Gradient Test

Introduction

The DeviceHub Gradient Test is a free online banding and dithering inspector that shows smooth color and gray ramps so you can visually judge whether your display path posts stair-steps, without claiming laboratory bit-depth certification. Photographers hunting posterized skies, video editors checking 8-bit versus 10-bit desktop pipelines, GPU tuners toggling dither options, and TV owners noticing contour lines in dark scenes use fullscreen gradients as a fast tell. Open the page, go fullscreen, and scan ramps for abrupt bands versus fine grain from dither. DeviceHub emphasizes visual judgment. Browser canvases, OS compositors, and panel processors all participate; remote desktop compression creates false banding. Pair with Gamma, Color, Black/White Level, Uniformity, and HDR Test (capability signals only, not a calibrated HDR meter). Posterized skies and contour lines in dark gradients are the quiet killers of photo and video confidence, often blamed on “the export” when the desktop pipeline’s bit depth and dither settings are the real culprits. DeviceHub’s Gradient Test shows smooth color and gray ramps so you can visually judge banding versus fine dither grain, without claiming laboratory bit-depth certification from a browser canvas. Photographers hunting rings in skies, editors checking 8-bit versus 10-bit desktop paths, GPU tuners toggling dither options, and TV owners noticing contouring in shadow rolls use fullscreen gradients as a fast tell. Visual judgment is mandatory: browser canvases, OS compositors, and panel processors all participate, and remote-desktop compression creates spectacular false banding. Pair with Gamma and Color Tests for tone and tint, Black/White Level Tests for end stops, Monitor Uniformity Test when blotches, not rings, are the issue, and HDR Test for capability signals only, not a calibrated HDR meter. Dead and Stuck Pixel Tools catch discrete dots that users sometimes mistake for gradient noise. Permissions stay none for easy classroom and office checks. Docking stations that renegotiate HDMI color depth are famous for introducing banding overnight; a DeviceHub ramp the morning of a grade session catches that before hours of wrong blame on the camera original. Teachers demonstrating 8-bit versus dithered output can show students stair-steps versus grain without claiming the tab certified panel bit depth. Visual judgment means sitting at a normal distance, nose-to-glass makes dither look like filth and smooth ramps look noisier than they are in real use. When HDR mode suddenly posters a ramp that was fine in SDR, record both passes and open HDR Test only for capability context, not as a nit meter. Uniformity Test is the wrong tool for rings and the right tool for blotches; keep that vocabulary straight in tickets. Permissions remain none; remote-desktop viewing is disallowed for conclusions because codecs fake contours aggressively. Color Test flat fields remain useful when a “band” is actually a tinted vertical smear from a failing column driver, rare, but worth a solid-field confirmation.

What this tool does

Horizontal, vertical, and perhaps radial gradients in gray and color channels help isolate where contouring appears. Copy explains dither versus true bit depth in plain language. The tool does not rewrite GPU pixel formats. It does not upload ramp screenshots. Use static viewing; animated gradients can hide banding. Horizontal, vertical, and related ramps in gray and color channels help isolate where contouring appears. Copy explains dither versus true bit depth in plain language so you do not confuse intentional grain with failure. The tool does not rewrite GPU pixel formats or HDMI deep-color flags. It does not upload ramp screenshots. Prefer static viewing; animated gradients can mask bands your still eye would catch. Change one pipeline setting at a time when A/B testing dither or HDR/SDR. If your GPU control panel exposes dither toggles, A/B them here with everything else held constant so the eyes see the only changed variable.

When to use it

Use Gradient Test after switching HDMI bit depth, after enabling HDR, when sky photos show rings, and when creative apps look posterized only on one monitor. Prefer Uniformity for blotches rather than rings. Prefer Color for flat field tint. Prefer Dead Pixel tools for discrete dots mistaken for noise. Use Gradient Test after switching HDMI bit depth, after enabling HDR, when sky photos show rings only on one monitor, and when creative apps look posterized on a docking station path. Prefer Uniformity for blotches rather than rings; prefer Color for flat-field tint; prefer pixel tools for discrete dots; prefer Gamma when midtones look wrong without clear stair-steps. Installers commissioning signage players can catch an 8-bit forced path before a client blames content. After driver updates that reset color settings, re-run ramps before a long grade session. After swapping from a direct DisplayPort cable to a dock hop, re-run the same gray ramp before blaming the creative application for contour lines that actually arrived with the link.

How it works

Gradients are drawn with CSS or canvas interpolation and composited. Limited output bit depth without dither shows bands. Some GPUs dither in the driver. Media queries and color management can alter ramps slightly. No permissions. Not an HEVC encode banding analyzer, though similar eyesight skills transfer. Gradients drawn with CSS or canvas interpolation are composited through the OS; limited output bit depth without dither shows bands, while some GPUs dither in the driver to hide them. Media queries and color management can alter ramps slightly across engines. No permissions are required. This is not an HEVC encode banding analyzer, though the same eyesight skills transfer. HDR tone mapping can posterize SDR-looking ramps; compare SDR baselines and use HDR Test only as a capability signal check. Phone photos add false contours, judge on the glass.

Step-by-step instructions

1. Open Gradient Test on the local display at native resolution, fullscreen, with temporary color filters off when practical.
2. Scan gray ramps slowly for stair-steps, then repeat on colored ramps if provided.
3. Change one pipeline setting, SDR/HDR, dither, or HDMI range, and recompare the same ramp.
4. If midtones look wrong without clear bands, open Gamma Test; if flat gray is blotchy, open Uniformity Test.
5. For HDR capability questions, use HDR Test as a signal check only.
6. Exit when finished; no revoke step is needed.

Common problems

Video-call compression invents banding. Phone photos add false contours. Sitting too close makes dither look like dirty noise. Assuming DeviceHub measured “true 10-bit panel” from a smooth ramp ignores processing. Vivid TV modes can both crush and band. Video-call compression invents banding that vanishes locally. Sitting too close makes dither look like dirty noise; back up to a normal viewing distance. Assuming DeviceHub measured “true 10-bit panel” from a smooth ramp ignores processing and dither. Vivid TV modes can both crush and band simultaneously. Windowed mode with browser chrome is fine for a glance, but fullscreen removes doubt about letterboxed scalers. Animated wallpaper behind a non-fullscreen window can distract; fullscreen the ramp before declaring the panel banded.

Privacy explanation

Gradients render locally. DeviceHub does not collect banding scores. No camera permission is requested. Close the tab to end the test UI. Gradients render locally; DeviceHub does not collect banding scores. No camera permission is requested. Close the tab to end the test UI; shared machines need no privilege revoke. Screenshots of ramps for docking-station tickets should crop unrelated desktop files; DeviceHub never scrapes those captures from your clipboard. Docking-station adapters that force 8-bit output often reveal themselves first as sky banding in Gradient Test even when solid Color Test fields look acceptable. Change one link in the chain, cable, dock, direct HDMI, and recompare the same ramp before blaming the panel. Creative apps that dither differently than the browser may disagree slightly; use DeviceHub as a desktop-pipeline check, not as a verdict on every export codec.