Even skilled editors miss problems: a halo around a building, banding in a sky, a cloned pattern, a cut-out edge, a strange AI-generated hand. Once delivered, printed or published, these defects reflect on your professionalism. Quality control is the final stage of editing, where you deliberately look for problems before anyone else sees them. This lesson lists the most common editing defects, explains how to find them and gives a practical checklist you can use on every job.
What is quality control?
Quality control (QC) is a systematic final check of images for technical and editing defects, consistency and correct output, before delivery. It complements the editing philosophy in editing philosophy and non-destructive workflows and the delivery steps in client delivery and file packaging.
Common editing defects
| Defect | Usual cause | Fix | Lesson |
|---|---|---|---|
| Halos around edges | Strong local contrast, clarity, HDR or sharpening | Reduce strength; refine masks | sharpening |
| Banding in gradients | Heavy edits on 8-bit files, strong compression | Edit in 16-bit, add fine grain | bit depth and digital tonality |
| Clipped highlights or shadows | Over-brightening or over-darkening | Check histogram and clipping warnings | levels and tonal adjustment |
| Color fringes and cut-out edges | Imperfect masks, color spill | Refine edges, decontaminate color | background removal |
| Clone and heal repeats | Same source used repeatedly | Vary sources; rework | healing and cloning |
| Plastic skin | Over-smoothing | Restore texture | advanced retouching |
| Oversaturation and odd skin color | Strong saturation or grading | Reduce, check skin | saturation and vibrance |
| Noise or over-smoothed detail | Too little or too much noise reduction | Rebalance | noise reduction |
| Remaining dust spots | Missed during cleanup | Spot heal with visualisation view | healing and cloning |
| Mismatched composite elements | Light, contrast or grain differences | Match systematically | compositing and image construction |
AI-specific artifacts
Generative and AI tools can introduce defects that traditional editing does not:
- Anatomy errors: extra or merged fingers, strange ears, distorted teeth.
- Garbled text and logos in generated or extended areas.
- Repeated or impossible patterns in fills and extensions.
- Mismatched light and perspective in generated objects.
- Over-smooth or invented textures from AI noise reduction or upscaling.
Warning:Inspect every AI-edited area at 100%. AI results can look convincing at small size and fail at print size.
A quality-control checklist
- View the whole set together for consistent color, brightness and style.
- Check each image at viewing size: composition, crop, horizon and overall look.
- Inspect at 100%, scanning systematically across the frame for dust, halos, banding, edges and repeats.
- Check skin, eyes and hands in people images.
- Check histograms and clipping.
- Check output settings: size, color space, sharpening and file format, as in export settings for web and social media and export settings for print.
- Check metadata and file names.
- Review on a second device or ask someone else to look.
Tip:Rest your eyes before QC, or check the next day. Fresh eyes notice what tired eyes accept.
Viewing conditions
Check work on a calibrated screen in consistent light, as in monitor calibration and consistent viewing. For prints, make a test print or soft proof, as in soft proofing and preparing for print. Many defects, such as banding and halos, are more visible in print or on large screens than on a laptop.
Common mistakes
- Skipping QC under deadline pressure: errors reach clients.
- Checking only at small size: defects hide.
- Not checking output files: wrong size, color space or compression.
- Trusting AI results without inspection: visible artifacts.
How professionals control quality
Professional studios build QC into every job with checklists, calibrated screens and sometimes a second person reviewing work. They inspect at 100% and at final size, check output files rather than only working files, and keep notes of recurring problems so they can prevent them. You have now completed the editing, retouching and advanced image processing course. Next, you will learn how to prepare photographs for print and display, starting with preparing photographs for print.
Practical examples
A banded sunset print
- Situation
- A large print shows visible steps in the sunset sky.
- What to do
- You re-edit from RAW in 16-bit, reduce extreme adjustments and add fine grain before printing.
- Why it works
- Banding comes from stretched tones and limited bit depth.
- Result
- A smooth gradient in the new print.
An AI-extended product image
- Situation
- An AI-extended background looks fine on screen but shows garbled text on a sign at 100%.
- What to do
- You remove the generated sign and rebuild the area manually.
- Why it works
- AI artifacts often appear only at full resolution.
- Result
- A clean image safe for large print.
Key points
- Quality control is the final, systematic check before delivery.
- Common defects: halos, banding, clipping, fringes, repeats, plastic skin, oversaturation, dust and composite mismatches.
- AI tools add anatomy errors, garbled text, odd patterns and invented textures.
- Use a checklist: set view, viewing size, 100% scan, people details, histograms, output, metadata, second view.
- Check on calibrated screens and test prints, with fresh eyes.
Frequently asked questions
What should I check before delivering photos?
Why do halos appear in edited photos?
How do I spot AI artifacts?
Why do problems appear in print but not on screen?
Conclusion
Quality control protects your work and reputation: know the common defects, inspect systematically at 100% and final size, check output files and AI results, and review with fresh eyes. You have now completed the editing, retouching and advanced image processing course.
Finished reading? Track your progress through Editing and retouching.