Include accepted finish variation as well as confirmed defects through the intended presentation range. Making the rejected plate look obvious is not enough if ordinary brushed finish becomes a reject everywhere else.

On an inherited reference set, we needed to distinguish the original physical judgement from later handling damage. Name who keeps these plates and when a reference must be re-examined, rather than assume its first label lasts forever

Quality has the retained plates and their region judgements now. Two accepted finish examples and one rejected nick are identified. We have kept the original guessed captions separately instead of quietly replacing the history.

Keep holder changes separate from lighting changes in the comparison.

And record the intended seat. If the plate can rest in two positions, an improved image might come from a different physical presentation rather than a better illumination condition.

Tooling confirmed the intended seat and we repeated it with the identified plates. The nick is visible in the current view, but an accepted brushed region sometimes saturates. Lighting comparison is next, not a holder redesign.

Retain the nick in every proposed lighting condition. In my small bezel study, keeping the real defect visible was the useful question; reducing brightness alone would have been too easy to call success.

Has the imaging owner recorded the required regions as well? A less saturated picture can still lose coverage near the edge if the view changes along with the light

Yes. Initial lighting comparison reduces saturation on the accepted regions without losing the identified nick in those tested views. Quality accepted that limited reference comparison; broader part variation and commissioning remain separate.

Keep the tested conditions and limits with those results. That gives the next imaging stage a useful starting point without turning three reference plates into a production performance claim.