GearLab：用瀏覽器量一量你的裝備GearLab: measure your gear right in the browser

專案是什麼

GearLab 遊戲周邊實驗室是一組免安裝、免註冊的電競裝備檢測網頁。它不引用廠商規格表，而是讓你在自己的電腦上，用瀏覽器提供的介面（例如畫面更新、Web Audio、Gamepad）量測螢幕、滑鼠、鍵盤、耳機與搖桿目前的實際表現。

為什麼做這個

• 高刷新率螢幕經常因為系統設定停在 60Hz，卻沒有人察覺。
• 滑鼠的回報率、鍵盤的按鍵延遲，在規格表上是數字，換成自己的環境又是另一回事。
• 想知道問題出在哪個環節，通常要安裝好幾個不同的測試軟體。我們想把這些檢查放進一個打開就能用的網頁。

可以做什麼

• 螢幕測試：螢幕資料健檢、動態殘影（UFO）比較、全螢幕壞點與均勻度檢查、色域覆蓋估算、暗部與亮部細節、OLED 烙印檢測。
• 音訊檢測：耳機延遲量測、左右聲道與相位、聽聲辨位、頻率掃描、麥克風品質評分。
• 操控測試：滑鼠回報率與穩定度、鍵盤延遲與全鍵無衝突（NKRO）、搖桿死區與飄移。
• 玩家挑戰：反應速度、動態跟槍、打字速度等小挑戰，可以與其他玩家比較。
• 硬體專欄：說明鍵盤、螢幕與軟體設定背後原理的文章，例如 OLED 烙印、遊戲掌機的功耗調度。

怎麼使用

以檢查滑鼠回報率為例：

1. 打開 GearLab，進入「操控測試」中的滑鼠回報率。
2. 依畫面提示在測試區內持續移動滑鼠。
3. 查看回報率數值與穩定度波形；換不同的 USB 孔、模式或軟體設定後再測一次比較。

多數測試都是這樣：先照指示操作，再比較設定前後的數字。

設計上的取捨

• 免安裝、免註冊：降低試用門檻；代價是只能使用瀏覽器開放的介面，量不到驅動層級的資訊。
• 運算在你的瀏覽器：測量運算在你的電腦上完成；玩家挑戰的成績可以上傳到全站排行榜。
• 用數據取代形容詞：每項測試盡量給出可以比較的數字或圖形，而不是只有「通過／失敗」。

技術與資料來源

測量使用瀏覽器本身的介面，例如逐幀的 requestAnimationFrame、音訊的 Web Audio API 與搖桿的 Gamepad API。螢幕資料健檢會與型號規格比對，專欄文章的來源與日期標註在各篇文章中。

限制與注意事項

• 瀏覽器量到的是「瀏覽器看得到的表現」，結果會受瀏覽器、作業系統、省電模式與背景程式影響，不等同專業儀器的實驗室量測。
• 耳機延遲等項目會受藍牙編碼器與接收環境影響，建議在相同條件下重複測量。
• 測試結果適合用來比較「設定前後」或「不同裝置之間」的差異，不適合作為購買糾紛的證明。

前往與回饋

前往 GearLab 遊戲周邊實驗室。有想要的測試項目或發現問題，歡迎寫信到 [email protected]。

What it is

GearLab is a set of install-free, sign-up-free web pages for checking gaming gear. It does not quote a manufacturer's spec sheet. Instead it uses interfaces the browser provides (frame updates, Web Audio, the Gamepad API) to measure how your monitor, mouse, keyboard, headset and gamepad actually perform on your own computer.

Why we built it

• High-refresh monitors are often stuck at 60 Hz because of a system setting, and nobody notices.
• Polling rates and keyboard latency are just numbers on a spec sheet until you check them in your own setup.
• Finding the weak link usually means installing several different test programs. We wanted these checks in one page you can open and use.

What you can do with it

• Display tests: a monitor data check, motion ghosting (UFO) comparison, full-screen dead pixel and uniformity checks, colour gamut estimation, shadow and highlight detail, OLED burn-in checks.
• Audio checks: headset latency, left/right channels and phase, directional hearing, frequency sweep, microphone quality scoring.
• Input tests: mouse polling rate and stability, keyboard latency and NKRO, gamepad deadzone and drift.
• Player challenges: reaction time, tracking, typing speed and similar mini challenges you can compare with other players.
• Hardware articles: explanations of what is behind keyboards, monitors and software settings, such as OLED burn-in and handheld power management.

How to use it

Checking a mouse polling rate as an example:

1. Open GearLab and go to the mouse polling rate test under input tests.
2. Keep moving the mouse inside the test area as instructed.
3. Read the polling rate and the stability graph, then change the USB port, mode or software settings and measure again to compare.

Most tests work like this: follow the instructions, then compare the numbers before and after a change.

Design decisions

• No install, no sign-up: it is easy to try. The cost is that only what the browser exposes can be measured, not driver-level information.
• Computation in your browser: measurements are computed on your own computer, and player-challenge scores can be submitted to the site-wide leaderboard.
• Numbers instead of adjectives: each test tries to give you something you can compare, not just a pass or fail.

Technology and data sources

Measurements use the browser's own interfaces, such as requestAnimationFrame for per-frame timing, the Web Audio API for sound and the Gamepad API for controllers. The monitor data check compares against model specifications, and article sources and dates are noted in each article.

Limitations

• A browser measures what a browser can see. Results depend on the browser, operating system, power mode and background programs, and are not equal to laboratory instruments.
• Headset latency and similar tests are affected by Bluetooth codecs and the radio environment, so repeat them under the same conditions.
• Results are best used to compare "before and after a setting change" or "between devices", not as proof in a purchase dispute.

Go further and give feedback

Visit GearLab. If you want a test added or find a problem, email [email protected].

其他專案

• KESI 工具箱：想轉一張圖、合併兩份 PDF，卻要安裝軟體、註冊帳號，或把檔案傳到不知道誰管理的伺服器。
• OmniRadar 山海路全方位雷達：出門前要在公路局、氣象署、地圖與潮汐網站之間來回切換，才能拼出一個決定。
• AfterHero 保固通：發票與保固卡散落在抽屜與信箱，東西壞了才發現過保，也找不到型號與客服電話。

Other projects

• KESI Tools: You just want to convert an image or merge two PDFs, but you are asked to install software, sign up, or upload files to a server you know nothing about.
• OmniRadar: Before a trip you have to jump between the highway bureau, the weather agency, maps and tide sites to piece one decision together.
• AfterHero: Receipts and warranty cards are scattered across drawers and inboxes, and you only find out a warranty has expired when something breaks.