Last updated: 08/10/2024 All future testing beyond the results in the table will be done with ThorLabs PDA100A2 photodetector as it will likely pickup even minute levels of PWM type flicker. After the latest Windows 11 update to build 22631.3155, I now see significantly more pixel level flicker on the Legion 7i, T480s, and on the LG 27gp95r-b monitor. Still testing... Update: As recommended by @jordan , using Windows 10 1809 (with a manual OS update, instructions below) has led to improvements. Using intel iGPU for T480s and 7i results in significantly less pixel flickering regardless of screen refresh rate (60Hz, 64Hz, or 240Hz). If using the Nvidia GPU on 7i with latest studio driver on 1809, I still see flickering with the screen is set to 60Hz but less so at 240Hz. Installation for Win 10 1809 for T480s and Lenovo Legion Pro 7i: Win10 Iot Enterprise LTSC 2019 (Build - 17763.107) en_windows_10_iot_enterprise_ltsc_2019_x64_dvd_a1aa819f.iso <[removed by moderation]> You may have to set secure boot to off for successful installation. Manually update with this version after installing with the previous iso. https://www.tenforums.com/windows-10-news/152009-kb4538461-cumulative-update-windows-10-v1809-build-17763-1098-march-10-a.html Microscope Videos (240 fps played back at 30fps): https://www.dropbox.com/scl/fo/47fw3ygxkxaq78ul0j19i/h?rlkey=a2fvkmlujs2m9tkmdw59ujwvl&dl=0 Hello folks. I am in the process of testing panels for pwm using Miniware DS213 oscilloscope and BPW34 photodiode + Opple Light Master 4 and temporal dithering and/or pixel inversion with high speed phone camera (240 fps) and microscope (carson microflip at 100-250x magnification). So far: Device Panel Model # PWM Dithering (or pixel inversion) OS Lenovo Legion 7i Gen 8 (Nvidia RTX 4080, 2560x1600 matte IPS) NE160QDM-NZ3 (BOE) Opple: 0.0141 Flicker Index, 7.61% Modulation at 50% brightness 33425Hz None detected with Miniware Yes (in 60hz mode), Higher frequency dithering less detectable in 240hz mode) More comfortable for me but YMMV Window 11 Home 23H2 Build 22631.2861 Thinkpad T480s (2560x1440 matte IPS, Intel UHD 620) LP140QH2-SPB1 None detected with Miniware, Opple: 0.0171 Flicker Index, 8.97% Modulation, 36342Hz at 52% brightness Yes (also with B140QAN02.3 ) Apparently after re-testing both panels show dithering so the T480s is no magic solution. B140 panel seemed to have less pixel dithering. Windows 10 Pro 22H2 Build 19045.4046 Thinkbook 13s G3 (2560x1600 matte IPS, AMD Ryzen 5600U) N133GCA-GQ1 None detected with Miniware Yes All (windows 10 potato edition linux mint xfce…) Nothing worked including ditherig, etc. 🙁 UPerfect 18 inch 144 hz 2560x1600 IPS portable display N/A None detected with Miniware Yes just less noticeable due to higher refresh rate Windows 11 Home 23H2 Build 22631.2861 (tested using nvidia 4080 laptop gpu via usb-c) Eizo ColorEdge CS2740 N/A ? Yes (both 8 bit and 10 bit modes) windows 10 potato edition (ryzen 5700g iGPU via HDMI, nvidia quadro k4200 via displayport), windows 11 23H2 (tested using nvidia 4080 laptop gpu via usb-c, hdmi) LG 27gp95r-b (not recommended due to panel horizontal line artifacts after less than a year of use) N/A Opple: 0.0089 Flicker Index, 4.53% modulation, 34008 Hz (144Hz, 35% brightness) None detected with Miniware (update: yes with thorlabs detector at 20dB gain: https://imgur.com/a/Vc4SVzk ) Yes just less noticeable due to higher refresh windows 11 23H2 (nvidia 4080 via usb-c) The older Thinkpad T480s seems good and never had problems with it. The Eizo also has a rougher matte surface that may be a problem for some. For the uperfect display I could not see dither at 144hz but not sure at 60hz setting. Given the limitations of equipment and threshold at which PWM may still exist, these data are considered measures of relative safeness of the tested devices. For devices with Nvidia GPU, I used their latest studio drivers and also the latest graphics drivers from AMD for the ThinkBook. For the Eizo I was not able to get clear info on whether it was true 10 bit or 8 bit + FRC. These two sites show conflicting info? I should also mention that I also used novideo_srgb and ColorControl but was unable see a difference after setting temporal dithering to disabled. https://www.displaydb.com/monitor/eizo-cs2740 https://www.displayspecifications.com/en/model/21d01e78 I also tested the keyboard backlight of the 7i and T480s. I did not detect any flicker from my T480s. The 7i keyboard light flickers intensely at low enough frequency to be visible in normal digital camera video at 1/8000s shutter speed. To be safer, I would recommend not using keyboard backlighting at all and focus on increasing your natural workspace/living space ambient light levels. The Opple Light Master 4 results seem promising. However, when doing repeated measurements, the reported frequency changes widely from high 30K to low thousands or occasionally even low hundreds for all of the tested panels. However, the results are always reported as "no risk". This is consistent with other's findings about the limitations of the Opple Light Master. 7i Color Rendering Index (CRI) at 96%? Light Master report on 7i laptop internal display measured at 100% brightness on pure white. The CRI is reported to be high at 96%. However the default color temperature of the display is way too blue/high at greater than 8625 Kelvin. https://iot2.opple.com/lightmaster/index.html#/report/01f5aa1814f8481f97d475d204dec4be ~~~~ 7i spectrum measured using HPCS-320 shows Ra of 87.9 and CCK (K) of 6427.

