four, because closure and drape are questions a macro cannot answer, which is why bracelet photography runs its own shot list.
The worn shot is where skin retouching becomes part of jewelry work rather than beauty work. A macro of a ring on a hand puts skin texture, cuticles and knuckle creases at the same magnification as the setting, and the fix is a light clean-up that keeps real texture rather than a smoothing pass that turns the hand into plastic. The hand is the scale reference. Erase its texture and you have erased the reference.
The FTC guidance on making a stone look bigger
This one is worth knowing before anyone briefs a hero image.
The FTC Jewelry Guides, at 16 CFR Part 23, cover how jewelry, gemstones and precious metals are described and marketed in the United States. The current text dates from a revision published in August 2018 (83 FR 40667). Section 23.1 says it is unfair or deceptive to misrepresent the size, weight or any other material aspect of a product in the category.
Note 3 to §23.1 addresses pictures specifically: “An illustration or depiction of a diamond or other gemstone that portrays it in greater than its actual size may mislead consumers, unless a disclosure is made about the item’s true size” (16 CFR 23.1, accessed September 2026).
The Guides set out the FTC’s current thinking and do not bind the FTC or the public. The FTC can still take action under the FTC Act if a claim is inconsistent with them, and it must then prove the claim is unfair or deceptive under Section 5 (16 CFR 23.0, accessed September 2026). This is general information, not legal advice.
Read plainly, that means an enlarged hero image of a stone is allowed and an enlarged hero image with no stated size is the risk. The practical response is not to avoid macro photography, which every jewelry listing needs. It is to make sure the true dimensions appear alongside it, and to keep a scale shot in the set.
This matters more with generated imagery than with a camera, because a camera can only enlarge a stone by moving closer, and a model can enlarge it by being asked to. The same rule that applies to a photograph applies to a render.
Marketplace specs for a very small product
Amazon asks for the main image on a pure white background at RGB 255, 255, 255, with the product at 85% of the image. The whole product sits in the frame, shown once, with no props that are not included. Only a few product types may use a lifestyle main image, and then with no text or extra logos. Images need at least 500 pixels on the longest side, up to 10,000, and 1,000 pixels or more turns on zoom on the detail page (Amazon product image guide, accessed September 2026). Amazon may remove a listing from search until the main image complies.
The 85% rule is the interesting one for jewelry. On a sofa it is trivial. On a pair of stud earrings it forces a very close crop, which means the main image is by definition an enlarged depiction with nothing in frame to indicate size. That is not a violation on its own, and it is a good reason to treat the scale shot as required rather than optional. The category-specific version of that squeeze, including the pair problem and the nickel limit on the post, is in earring photography.
Getting a clean pure white behind a reflective object is also harder than it sounds, because the white surround you built to control reflections is the same white that has to read at maximum in the file. That is a lighting problem before it is a background problem, and it is covered in how to fix product photo lighting.
Etsy runs differently. Its help page asks for a first photo with enough border to crop to square, portrait and landscape thumbnails without losing the product (Etsy Help, September 2026). If you sell across both, treat position one on each channel as a different shot rather than a different crop. The channel-by-channel breakdown is in ecommerce product photography specs.
Where generated imagery helps, and where it must not go
The useful split is presentation against product.
Safe to generate. Backgrounds and surfaces. Styled scenes and props. Seasonal treatments of the same piece. Hands, necks and ears for worn shots, where the body is a display stand rather than a claim. Multiple angles of the same physical piece from one good original, which is what a multi-angle template makes. Overhead compositions of a collection, which is a job for a flat lay template, and works well for jewelry because the pieces are naturally flat and the whole range fits in one frame.
Not safe to generate. Carat weight and stone size. Cut and facet pattern. Metal colour and finish. The number of stones. The setting type. Hallmarks and stamps. Anything a buyer would use to identify the specific item they are purchasing.
The line is the same one the FTC note draws. Change the room and you have styled a photo. Change the stone and you have described a different product.
