How to use Image Quality and Blur Checker
- Choose up to 20 JPEG, PNG or WebP images within the combined limits. Use images with similar content and scale when comparing measurements.
- Set Texture review cutoff. This is the only adjustable threshold; dark/light pixel boundaries are fixed.
- Read the ranked table alongside the numbered thumbnails. Rank is the sorted position; Input number identifies the original file-list position.
- Inspect the original images at their intended size. Download the JSON record if you need the measurements and chosen cutoff for a later comparison.
Example: Image Quality and Blur Checker
An 80 × 60 PNG with a solid red left half and solid green right half, measured at the default cutoff of 100.
Options
- Threshold
- Texture review cutoff is your rule for measured Laplacian variance. It cannot certify sharpness or quality; a smooth scene can have a low value without blur.
- Sampling
- Images are reduced to the documented longest side before measurement. Scale, texture, noise and transparency can change the result.
Supported inputs and limits
- Up to 20 JPEG, PNG or WebP files, 64 MiB each, 128 MiB and 24 megapixels combined.
- Each measurement raster has a longest side of at most 512 pixels and must provide a usable neighborhood of at least 3 × 3 pixels.
- Alpha below 242/255 is excluded; a Laplacian sample is omitted if it or any of its four neighbors is excluded.
- The metric is population variance of the four-neighbor Laplacian on gamma-encoded grayscale.
- Dark is below 30 and light above 225 on the 0–255 gray scale; those boundaries are fixed.
More details
The entered cutoff is an uncalibrated review aid, not proof of blur or quality.
Noise, texture, sampling and alpha change the numbers.
Thumbnails label input positions, while table rows sort by decreasing metric.
Where your input is processed
This tool does not upload your inputs. See the limits for anything saved on your device.
Compare like images under the same conditions
A noisy photo can have a high edge metric without useful detail. A smooth subject can have a low value despite being in focus. Use the measurement to order a manual review, then inspect the original at its intended size.
Questions about Image Quality and Blur Checker
Does a low metric always mean an out-of-focus photo?
No. A smooth sky or other textureless subject can also have little edge variation.
Can I compare values from differently sized images?
Sampling and scale affect the measure. Compare results under the same stated conditions and inspect the images.
Do dark pixels prove underexposure?
No. A dark subject can be intentional. The fraction is a review cue, not a camera diagnosis.