T-shirt product photography: flat, ghost, or on a body

A t-shirt looks like the easiest product in the store to photograph and quietly punishes the attempt — wrinkles read loud on jersey, the print carries the sale, and the bestselling color fights the standard background. Here is when flatlay, ghost mannequin, or on-model earns its cost, and how to keep a thirty-design drop looking like one catalog.

Key takeaways

  • Jersey broadcasts every wrinkle, the print carries the purchase decision, and white tees fight white backgrounds — the three problems that make tees deceptively hard to photograph.
  • Flatlay is the volume default for graphic tees, ghost mannequin earns its cost where the cut is the selling point, and on-model belongs on heroes — not the whole catalog.
  • Print accuracy is the non-negotiable: inspect every generated tee at 100% zoom and lock the artwork and label with fabric and tag references.
  • A thirty-design drop stays consistent by shooting one framing, generating one style with one instruction, and verifying each print individually.

Why a t-shirt is harder than it looks

A tee is the simplest garment in most catalogs and one of the least forgiving in front of a camera. The difficulty hides in three places, and all three punish exactly the casual phone shot a tee seems to invite.

  • Jersey broadcasts every wrinkle. Cotton jersey is a thin, structureless knit — it holds whatever shape it was last crumpled into, and light coming from the side turns each fold into a hard shadow line. A wool blazer can carry a soft crease and still read as tailored; the same crease on a tee reads as pulled from the laundry pile. Worse, jersey re-wrinkles as you handle it, so the steaming you did ten minutes ago is partly undone by the time the frame is set.
  • The print is the product. On a graphic tee the blank is close to a commodity; the customer is buying the artwork — this artwork, at this size, in this position on the chest. A photo where the print is slightly soft, warped over a fold, or a shade off in one ink is a photo of a different product, and it comes back as one. No other garment concentrates so much of the purchase decision into a few hundred square centimeters of ink.
  • White on white disappears. The bestselling tee color is the hardest to photograph for the standard listing format. Expose bright enough for a clean background and the shirt's edges dissolve into it; expose to hold the edges and your "white" tee publishes as grey. The white background guide covers the separation problem in depth — for tees it is not an edge case, it is the bestseller.

Prep that actually changes the photo

Ten minutes of physical prep does more for a tee photo than any edit afterwards, because two of the three problems above are solved at the garment, not the camera.

  • Steam last, shoot immediately. Steam the shirt after it is already positioned on the shooting surface, or carry it there flat on both forearms like a tray. Fold it over one arm on the way and you have re-creased it. Jersey holds a press for minutes, not hours — the steamer belongs next to the table, not in another room.
  • Lint-roll under raking light. Dark jersey collects lint and hair that is invisible face-on and obvious the moment light skims across the fabric — which is exactly what even studio-style lighting does. Roll the shirt, tilt it against the light to check, then shoot. Retouching forty specks off a black tee later is slower than thirty seconds with the roller.
  • Shape it symmetrically. Sleeves folded at matching angles, side seams straight, hem parallel to the bottom of the frame, collar a smooth even ring. The collar is the first place the eye lands and the first place asymmetry reads as carelessness. A tee has perhaps six visible lines — make them agree with each other, and make them agree across every shirt in the session.
  • Get the price tag out of the frame. Tuck it inside the collar or snip it — a swing tag on its string is the loudest "shot in the stockroom" signal a listing can send.

Flat, ghost, or on a body

There are three honest ways to present a tee, at three costs. Most catalogs should pick one as the default for every listing and reserve a second for heroes — running all three on every design is a campaign budget, not a catalog workflow.

  1. Flatlay: the volume default. The shirt laid flat, pressed, shaped, shot straight down. It is the fastest style to produce at scale, and for graphic tees it carries a quiet technical advantage: a flat shirt shows the print flat — actual size, actual placement, no distortion from drape. The flatlay guide covers the shaping and lighting in detail. For a print-led catalog this is the style that belongs on every listing.
  2. Ghost mannequin: shape without a model. The invisible-mannequin look shows the tee holding its worn silhouette — where the shoulder seam falls, how much the body boxes out, how the hem hangs. On a basic blank it is more than the listing needs. On a heavyweight 240 gsm tee with a drop shoulder, the cut is the product, and a flat photo hides exactly the thing the customer is paying extra for. The ghost mannequin guide explains the look; it earns its cost precisely where fit, not print, closes the sale.
  3. On-model: context, at the highest cost. A body answers questions no other shot can — where the hem lands, how the fabric moves, what size the person in the photo wears. It also costs the most: a person, a session, and a re-booking every time the line grows. And it is the hardest style to keep consistent, because the same tee on two different days with two different people photographs as two different products. Use it for hero images and lifestyle content, not as the workhorse.
StyleWhat it shows bestCost to produceWhere it earns it
FlatlayThe print — flat, undistorted, at real placementLowest: a table, window light, a phone held levelThe default for graphic tees and volume catalogs
Ghost mannequinCut and drape — the worn silhouette with nothing in itTraditionally a mannequin shoot plus compositing; with generation, produced from a flat photoPremium blanks where fit is the selling point
On-modelFit on a real body, sizing reference, styling and moodHighest: a person, a session, and a re-booking per dropHero and lifestyle images — not all thirty listings

The print is the one thing you cannot let drift

Generation handles the presentation half of this guide well — pressing the wrinkles out of jersey, converting a flat phone photo into a ghost mannequin shot, lifting a white tee cleanly off a white background. PhotosynthAI does that work inside the Shopify admin, on photos already uploaded to the product, in under a minute per photo. The honest caveat concentrates in one place: the artwork.

