Micru Logo

Diff Statistics

Get a complete quantitative comparison of two images: pixel-level differences, dimensions, file sizes, perceptual hash values, and visual similarity score.

Image A

Click or drag image here

PNG, JPG, WebP, GIF

Image B

Click or drag image here

PNG, JPG, WebP, GIF

5

Understanding the Metrics

Pixels Changed

The number and percentage of pixels whose Euclidean sRGB distance exceeds the configured tolerance. This metric is sensitive to compression, rendering engines, and font antialiasing. Use a tolerance of 3-10 to filter noise from JPEG encoding.

Perceptual Hash (pHash)

A 64-bit hash derived from the low-frequency DCT components of the image. The hash is stable under minor resizing, light compression changes, and small brightness shifts. Two identical images always produce the same hash. The hash is displayed as a 16-character hexadecimal string.

Hamming Distance

The number of bit positions where the two pHash values differ (0-64). A distance of 0 means the images are perceptually identical. Distances up to 10 typically indicate the same image saved at different quality settings or slightly different dimensions. Distances above 20 usually indicate substantially different images.

Visual Similarity

A derived percentage calculated as (64 - hammingDistance) / 64 * 100. 100% means identical pHash values. This is an approximation of human-perceived similarity for the overall tonal and structural content of the image, not a pixel-perfect measure.