Separate the person, lighting, lens-like appearance and background.
Suggested reading time: about 8 minutes. This is an editorial production guide. Generation, performance and pricing comparison experiments have not been run.
Read first: Turn visual direction into words and specifications
What you will learn
- Evaluate realism through more than skin detail.
- Translate photographic intent into a prompt.
- Check how human subjects are used and presented.
Realism depends on consistent relationships
More pores alone do not make an image realistic. Relationships must fit: the direction of the face and light, hair and shadows, hands and objects, background and focus. Begin with fewer relationships, such as soft window light, a simple background and a chest-up composition, to make defects easier to find.
Connect camera terms to their purpose
Terms such as “85 mm” and “shallow depth of field” should express an intention: avoid an extreme wide-angle appearance of the face, or blur the background to separate the subject. In a generated image, these terms are not evidence that a real lens was used. Even when the appearance suggests a particular focal length, inspect the result rather than claiming a precise photographic simulation.
Avoid misleading presentation when publishing
When using a photograph of a real person as a reference, check that person’s consent and the asset’s usage terms. Present fictional people as fictional; do not use them as if they were testimonials from real customers or employees. Check the publishing platform’s disclosure requirements for AI images. The exercises here use fictional adults and do not require reproducing a real person.
Inspect relationships in the portrait
- 1Purpose and fictional subject
- 2One main light and composition
- 3Candidate
- 4Inspect face, hands and shadows
- 5Correct a region
- 6Review final presentation
This original inspection diagram connects the photographic intention to the public portrayal. Regional editing is followed by another full-image check because a corrected detail may disagree with the remaining light or pose.
Work and responsibilities
- Photographer: makes composition and lighting decisions.
- Retoucher: adjusts skin, color and unwanted objects.
- Brand reviewer: checks the portrayal and permitted uses.
From input to output
| Input | Workflow | Output |
|---|---|---|
| A fictional adult subject.; The intended use and expression.; The background, clothing and light direction.; Permission and provenance if a real-person reference is used. | 1. Choose an expression and composition suited to the purpose.; 2. Generate candidates with one main light source.; 3. Inspect the eyes, teeth, ears, hands, accessories and shadows.; 4. Edit only the parts needing correction.; 5. Check the final screen and how the image will be presented publicly. | A portrait asset.; Reasons for selection and edits.; Notes on permitted uses and disclosure. |
Quality checks
- The eyes and accessories are consistent on both sides.
- Light direction agrees across the face, neck and hands.
- Check for unnaturally uniform skin or teeth.
- Do not present fictional people as real-world facts.
Diagnose failures
| Symptom | Possible cause | Correction |
|---|---|---|
| Skin looks like plastic. | Excessive smoothing, uniform light or an unbalanced emphasis on texture. | Retain natural variation in the skin and simplify the lighting. |
| Hands are malformed. | The relationship between the pose and held objects is complex. | Simplify the pose, make a local correction and inspect again. |
Exercise: Compare lighting for the same fictional person
Status: not run (not-run). This is a plan with completion criteria. No artifacts, duration, cost, performance or success rate have been recorded.
Plan two versions: soft window light and hard side light. When running the comparison, keep the model, reference and composition aligned, and evaluate the lighting impression separately from defects.
Deliverable: Photographic intent and a comparison record.
Completion criterion: You can describe personal preference separately from structural defects.
When you run it, separate predictions from results and record conditions and evidence using the experiment worksheet.
Related Catalog candidates
- Attractive adult woman in natural light (Catalog candidate:
photo-woman-natural) - Handsome adult man in natural light (Catalog candidate:
photo-man-natural) - Beauty portrait with natural skin texture (Catalog candidate:
photo-beauty-editorial) - Photorealistic landing-page image with copy space (Catalog candidate:
photo-hero-copyspace)
Technical references and evidence scope
Technical references were checked on 2026-10-03. They support the technical concepts. The exercises, workflows and evaluation criteria are editorial proposals, rather than experiments performed by the source authors. Recheck changing APIs, prices and publication rules when using them.
- OpenAI Image generation guide — Image generation and editing, output settings, and limits on consistency and precise placement.
- YouTube GenAI content disclosure — Disclosure for realistic generated content or meaningful alterations.
Source scope checked on 2026-10-04
YouTube’s disclosure guidance distinguishes realistic generated content and meaningful alterations from minor edits. Decide from the actual published scene, rather than treating every production-assistance task as the same case.
MENTAL MODEL / REASONING ORDER
Change one thing at a time.
Decide what the image communicates, its medium and dimensions, and the meaning you want to preserve.
Sources
Publication dates belong to the source; access dates record when it was checked. Community observations are separate from official statements.
01