1. Four different effects

CIE TN 006 separates measurable temporal modulation from the perceptions it may produce [S2]. Camera artefacts are another system interaction and cannot be cleared by eyesight alone.

2. PWM timing calculation

Illustrative values, not LIKELIGHT specifications: at 2,000 Hz, T=0.5 ms. At 25% duty, on-time=0.125 ms. A 1/50 s exposure lasts 20 ms and spans about 40 PWM periods; a 1/1000 s exposure lasts 1 ms and spans about 2 periods. More cycles may improve averaging, but do not guarantee band-free footage because phase, waveform, rolling/focal-plane readout and color-channel timing remain relevant [S4].

3. Why low brightness is often harder

PWM dimming commonly reduces duty cycle. Shorter pulses can expose timing interaction even when full output appears clean. Driver design or controls can also add temporal modulation [S2]. Test at 100%, 50%, 10% and the lowest intended show level, plus slow fades and representative RGB mixes.

4. A 48-clip camera matrix

For one approved camera body and recording mode, a practical screening matrix can use:

• 3 frame rates: 24, 25 and 30 fps;
• 4 exposure settings selected by the production team;
• 4 light levels/scenes: 100%, 50%, 10% and a representative mixed-color scene.

That gives 3 × 4 × 4 = 48 clips. This is a planning example, not a universal pass test. Repeat for every camera/readout mode required by the project. Record lens, ISO/gain, codec, anti-flicker/variable-shutter setting, fixture firmware, controller, scene file, supply condition and test distance.

5. Measurement and filming workflow

1. Freeze fixture model, driver/IC, firmware, controller, refresh settings and scene.
2. Measure light output waveform with suitable bandwidth and document setup per CIE TN 012 [S1].
3. Report frequency, modulation depth, duty cycle, pulse shape and any beat/non-periodic behavior.
4. Run the camera matrix using the final production body and sensor mode.
5. Inspect full frame and crops for bands, color separation, moving bars and exposure variation.
6. Adjust only approved variables: PWM/refresh mode, camera variable shutter, exposure, or scene floor.
7. Retest all changed scenes and archive raw clips plus settings.

6. Product evidence boundary

LIKELIGHT's official collection publicly lists a 36 W, DC24 V, RGB, DMX512 addressable linear product [S5]. The cited page does not publish PWM frequency, per-channel waveform, refresh architecture, bit depth at low level, camera compatibility or test clips. These items are 待确认 for the chosen product and controller. Do not infer them from voltage, protocol or LED count.

7. When not to approve

• The only test is a phone video in automatic mode.
• The supplier gives only a PWM-frequency number without waveform and test conditions.
• Testing covers 100% output but not low levels, fades or mixed colors.
• Camera body is named but sensor/readout mode, frame rate and exposure are absent.
• A single successful shutter setting is advertised as universally camera-safe.
• Firmware or controller settings change after acceptance without retest.

8. Evidence ledger

9. FAQ

Does a higher PWM frequency always eliminate banding?

No. It can help, but exposure, sensor readout, phase, waveform and channel timing still matter.

Why does a phone look clean while a cinema camera shows bands?

Different sensors, automatic exposure, readout and processing can produce different results.

Can variable shutter fix every scene?

No. It may reduce interaction for a defined frequency, but scenes and timing can change [S4].

Should every color be tested?

Test representative primaries, whites, low levels, fades and critical show content because channels may behave differently.

Is IEEE 1789 a camera-banding standard?

No. It addresses LED modulation practices for viewer risk, not a universal cinematography acceptance test [S3].

Who authored this guide?

Author: LikeLight Tom, Operations. Technical reviewer: 待确认.