Touch Sampling Rate vs Refresh Rate: Input Latency in Mobile Displays

Every gaming phone now advertises two numbers: 120Hz refresh and 240Hz, 480Hz, even 1000Hz touch sampling. They look like the same spec — they're not. One determines how smooth motion appears, the other how quickly your finger intent becomes action. Confusing them is exactly what marketing wants you to do.

We measured the full input-to-photon pipeline on nine current flagships. The delta between a 120Hz/120Hz phone and a 120Hz/240Hz phone can be 28ms of touch latency. The delta between 240Hz and 480Hz sampling? Often under 4ms — and swamped by software debouncing. If you're buying for responsiveness, Hz alone will mislead you.

1. Refresh Rate: The Output Clock

Refresh rate is the fixed cadence at which the display controller pulls a new frame from the framebuffer and scans it to the OLED. At 60Hz, that's every 16.6ms. At 120Hz, every 8.3ms. At 144Hz, every 6.9ms.

It defines motion clarity, not responsiveness. A 120Hz panel shows twice as many unique positions per second as a 60Hz panel, halving persistence blur and judder. Modern LTPO OLEDs can dynamically switch between 1Hz–120Hz to save power, but during interaction they lock high.

Important nuance: refresh is a ceiling. If the SoC only renders at 45 fps, a 120Hz display still scans at 120Hz but repeats frames. You get smooth scrolling of the OS compositor (SurfaceFlinger / Core Animation does 120 fps easily), but heavy 3D games may not benefit without sustained GPU headroom — which is why our Dimensity 9400 review stressed sustained vs peak frame delivery.

What refresh actually costs

Doubling refresh from 60Hz to 120Hz increases display pipeline power by ~180–250mW and GPU load by nearly 2x for the same scene. That's why variable refresh matters: the OIS/EIS camera stack isn't the only subsystem that benefits from adaptive clocking — the display driver IC does too, modulating VCOM and scan timing.

2. Touch Sampling Rate: The Input Clock

Touch sampling rate is the scan frequency of the capacitive digitizer — the transparent grid of TX/RX electrodes laminated above the OLED. The touch controller (often Goodix, FocalTech, or Synaptics) drives TX lines and reads mutual capacitance changes at each intersection.

At 120Hz sampling, the controller takes a snapshot every 8.3ms. At 240Hz, every 4.16ms. At 480Hz, every 2.08ms. Each scan produces a raw capacitance map that firmware filters for noise, palm rejection, and moisture.

Three factors make sampling rate slippery as a market spec:

• Instantaneous vs average: Many phones advertise "480Hz touch" but only achieve it in Game Mode or with one-finger touch. Two-finger scan drops to 240Hz due to TX/RX time-division.
• Reporting rate ≠ sampling rate: Some controllers sample at 480Hz internally but report to the AP at 240Hz over SPI/I²C after filtering. You're buying internal oversampling.
• Single-point vs multi-point: Peak numbers often reflect single-touch polling. Real gaming with two thumbs + shoulder triggers runs slower.

Unlike refresh, touch sampling is bursty. The scheduler boosts the touch IRQ thread priority when a finger is down. Lift your finger and sampling idles to 60Hz to save power.

3. The Latency Pipeline: From Finger to Photon

End-to-end input latency is the sum of six stages. Hz numbers capture only two of them.

So why does 240Hz touch feel better even though it's only a 4ms theoretical win over 120Hz? Because jitter matters more than average. With 120Hz sampling, your finger can land up to 8.3ms after the last scan — that's a uniform random delay. At 240Hz, worst-case wait is 4.16ms. The game receives a fresher coordinate, and the engine interpolates fewer extrapolated points between hardware events. Your swipe curve is sampled 2× finer, reducing aliasing.

But Stage 2 filtering kills much of that gain. To reject charger noise and greedy palm contacts, firmware applies IIR filtering and a 2–3 sample confirmation. A controller scanning at 480Hz but requiring 3 consecutive hits effectively reports at 160Hz. This is why independent latency measurements matter more than Hz.

4. Why “480Hz Touch” Rarely Means 480Hz

We logged touch event timestamps via getevent -lt and high-speed 1000 fps camera tests on three advertised 480Hz phones. Results:

• Phone A delivered 480Hz only with one finger idle-swiped at <200mm/s in Game Mode. Two fingers: 235Hz avg.
• Phone B sampled at 480Hz but throttled reporting to 240Hz over the SPI bus; Android's MotionEvent batching showed 4.16ms intervals, not 2.08ms.
• Phone C hit 480Hz genuinely, but display was spaced 8.3ms VSYNC, so app only consumed every second sample. Effective latency win: 2.1ms.

The same pattern repeats in our cross-platform framework benchmark — framework overhead (React Native / Flutter gesture pipelines) adds 6–14ms that dwarfs controller differences. A well-optimized native game at 120 fps with 240Hz touch will beat a poorly batched 480Hz-touch app every time.

For developers, stop polling at display refresh rate. Use InputManager batching and predictive touch (Android 12+ VelocityTracker with MotionEvent.ACTION_PREDICTED) rather than assuming raw Hz will fix latency.

5. Measured Data: What Actually Ships

* Tap latency = finger contact to first bright pixel change, 1000 fps camera, native touch tester app. Lower is better. ** Firmware throttled when temperature >38 °C.

Takeaway: the fastest photon latency wasn't the highest Hz. ROG Phone wins because its touch firmware runs with near-zero debounce and its Snapdragon 8 Gen 4's touch IRQ is pinned to a prime core — not because 720Hz alone is magic. Pixel's 121Hz sampling looks modest but its well-tuned compositor and passkey-optimized secure pipeline scheduling keeps latency competitive with 240Hz rivals in UI tasks.

6. Gaming vs Everyday Use: Where It Matters

For competitive shooters, rhythm games, and MOBA: 240Hz touch on a 120Hz+ panel is the sweet spot. You get ~4ms lower worst-case sampling delay plus finer trajectory data for aim assist. Enable Game Mode — it raises touch priority, disables palm filter aggressiveness, and pins sampling high. Pair with a 120 fps game build; a 60 fps game wastes half your touch samples waiting for VSYNC.

For scrolling, typing, casual games: Refresh rate dominates perception. A 120Hz/120Hz phone with excellent touch prediction feels better than a 60Hz/240Hz phone. Typing latency is dominated by haptic + audio sync anyway.

For stylus / drawing: Neither spec captures pressure sampling or tilt. Apple Pencil samples at 240Hz but interpolates with 9-axis prediction; many Android styluses sample at 360Hz. Check pen latency, not finger touch.

7. Buying and Tuning Advice

• Prioritize 120Hz LTPO first. It's the biggest quality-of-life jump. Then demand ≥240Hz sustained touch.
• Ask for sustained, multi-touch numbers. "480Hz instantaneous" is brochureware. Look for third-party latency tests.
• Use the right mode: Force 120Hz in Developer Options if battery allows, enable Game Mode for touch boost, disable battery saver during tournaments.
• Reduce pipeline fat: Close overlay apps, use wired or 1000Hz polling peripherals where supported, and keep touch area clean — moisture forces heavier filtering which adds 5–8ms.

Ultimately, touch sampling and refresh are complementary clocks. Refresh decides how fine the flipbook is; sampling decides how often the pen writes a new line. The best phones, like those with next-gen silicon we covered in our Dimensity 9400 analysis, co-optimize both ends — and the scheduler in between.