← Back to the full Webcam Test

2-Minute Endurance Benchmark

Hardware Stress & Headroom Test

External 4K webcams and capture cards can drop frames or bog down once they're running alongside a resource-heavy call. This drives sustained high-resolution encoding for two minutes and tracks real, measurable signals of your system running out of headroom.

Camera off
Honest scope: a browser tab genuinely cannot read your CPU/device temperature or tell which physical USB port something is plugged into — there's no web API for that, and anything claiming otherwise is guessing. This measures what a browser can actually see — dropped frames, main-thread stalls, and render/encode timing — and reports how those change over the run. If performance degrades over the two minutes, that pattern commonly correlates with thermal throttling, USB bandwidth contention on a shared hub, or background CPU load — but this test can't tell you which one.

Status

Test Progress

Click "Start 2-Minute Test" to begin. Keep this tab active and avoid switching apps for an accurate read — background tabs get throttled by the browser itself, which would skew results.

Readout

Hardware Stress Test Results & Headroom Score

/100
Hardware Headroom Score

Signals Measured

  • Frames Encoded / Dropped:
  • FPS: Start → End:
  • Main-Thread Long Tasks:
  • Longest Main-Thread Stall:
  • Encode Queue Backlog (peak):

FPS Over Time

Share

Share Your Result

Related Tools

Take This Further

A low headroom score rarely travels alone. To see this result ranked against your other diagnostics, open your fix plan.

Notes

Why calls degrade after the first twenty minutes

A common and frustrating pattern: the call starts fine, and an hour later your video is stuttering while everyone else looks normal. This is rarely a network problem, and it is the specific failure this benchmark is designed to reproduce.

Sustained video encoding keeps your processor working continuously. Laptops in particular respond by reducing their own clock speed as they heat up, protecting the hardware at the cost of performance. Add a virtual background, a screen share, and twenty browser tabs, and the machine has less capacity available exactly when the call has been running longest. A short test never reveals this; two minutes of genuine sustained load does.

What the headroom score is built from

The score starts at one hundred and subtracts penalties for four measured behaviours: how far frame rate fell between the start and end of the run, what proportion of frames were dropped entirely, how many long tasks blocked the main thread, and how deep the encoder queue backed up. A machine coping comfortably loses very little. One struggling loses points in several categories at once, which is why a low score usually indicates a general capacity problem rather than one specific fault.

Acting on a low score

Close background applicationsCloud sync clients, browser tabs, and chat applications all wake periodically and compete for the same processor.
Disable video effects for long callsBackground blur and beauty filters are among the most expensive things a call can do, and the first thing worth sacrificing.
Check airflowA laptop on a soft surface throttles sooner. Hard surface, clear vents, and the problem often reduces measurably.
Lower your send resolutionA steady 720p is a better call experience than an unstable 1080p, both for you and for everyone watching.

What this cannot tell you

Browser JavaScript cannot read your processor temperature, inspect your cooling system, or identify which component is the bottleneck. This measures the observable result — whether performance degrades under sustained encoding load — not the underlying cause.

That is still the useful answer for the decision most people face, which is whether their current machine can carry an hour-long call with effects enabled. If it cannot, the remedies above are worth trying before concluding you need new hardware.

Benchmark Reference

Headroom Score at a Glance

GradeScoreWhat it means
Comfortable70 – 100Room for effects, screen sharing, and long calls.
Moderate40 – 69Workable, but effects and long calls will show some strain.
Constrained0 – 39Likely visible degradation during long calls.

The score starts at 100 and subtracts real penalties for FPS drop, frame loss, main-thread blocking, and encoder queue depth — all measured during your two-minute run.

Step-by-Step

How to Run the Test

  1. Close anything you'd normally have open during a real call, or leave it open to test your actual typical conditions.
  2. Click "Start" and allow camera access when prompted.
  3. Let the full two-minute test run without interacting with the page — this reproduces sustained load the way a real long call would.
  4. Read your headroom score and the four measurements it's built from.

Common Questions

Frequently Asked Questions

Why does my video call get worse after 30 minutes?
Sustained encoding keeps your processor working continuously, and laptops reduce clock speed as they heat up. This two-minute benchmark reproduces that pattern so you can see it before a real call.
How do I know if my laptop can handle video calls?
Run the stress test and read the headroom score. Below 40 means you will likely see degradation during long calls; above 70 means you have room for effects and screen sharing.
Does background blur slow down my computer?
Yes, considerably. Background effects and beauty filters are among the most expensive things a call can do, and disabling them recovers more headroom than lowering your resolution.
Can this test read my CPU temperature?
No. Browser JavaScript cannot access temperature sensors or identify which component is the bottleneck. It measures the observable result — whether performance degrades under load.
Why does the test need to run for two full minutes?
Thermal throttling and background-task interference both build up over time — a short test never gets hot enough or busy enough to reveal them. Two minutes of genuine sustained load reproduces the pattern that actually shows up on a real long call.