How browser observations, public standards, and project calibration inform Motion Audit.
Motion Audit combines public browser standards, observable Chromium/CDP telemetry, and project-defined calibration. This page explains what each source can support and where the evidence stops.
This protocol separates:
Public facts and browser observations support measurement capabilities. Calibration values are explicitly labeled as project methodology and validated with independent fixtures.
| Input | Permitted use | Boundary |
|---|---|---|
| W3C Web Animations API | Read Document.getAnimations(), keyframes, and timing | Describes animation state; it does not prove compositor execution |
| W3C Long Animation Frames API | Observe long-animation-frame entries and script/render timing | Availability varies by browser version |
| CDP LayerTree | Observe layer changes and request compositing reasons | CDP output is browser-specific telemetry |
| CDP Performance/Tracing | Collect timing and rendering events | Dropped frames require interpretation of trace/timing data |
| CSS and graphics definitions | Derive CSS-pixel, DPR, and raw RGBA estimates | Raw estimates are not driver VRAM measurements |
| Public rendering guidance | Identify forced-layout patterns and mitigation | Guidance does not provide numeric score weights |
These equations describe measurement primitives. They do not define a universal performance score.
For a CSS surface of width , height , and density , the raw texture estimate is .
This estimates an uncompressed RGBA8888 surface. Alignment, tiling, compression, swap chains, and driver allocation are outside the equation.
For sorted values , linear interpolation at percentile uses and .
The percentile implementation must define its interpolation and keep that choice stable across releases.
Given observed frame indices, begin a run only where (f-1) is absent, then count (f+1, f+2, \ldots). This is a deterministic (O(N)) hash-set calculation for the longest observed run; it is not a claim about frame duration.
Calibration values such as pillar weights, penalty points, memory bands, duration floors, and tier boundaries are project parameters. Each parameter must have:
src/;Synthetic fixtures may verify deterministic score behavior. They must not be described as browser ground truth or user-perceived performance evidence.
Real-site benchmarks must report raw telemetry beside the operational score: browser/version, viewport, DPR, run count, warm-up policy, frame timings, LoAF/long-task observations, layer telemetry, and modeled memory. Scores are comparative indicators within that controlled environment.
Implementation decisions are based on the public foundations and observations listed above, not on undocumented third-party internals or output values used as calibration oracles.
Built by Axyl. A performance auditing toolkit for motion-heavy web interfaces.
Last updated: 10/1/2026