Documented preview case

Strong shadow

Shadow-to-depth confusion caught in the free preview.

免费预览中发现的阴影被当成深度。

What the preview returned

Every file below came from one free preview run on the licensed source image. The numbers quoted are the values the checks reported, plus measurements taken from the returned depth field. No physical result is claimed for this case.

Licensed source image for the Strong shadow case — Source image. Photo by Shixart1985, CC BY 2.0, via Wikimedia Commons License

Approved depth output for the Strong shadow case — Approved depth output, exported before the job was approved.

Mesh check for the Strong shadow case — Workbench check view captured during the same preview run.

Source readiness at approval

Reported 1920 px input, 97% usable tonal range, edge detail 0.008, and border clutter 0.191, which passed. A source like this looks acceptable to every cheap check, so the dispute is not whether it is sharp but whether the returned field describes shape or light.

Manufacturing checks at approval

Reported 1.68 mm tool-to-grid detail, 61 effective tonal slices, and a maximum relief slope of 85.1°. Measuring the returned field against the source afterwards: brightness and depth correlate at r = 0.580 (Spearman 0.420) over 512 x 341 samples, with the field using the full 0-254 range. Strong sunlight and its cast shadows therefore carry substantial weight in the estimate, which is the confusion this case exists to show, and why direction has to be confirmed by eye before approval.

Failure notes

No physical result is claimed for this case. It documents the free preview and a measured correlation between source brightness and the returned depth field.