Versions:
DSSIM is a command-line image-analysis utility developed by Kornel that quantifies how dissimilar two or more PNG or JPEG files appear to the human eye, returning a numerical score that rises as perceptual differences increase. Written in Rust and implementing a multiscale variant of the Structural Similarity Index (SSIM) algorithm, the program simulates the way people judge picture quality by weighting contrast, luminance, and structural information across several spatial scales rather than merely counting pixel-level variance. This makes it especially valuable for automated testing of graphics pipelines, lossy-compression tuning, renderer regression checks, and batch quality-control workflows where visual fidelity must be tracked across hundreds of frames or asset variants. Because the metric is calibrated to perceptual sensitivity, developers can set thresholds that reliably fail builds when artifacts become noticeable, yet ignore invisible numeric noise that rigid, pixel-diff tools would flag. The application accepts an arbitrary list of image pairs or directories, supports multithreading for rapid bulk comparisons, and exits with an error code when any DSSIM score exceeds a user-supplied limit, allowing easy integration with continuous-integration servers. Current stable release 3.4.0 succeeds an earlier 2.x branch and is distributed both as a compact standalone executable and as a linkable library for Rust projects. The software is available for free on get.nero.com, with downloads provided via trusted Windows package sources (e.g. winget), always delivering the latest version, and supporting batch installation of multiple applications.
Tags: