ImageReasonLocal Image Analysis Lab
How much fine-scale variation is in this flat region?
Choose two visually flat regions in your local image and compare their measured fine-scale variation, a bounded proxy, not a noise or ISO diagnosis.
Answer
Your image is decoded and analyzed in this browser tab only. It is never uploaded, transmitted, or stored anywhere else, and nothing about it reaches analytics.
Visualize
| Region | RMS residual | Robust MAD |
|---|
Plan
Analyze an image to generate a plan.
Fine-scale variation is not proof of one cause
A finished image can show fine-scale variation, but it cannot reveal one certain cause. A smooth dark wall may contain sensor variation, JPEG artefacts, sharpening residue, demosaicing, resampling, or subtle real texture. This tool therefore makes a narrower local measurement in regions you identify as visually even and uses it only for a comparison inside the same decoded image.
Choose two similarly smooth shadow areas from one image. If one shows twice the residual variation, use that as a question: did one region receive less exposure, undergo a different local edit, contain compression artefacts, or simply have more texture? Do not use the ratio as an ISO verdict or compare it with a different camera file. Check the histogram and clipping, review the export path, and run a controlled repeat capture if the distinction matters.
Your selected file is processed in this browser. Image bytes, metadata, filename, pixels, and derived results are not uploaded or sent to analytics. A rendered-file measurement is not proof of RAW data, scene conditions, provenance, or artistic quality.
The optical values here are ideal geometric limits, not measured lens performance. Focus, motion, aberrations, subject contrast, sensor processing, crop, and output can change practical detail.
Make one controlled test with the important variable held constant. Inspect the subject detail, motion, highlights, and intended output, then adjust the setting that protects the priority you named.
FAQ
- Can this tell me which ISO is too noisy?
- No. It measures local variation in a finished decoded image. Sensor, exposure, processing, texture, and compression cannot be separated reliably from that one file.
- Is this a guarantee?
- No. The result is bounded by the inputs and assumptions shown on the page. It is designed to support a better next decision, not replace a test, measurement, proof, or professional judgement where one is needed.