Zenmetrics

Compare image quality and explore spatial error. Images stay on your device.

SourceDownloads

Images

40

JPEG, PNG, WebP and JPEG XL use Zen decoders. SDR RGB8 ingress matches the CLI: transparency is preserved and scoring refuses non-opaque pixels. ICC metadata is not applied. Use already composited sRGB images for color-managed comparison.

Metric

Score–
Compute time–

Reference, score and map timings appear after comparison.

Worker startup: starting…

Checking browser capabilities…

Brave: turn Shields off · Browser requirements

In Brave, turn Shields off for this site, then reload. Relaxed SIMD128, shared memory and working Web Workers are required. HTTPS and cross-origin isolation enable the Rayon worker pool. WebGPU readiness is checked separately; the selectable metrics in this build run on the CPU.

Starting threaded SIMD WASM…

Run the same native metric

Use the same opaque sRGB files. Generated JPEG candidates can be downloaded below. Zensim scoring and its map use the selected SDR model.

Image comparison

100%
LowHigh error
Load a pair or try the demo sceneMove over the image to inspect a pixel

Map scaling changes only the visualization. Scores and downloaded raw values retain their metric definitions. Native similarity maps retain their metric sampling grids; visualization uses 1−similarity. VSI displays saliency-weighted quality contributions. FSIM/FSIMc maps show luminance similarity; MS-SSIM shows level-0 SSIM. Hatched cells mark uncovered borders or undefined native values; HaarPSI has undefined local similarity in zero-weight windows. NLPD, PSNR-HVS, MAD, VIF and VMAF maps show one native component of their score (NLPD level 0, PSNR-HVS per-block CSF error, MAD's high-quality strategy, VIF's finest information loss, VMAF's SPEED chroma-U blocks), not a per-pixel score; IW-SSIM shows scale-0 contrast-structure and mDCT-PSNR each window's detection probability. IW-SSIM maps need both sides of at least 176 pixels; VMAF v1 needs even dimensions, both sides of at least 33 and one of at least 216. SSIMULACRA2 currently exports a scalar only. Browser timings are interactive feedback, not accepted fleet benchmark observations.