Sharpness analysis scans every frame and flags sections where focus hunting, subject blur or camera movement has softened the image below broadcast quality.
The camera failed to achieve sharp focus on the primary subject. Common with autofocus hunting in low light, very shallow depth of field, or manual focus racks that missed the mark. These sections score the highest blur severity.
Autofocus oscillates through focus as the lens searches for the correct plane. This produces a rhythmic in-out blurring pattern that's detected by analysing the variance of sharpness scores across a short window. Particularly common on mirrorless cameras in continuous AF mode.
Excessive shutter speed (180-degree rule violated) or fast subject movement creates directional blur that softens the frame. Detected separately from focus-based blur by analysing the directionality of the low-sharpness regions in the Laplacian output.
The default sharpness threshold is calibrated for broadcast-quality delivery. Narrative fiction editors working with intentional shallow-focus shooting may want to raise the threshold to avoid false positives on creative rack focuses. Documentary and corporate editors working to strict QC standards can lower it.
Sharpness analysis runs at a sampled frame rate (default every 6 frames) to keep analysis fast. Full-frame analysis is available for critical QC at the cost of longer processing time.
EditLint AI decodes frames from your source media at the configured sampling rate. No render or export needed — the decoder reads directly from your NLE's linked media files. For a 10-minute clip at default settings, roughly 1,000 frames are sampled.
Each sampled frame is analysed using a Laplacian variance operator — the standard computer vision method for measuring image sharpness. The variance of the Laplacian kernel response is computed across the analysis region (centre, face area, or full frame). Low variance means a soft, blurry frame. High variance means a sharp, in-focus frame.
Consecutive low-scoring frames that exceed the minimum duration are grouped into a blurry section. Each section receives a severity classification. Markers are placed on the timeline with the sharpness score, detected blur type and a suggested action in the comment field.
Sharpness analysis catches every soft-focus section automatically. You decide whether to cut, stabilise or cover — EditLint finds the problems.