Focus Stacking Food Photos: Keep the Dish Sharp Without Inventing Detail
Use focus stacking when a stationary dish needs more front-to-back sharpness than your preferred single exposure provides. Photograph the same composition at several focus distances, then combine the usable sharp areas in a stacking editor. The result should preserve the food you photographed, including its actual ingredients and portion.
A stack needs captured detail. If an important area is soft in every source frame, merging those frames cannot supply a faithful sharp version of it. Before starting, decide whether blur needs a reshoot: camera shake, subject movement, and incomplete focus coverage require different corrections.
Try one frame before planning a stack
Define what must be readable in the menu image. A garnish in the background may not need the same sharpness as the main ingredient. Review a single exposure at the intended output size before adding a composite workflow.
If your camera allows aperture control, test a smaller aperture while keeping the composition consistent and compensating exposure appropriately. Check the resulting image rather than assuming the smallest available aperture will look best. There is no universal aperture or frame count for every dish, lens, and shooting distance.
| What the photograph needs | First option to try | Reason to change approach |
|---|---|---|
| One clear hero ingredient with a soft background | A single focused frame | Required ingredients remain unreadable |
| An entire stationary plate in focus | A deeper-focus single exposure | Important near and far details cannot both meet your needs |
| A tall, still arrangement with several important depths | A planned focus stack | Food changes position during the sequence |
| Pouring sauce, lifting noodles, or visible moving steam | A single timed exposure | A stack would combine different moments |
Keep the best single frame as a fallback. A clean single exposure can be more useful than a composite with distracting seams.
Capture a sequence you can audit
Set the plate, utensils, and lighting before taking the sequence. Support the camera securely, leave the viewpoint unchanged, and keep exposure and white balance consistent. For an overhead composition, keep the camera level and fixed before starting; do not correct the camera angle halfway through the stack.
Make a simple focus map of the dish. Identify the nearest required detail, the middle ingredients, and the farthest required detail. Work through those depths in order, changing focus while leaving the camera and plate in place.
Review adjacent frames at a useful magnification. Their acceptably sharp areas should overlap, with no missing strip between them. Add an intermediate capture when a required region is not covered. The number of positions comes from that review, not a preset promise that three or five photographs will cover any plate.
For example, consider a hypothetical layered sandwich photographed at an angle:
| Focus region | Detail to inspect in the source | Sequence problem to watch for |
|---|---|---|
| Front | Cut bread edge and nearest filling | Foreground edge never becomes sharp |
| Middle | Main filling and central bread surface | A soft gap remains between focus positions |
| Back | Rear crust and required garnish | Garnish shifts before the last capture |
These are inspection regions, not instructions to take exactly three frames. If the filling slips or a leaf falls, stop and establish a new consistent sequence. Avoid mixing captures from before and after rearranging the food.
Align and blend in a stacking editor
Adobe documents a Photoshop workflow for extending depth of field: load the photographs as layers, select them, run Edit > Auto-Align Layers, then Edit > Auto-Blend Layers and choose Stack Images. Its documented loading route uses Adobe Bridge's Tools > Photoshop > Load Files into Photoshop Layers command. These are Photoshop controls, not DishVivid controls. Adobe's extended-depth-of-field procedure
Keep an editable copy with the source layers. Treat the first blend as a candidate. Alignment does not establish that every ingredient stayed still or that every selected edge is correct.
Compare the candidate with individual captures before flattening it. If a seam needs correction, use a genuinely sharp source area that matches the dish's geometry. When no suitable source exists, return to the single-frame fallback or reshoot. Do not replace missing texture with invented seeds, crumbs, or filling.
Accept the food, not just the sharpness
Inspect the composite both enlarged and at delivery size. Use this checklist before export:
- Plate outline: Is there one continuous rim, without doubled edges or local bends?
- Ingredient boundaries: Do leaves, bread edges, and thin stems retain their original shapes?
- Depth transitions: Is there a soft band or halo between otherwise sharp areas?
- Sauce and reflections: Are highlights continuous, without abrupt patches from different frames?
- Moving elements: Has steam become broken or repeated? Has a garnish acquired a ghost edge?
- Portion and identity: Does the composite show the same ingredients and quantities as the actual serving?
Use the original captures to check ingredients and portion size, especially where automatic blending crosses overlapping food. Reject a technically crisp result if it duplicates a topping or changes the visible portion.
Keep a clear handoff
Retain the original sequence, the layered composite, and the approved rendered export separately. Name them so another team member can identify which serving and capture sequence produced the finished image.
After completing the stack externally, upload a JPEG, PNG, or WebP copy up to 20 MB to DishVivid Studio for brightness, contrast, saturation, or crop. Studio does not align frames, stack focus, accept camera RAW, or export a layered file; its reviewed download is JPEG. Compare any later edit with the approved composite and real dish. Adobe's linked Photoshop procedure was checked October 10, 2026; the capture sequence and inspection rubric are editorial suggestions.