In practice this means the input matters more than the prompt. Everything should start from one accurate photograph of the actual piece, so the generated set keeps the real geometry, the real setting and the real proportions. Check each result against the piece anyway. The jewelry photography page shows the set across a collection, and a hero product shot template is a starting point for the packshot.
The general version of this constraint, and how to test whether a generated image still represents the physical object, is in what to check before you ship a product photo.
Build the four-shot set once, as a workflow, on your hardest piece: the most polished metal and the smallest stone. Settle the surround, the light direction and the crop there. A saved workflow runs the same way on the next piece, with the stone and the setting taken from the original photograph each time. The free plan takes no card. Sign up for the free plan, start from a template, and see the cost before you run it.
FAQ
How do you photograph jewelry without reflections?
You cannot remove reflections from polished metal, because a mirror always shows something. You replace what it reflects. Surround the piece with white bounce cards or a light tent so the metal mirrors a large even white field, then add one narrow black card to create a dark band that reads as an edge. Block the camera’s own reflection with a black card with a hole cut for the lens.
What lighting is best for jewelry photography?
Two sources doing different jobs. A large diffused source close to the piece controls the metal, because a big soft source relative to a small object produces a broad gradient instead of a hard streak. A small directional source adds sparkle to faceted stones, which soft light removes entirely. Metal wants soft, stones want hard, and most pieces need both.
How do I show the size of a ring or necklace in a photo?
Put it on a body. A ring on a hand and a necklace at the collarbone are the most readable scale references, and they double as the shot that sells. Back that up with exact measurements in the listing text and a physical reference point such as chin length. A coin or fingertip in frame works as a second option and reads more clinical. Rings need one frame more than that, because band width and profile height are absent from the size number, and the ring shot list covers what those frames have to show.
Is it legal to show a gemstone larger than it is?
Showing an enlarged stone is not prohibited on its own. The FTC Jewelry Guides note that a depiction portraying a diamond or other gemstone in greater than its actual size may mislead consumers, unless a disclosure is made about the true size (16 CFR 23.1, accessed September 2026). The Guides do not bind anyone, but the FTC can act under the FTC Act against a claim inconsistent with them. Publish the dimensions with the image. This is general information, not legal advice.
Can I use AI-generated images for jewelry listings?
For the scene, the background, the styling and the body wearing the piece, yes. For the stone, the metal, the setting and anything a buyer would use to identify what they are buying, no. Generate from an accurate photograph of the actual piece so the image keeps the real geometry, and check carat, cut, colour and setting against the physical item before anything is published.
How many photos does a jewelry listing need?
Four types as a minimum: packshot, macro detail, scale reference and worn. Marketplaces allow far more than that, and many brands fill the slots with variations of the packshot, which adds files without answering any new question. A fifth image showing the piece against a competing size, such as a stacked set or the full collection, is worth more than a fourth angle of the same band.
Why does jewelry get returned even when the photos are accurate?
Usually scale, occasionally colour. A macro shot with no size cue leaves the buyer to guess, and buyers guess large because the image is large on their screen. Colour is second, because screens render metal tone and stone hue inconsistently. A scale shot fixes the first. A colour reference frame shot at capture time reduces the second.
Sources
- FTC Jewelry Guides, 16 CFR 23.0 and 23.1, including Note 3 to §23.1 (law.cornell.edu, reproducing the eCFR, accessed September 2026)
- Amazon Seller Central, product image guide (sellercentral.amazon.com, accessed September 2026)
- Etsy Seller Handbook, 6 tips for selling accessories (etsy.com, accessed September 2026)
- Etsy Seller Handbook, 4 top tips for selling clothing (etsy.com, accessed September 2026)
- Etsy Seller Handbook, 5 tips for writing product descriptions (etsy.com, accessed September 2026)
- Etsy Help, requirements and best practices for images (help.etsy.com, accessed September 2026)
This is general information, not legal advice. Individual results vary.
