Skip to main content
  • Pricing
AI video

AI Video Defocus Benchmark for an Unsuspecting Subject

By Oakgen TeamPublished 1 min read

Share
AI Video Defocus Benchmark for an Unsuspecting Subject

Lens defocus can help a cut when it moves toward an intended subject. A clip may instead look like its whole image loses fine detail briefly. Test reference markers at different distances so the difference is measurable. Make focus score independent of color or motion preference.

Build the test input

Create a still with a sharp foreground fabric edge, a middle paper bookmark, and a distant plain shape. Ask for the foreground to soften and the bookmark to become sharp during a short hold. Keep all objects and camera positions static. Save a copy marking where each focus region sits.

Run the proposed comparison

Render a controlled condition with focus change and a separate condition with a fixed-focus brief. Export both and inspect at the same playback scale. Sample the fabric fibers and printed lines near the bookmark. Compare whether sharpness moves in an expected order rather than dropping across every element simultaneously.

Decide what passes

Pass when one region becomes clearer while the initial region softens, with stationary geometry that proves the image was not just zoomed or morphed. Reject a global blur with unclear destination. Record the exact frame where the destination becomes sharp enough to read for editing purposes.

Act on the outcome

If the generated motion is strong but focus looks uniformly soft, use an editor depth or mask technique on a stable plate where available. In a shot that depends on focus moving attention, select the model with the more stable visual transition even if another creates more dramatic background treatment.

Model references

Share