Products to try or avoid? PWM Flicker and Temporal Dithering Testing

jordan

So frustrating to add now motherboards as eyestrain contributor. You would think that boards supporting the same chip should be more consistent. At least you don't have to config those rigs often. Set it and forget it. Next time maybe you could run the iGPU output through a hdmi converter and OBS to get rid of the dithering on whatever computer you find most comfortable.

Don't laugh, I'm being serious unless you have food allergy.
https://www.quora.com/Is-chicken-liver-good-for-the-eyes
https://www.livescience.com/1021-scientists-eye-benefits-spinach.html
I need to find a good recipe.

Here is the spectrum for the 7i
https://ibb.co/7z6CvRv

and the "good" M2 MBP Touch Bar:
https://ibb.co/TKWmjyQ

Both are spikey but the red seems unbalanced on the Mac. Night light off and default settings on 7i. Mac is running sRGB color profile with true tone and night light off. Sunlight is a gentle curve with no big spikes.

13 inch M2 MacBook Pro Touch Bar testing:

I measure the same red peaking spectrum when on Color LCD profile. I measure on pure white like RGB (255, 255, 255). Not a concern for me. Being using this base model machine for the last few days and so far it feels comfortable and dare I say more comfortable than the 7i screen at 240 hz refresh rate even with all the dithering removal workarounds enabled.

So I used an industrial/scientific grade Thorlabs PDA100A2 photodetector and I see the threshold between almost flat signal to high frequency PWM or PWM like fluctuations at 63 to 64% brightness as adjusted using BetterDisplay display. Switching between 1280 and 1440 Rez didn't affect these readings.

What is strange is the notebookcheck reported 56% brightness threshold for PWM. Switching from 56% to 57% brightness sometimes but not always results in suddenly less high frequency peaks in the signal but it's not a consistent occurrence. Sometimes going from 57% to 58% higher brightness would increased not decrease PWM-like fluctuations. What is consistent on my cheap oscilloscope is the 63% to 64% brightness threshold change. Need to recover budget wise then someday will upgrade my oscilloscope. I use 20dB gain setting on the photodetector to account for lower backlight and the photodetector noise is still low as stated in the manual at 219 μV RMS and the bandwidth is 800kHz.

My general impression using this pro photodetector is that the Opple4 is still giving a roughly useful waveform and not just pure low light noise in my previous testing at 50% brightness. The Thorlabs device is awesome and should be. It cost almost as much as a used laptop!

simplex

I was not able to get the Opple4 to reliable distinguish PWM between 64% and 63% like I can with the Thorlabs device. As I increase brightness, the Opple waveform decreases in thickness until about 95% brightness or above 400 nits of brightness so I think the Opple is very noisy.

Edit: Repeating Opple measurements might yield slightly thicker or thinner waveforms even when the actual signal should be stable. The device seems suitable for checking larger fluctuations.

16 inch Lenovo Legion Pro 7i 2023 laptop testing:

Using PDA100A2 photodetector, the switch between PWM-like display backlight fluctuations and relative flat line is at 85-86% brightness as adjusted in Windows 11. Overall, the 7i backlight fluctuations are less than those of the M2 MBP TB as I needed to set the photodetector gain to 50dB to notice these tiny levels of flicker. On the MBP, I only needed 20dB gain.