MediaCapabilities powerEfficient agrees with WebCodecs prefer-hardware
For the same codec at 1920×1080, decodingInfo's powerEfficient equals VideoDecoder.isConfigSupported's answer with prefer-hardware. N/A without WebCodecs or MediaCapabilities (severity: warn).
What a detector infers
MediaCapabilities.decodingInfo() reports whether decoding a format would be power-efficient, and WebCodecs' VideoDecoder.isConfigSupported() with hardwareAcceleration 'prefer-hardware' reports whether a hardware decoder accepts it.
In a window, for unencrypted file playback, Chrome answers both by asking the GPU process the same question — whether its hardware decoder supports that codec, profile and size (IsDecoderConfigSupportedOrUnknown) — so for one codec at one size the two must agree. Measured agreeing on every stock arm: all true on Windows with an RTX 3070, all false on a Linux machine without hardware decoding and in chrome-headless-shell.
The failing shape is powerEfficient taken from a stored table while WebCodecs keeps asking the real decoder: AV1 marked not power-efficient while the hardware decoder accepts it. The probe tests at 1920×1080, because hardware decoders have minimum sizes of 64 to 144 pixels and tiny sizes legitimately report false, and it runs in the window without encrypted media, because workers and EME answer from Chrome's decode-statistics database instead.
Comparing two APIs is deliberate: in the audit's corpus, real hardware GPUs contradict any table of which GPU decodes which codec power-efficiently.
How to resolve it
Let both APIs ask the real decoder. Filling MediaCapabilities from a table while WebCodecs keeps querying the hardware describes two different machines; if hardware decoding should look absent, disable it in the browser so both see the same answer.
Nearby checks in Environment & locale
- Google's network voices arrive as Google's exact set
google-voices-exact-setBranded Google Chrome on Windows, macOS and Linux adds a set of network voices from a component extension whose manifest lists exactly 19 of… - The local voices come from the claimed OS's speech engine
local-voice-vendor-vs-osVoices marked localService are the host's own speech engine, enumerated by the browser from the operating system, and their names say where… - Intl's default locale is a UI locale desktop Chrome ships
intl-is-a-chrome-ui-localeChrome's Intl default locale — what new Intl.DateTimeFormat().resolvedOptions().locale returns — is not the operating system's region. - The browser's own UI strings are in the language Intl resolves
validation-message-vs-intlThe messages a browser writes itself — 'Please fill in this field.' under an empty required input — come from the renderer's resource… - A playback AudioContext buffers more than an interactive one
playback-hint-buffers-moreAn AudioContext created with latencyHint 'playback' asks for a larger buffer than one created with 'interactive', and Chrome honours the… - AudioContext latencies are values Chrome's audio stack can produce
audio-latency-render-stateTwo AudioContext latencies are produced by the audio stack at particular moments, and their values are constrained in ways a constant is… - A granted geolocation fix carries a real accuracy
geolocation-has-accuracyEvery position a real location provider returns comes with an accuracy: the radius, in metres, within which the true position probably lies. - The field-trial state of this build is reported (XSLT warning banner)
xslt-banner-readoutChrome is deprecating XSLT, and a fresh Chrome 155 profile shows it: an XML document styled with XSLT is rendered with a red deprecation…
Clearcote is a browser built for fingerprint coherence
It is a Chromium fork, maintained by the same people who wrote this reference. It ships as a compiled browser rather than as a stealth script injected into someone else's — which is a description of how it is built, and is not an argument about how it behaves on this check.
This audit takes no position on how Clearcote scores on Environment & locale checks, on this one, or anywhere else. It has no baseline corpus of other people's fingerprints to rank you against and no vendor scoreboard — nearly every check is self-referential, asking one browser the same question through two independent APIs and reporting whether both answers can be true at once. It runs identically on any browser, including ours. Run it on yours and read the result yourself.
See the other checks in Environment & locale — the family power-efficient-vs-webcodecs belongs to.