The model is instructed to preserve color, stitching, and printed branding exactly, and on a bold three-color chest print it generally does — the artwork is in the source photo, so nothing has to be invented. But instructed is not guaranteed, and the failure mode is specific: small dense lettering and fine linework drift first. A serif thickens, a hyphen disappears, a line of six-point text becomes plausible instead of exact. On a garment where the print is the product, plausible is wrong. Four habits close the gap:

  • Inspect every generated tee at 100% zoom, on the print. Read every word of the artwork out loud. Count its elements — stars, letters, outline strokes — against the real shirt in your hand. This takes under a minute and it is the entire difference between publishing a photo of your product and publishing a photo of something similar to it.
  • Lock the artwork with references. Up to three reference photos ride along with generation: a fabric reference locks the body color and the ink colors so sage does not return grey, and a tag reference reproduces your real neck label as it actually reads instead of as approximate lettering. For ghost shots, a fit reference locks the silhouette to your actual cut rather than a generic one.
  • Refine instead of settling. A result that is right except for one thing does not need regenerating from scratch — a custom instruction like "softer shadow" or "pure white background" adjusts it, and every version is kept side by side so you pick rather than gamble.
  • Dense text gets a real close-up. If a design carries paragraph-sized copy — a manifesto print, a care-instruction parody — put a straight photo of the print itself in the listing as a secondary image, whatever the primary is. That is the same rule that applies to logos and labels everywhere, and the honest assessment of AI product photos covers the full pre-publish checklist.

Thirty designs, one catalog

Print-on-demand and drop-based tee brands hit a problem a general boutique never sees at this intensity: thirty designs on the same blank need thirty photos that look like one photographer shot them in one session. A collection grid where framing wanders, whites shift between cream and blue, and shadows fall three different ways reads as a reseller, not a brand — and with tees the grid is the storefront.

  • Remove every variable except the print. Same blank, same surface, same distance, same shaping, one session per drop. When the only thing that changes between frames is the artwork, the grid does the selling for you.
  • Generate in batches, in the same run. From the product in the admin, queue up to ten photos per run at under a minute each — a thirty-design drop is three runs, not thirty separate editing sessions. Photos processed together, in the same style, come back looking like siblings.
  • Reuse the exact instruction wording. If design one was finished with "pure white background, soft shadow low and to the right", designs two through thirty get the same words. Consistency lives in the repetition.
  • Verify per design, not per drop. The consistency check is one look at the grid; the accuracy check is per shirt, because every artwork fails differently — the fine-line design drifts where the bold one cannot.

The arithmetic works out cleanly for a drop this size. The Starter plan at $29 per month covers 50 photos — thirty designs plus twenty retries and refinements — and it is the plan with the 7-day free trial, so the worst photo in the current drop is a free test. For a one-off launch without a subscription, a one-time 50-photo pack costs $25 — the pricing page has both.

Common questions

Should I photograph t-shirts flat, on a ghost mannequin, or on a model?

Pick one default and one hero style. Flatlay is the volume default for graphic tees because a flat shirt shows the print at its real size and placement, undistorted by drape. Ghost mannequin earns its extra effort on premium blanks where the cut — a drop shoulder, a boxy body — is the selling point. On-model costs the most and is hardest to keep consistent, so reserve it for hero and lifestyle shots rather than the whole catalog.

How do I photograph a white t-shirt for a white background listing?

It is a separation problem: expose bright enough for a clean background and the shirt's edges dissolve into it; expose for the edges and the shirt publishes as grey. Keep a soft contact shadow under the shirt so it reads as an object, and judge the result next to the real tee — the fabric should read white, not silver. Generating the studio white version from a decent phone shot handles the separation, but the same check applies before publishing.

Will AI generation keep my t-shirt graphic accurate?

Only with checks. The model is instructed to preserve printed branding exactly, and a fabric reference locks the body and ink colors while a tag reference reproduces your real neck label — but dense small lettering and fine linework are where generation drifts first. Inspect every generated tee at 100% zoom, read each word of the artwork, and count its elements against the real shirt. If a design carries paragraph-sized text, keep a real close-up as a secondary image.

How do I keep photos consistent across a thirty-design drop?

Remove every variable except the print. Shoot each design on the same blank, surface, and framing in one session, then generate one style with the same written instruction for all of them — PhotosynthAI processes up to ten photos per run at under a minute each, so a thirty-design drop is three runs. Consistency you check once across the grid; print accuracy you check per design, because every artwork fails differently.

Thirty designs, one look.

Pressed flatlays and ghost mannequin tees from the phone photos already on your products — with references keeping every print exact.

Add to Shopify — free trial