A small, self-contained Python script that estimates how soft a frame is WITHOUT a reference image, using three classic no-reference measures: the Crete-Roffet perceptual blur metric (re-blur the frame and measure how much local variation drops — 0 sharp to 1 blurred), the mean edge width in the spirit of Marziliano et al. (blur spreads edges), and the variance of the Laplacian (a focus measure). Given two renditions of the same content (two encodes, or source vs encode), it reports which kept m
Download free PDF
The handbook · 7 volumes
Fora Soft Video Engineering Handbook
This blur detector scores frame softness with no reference image, using the Crete-Roffet metric, edge width and Laplacian variance to compare two renditions. Our seven-volume handbook covers the full video stack; start with Volume 1 for compression artifacts, rate control and quality metrics, then Volume 3 for delivery and Volume 6 for AI-based quality models. Written by Nikolay Sapunov, CEO at Fora Soft.
Losing detail in your encodes and not sure which rendition is softer?
Send us two renditions of the same clip, and we will measure blur, find where detail is lost and suggest encoder settings that keep it. 250+ video products shipped since 2005.
You get a realistic estimate, technical recommendations and next steps. Free, no obligations.