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

photon78s

Last updated: 08/10/2024

All future testing beyond the results in the table will be done with ThorLabs PDA100A2 photodetector as it will likely pickup even minute levels of PWM type flicker.

After the latest Windows 11 update to build 22631.3155, I now see significantly more pixel level flicker on the Legion 7i, T480s, and on the LG 27gp95r-b monitor. Still testing...

Update:
As recommended by @jordan , using Windows 10 1809 (with a manual OS update, instructions below) has led to improvements. Using intel iGPU for T480s and 7i results in significantly less pixel flickering regardless of screen refresh rate (60Hz, 64Hz, or 240Hz). If using the Nvidia GPU on 7i with latest studio driver on 1809, I still see flickering with the screen is set to 60Hz but less so at 240Hz.

Installation for Win 10 1809 for T480s and Lenovo Legion Pro 7i:

Win10 Iot Enterprise LTSC 2019 (Build - 17763.107) en_windows_10_iot_enterprise_ltsc_2019_x64_dvd_a1aa819f.iso <[removed by moderation]>
You may have to set secure boot to off for successful installation.

Manually update with this version after installing with the previous iso.

https://www.tenforums.com/windows-10-news/152009-kb4538461-cumulative-update-windows-10-v1809-build-17763-1098-march-10-a.html

Microscope Videos (240 fps played back at 30fps):

https://www.dropbox.com/scl/fo/47fw3ygxkxaq78ul0j19i/h?rlkey=a2fvkmlujs2m9tkmdw59ujwvl&dl=0

Hello folks. I am in the process of testing panels for pwm using Miniware DS213 oscilloscope and BPW34 photodiode + Opple Light Master 4 and temporal dithering and/or pixel inversion with high speed phone camera (240 fps) and microscope (carson microflip at 100-250x magnification).

So far:

Device	Panel Model #	PWM	Dithering (or pixel inversion)	OS
Lenovo Legion 7i Gen 8 (Nvidia RTX 4080, 2560x1600 matte IPS)	NE160QDM-NZ3 (BOE)	Opple: 0.0141 Flicker Index, 7.61% Modulation at 50% brightness 33425Hz None detected with Miniware	Yes (in 60hz mode), Higher frequency dithering less detectable in 240hz mode) More comfortable for me but YMMV	Window 11 Home 23H2 Build 22631.2861
Thinkpad T480s (2560x1440 matte IPS, Intel UHD 620)	LP140QH2-SPB1	None detected with Miniware, Opple: 0.0171 Flicker Index, 8.97% Modulation, 36342Hz at 52% brightness	Yes (also with B140QAN02.3) Apparently after re-testing both panels show dithering so the T480s is no magic solution. B140 panel seemed to have less pixel dithering.	Windows 10 Pro 22H2 Build 19045.4046
Thinkbook 13s G3 (2560x1600 matte IPS, AMD Ryzen 5600U)	N133GCA-GQ1	None detected with Miniware	Yes	All (windows 10 potato edition linux mint xfce…) ~~Nothing worked including ditherig, etc. 🙁~~
UPerfect 18 inch 144 hz 2560x1600 IPS portable display	N/A	None detected with Miniware	Yes just less noticeable due to higher refresh rate	Windows 11 Home 23H2 Build 22631.2861 (tested using nvidia 4080 laptop gpu via usb-c)
Eizo ColorEdge CS2740	N/A	?	Yes (both 8 bit and 10 bit modes)	windows 10 potato edition (ryzen 5700g iGPU via HDMI, nvidia quadro k4200 via displayport), windows 11 23H2 (tested using nvidia 4080 laptop gpu via usb-c, hdmi)
LG 27gp95r-b (not recommended due to panel horizontal line artifacts after less than a year of use)	N/A	Opple: 0.0089 Flicker Index, 4.53% modulation, 34008 Hz (144Hz, 35% brightness) None detected with Miniware (update: yes with thorlabs detector at 20dB gain: https://imgur.com/a/Vc4SVzk)	Yes just less noticeable due to higher refresh	windows 11 23H2 (nvidia 4080 via usb-c)

The older Thinkpad T480s seems good and never had problems with it. The Eizo also has a rougher matte surface that may be a problem for some. For the uperfect display I could not see dither at 144hz but not sure at 60hz setting. Given the limitations of equipment and threshold at which PWM may still exist, these data are considered measures of relative safeness of the tested devices. For devices with Nvidia GPU, I used their latest studio drivers and also the latest graphics drivers from AMD for the ThinkBook.

For the Eizo I was not able to get clear info on whether it was true 10 bit or 8 bit + FRC. These two sites show conflicting info? I should also mention that I also used novideo_srgb and ColorControl but was unable see a difference after setting temporal dithering to disabled.

https://www.displaydb.com/monitor/eizo-cs2740

https://www.displayspecifications.com/en/model/21d01e78

I also tested the keyboard backlight of the 7i and T480s. I did not detect any flicker from my T480s. The 7i keyboard light flickers intensely at low enough frequency to be visible in normal digital camera video at 1/8000s shutter speed. To be safer, I would recommend not using keyboard backlighting at all and focus on increasing your natural workspace/living space ambient light levels.

The Opple Light Master 4 results seem promising. However, when doing repeated measurements, the reported frequency changes widely from high 30K to low thousands or occasionally even low hundreds for all of the tested panels. However, the results are always reported as "no risk". This is consistent with other's findings about the limitations of the Opple Light Master.

7i Color Rendering Index (CRI) at 96%?
Light Master report on 7i laptop internal display measured at 100% brightness on pure white.
The CRI is reported to be high at 96%. However the default color temperature of the display is way too blue/high at greater than 8625 Kelvin.
https://iot2.opple.com/lightmaster/index.html#/report/01f5aa1814f8481f97d475d204dec4be~~~~

7i spectrum measured using HPCS-320 shows Ra of 87.9 and CCK (K) of 6427.

JTL

photon78s For the Eizo I was not able to get clear info on whether it was true 10 bit or 8 bit + FRC. These two sites show conflicting info?

An internal source of mine at EIZO claims the CS2740 is true 10-bit.

However AFAIK even if it was 8-bit+FRC, in theory feeding it an 8-bit signal would still be okay.

JTL

photon78s These two sites show conflicting info?

I don't know about displaydb.com but as far as I know the owner of displayspecifiations.com is anonymous, and does not disclose how they obtain the information displayed on their site.

AlanSmith

A couple of tips, (1) Isn't it easier to go to the site notebookcheck.net read reviews for laptops? (2) Intel iGPU is responsible for displaying a 2D picture (desktop) in a laptop. Hence the dithering in 1440/60Hz mode. Just disable the Intel iGPU and the problem will go away. (2) Only the IPS matrix does not have pixel flickering? Do not buy laptops with other panels. (3) The problem with the Intel iGPU is not only forced anti-aliasing (pixel flickering), but also de-synchronization (Input Lag) of image output to the monitor. After turning on FreeSync Premium the problem is solved.

photon78s

AlanSmith

Yes, I know about notebookcheck for many years. It would be easier but I had access to these devices before discovering this site only recently and recognizing symptoms.

I have tried HDMI, displayport, usb-c, and from nvidia 4080 and hdmi from ryzen 5700g iGPU to the Eizo and I can still measure dithering even when output is set 8 bit mode to avoid potential 8 bit + FRC feature. I also tested a quadro k4200 running on windows 10 potato edition via displayport and still see dithering. I no longer have the Eizo but I still think it is more likely my gpu/ software combination doing the dithering.

For the legion laptop, I have the igpu disabled in bios for testing all external monitors. In NVIDIA control panel, I have tried both 8 bit and 10 bit output. I have tested two T480s both running on intel 620 igpu, same screen specs, same os build versions and drivers, but different panels and differing dithering results. I do this in case it is helpful for others and I do hope more review sites provide more testing beyond just backlight flicker.

Based on posts on this forum, it looks like dithering depends on a combination of software (OS, drivers), video card model and the firmware/bios, display, and perhaps even sample variation from manufacturing defect but I not sure about that. For the uperfect, I tested using usb-c at 144hz. If I understand correctly, others are saying dithering is "baked in" to these newer nvidia and amd GPUs now.

No idea what AAS matrix is. Still looking…*

*https://ledstrain.org/d/2460-ah-ips-ips-ahva-ads-aas-my-experience-with-different-lcd-panels

photon78s

AlanSmith

And notebookcheck.net does not do temporal dithering measurements for all devices. Disabling intel iGPU in bios resulted in no visible change under the microscope. On the Legion, the 240 hz setting can be enabled while the nvidia GPU activity icon shows no activity. As someone still doing and learning 3D web graphics programming, 240hz is only a slight inconvenience for clients but that's another issue for another time.

An old arduino demo of dithering (what is the frequency of these algorithms?)

https://www.youtube.com/watch?v=1-_JtRl2ks0

JTL

JTL Also I forgot to mention. Same source claims CS2740 is flicker/PWM free.

AlanSmith

photon78s You understand, manufacturers are increasingly lying that the 8bit panels (although in fact they are 6bit + FRC). There should be no dithering on honest 10bit. You can check your monitors/laptops on Android x86 (for your computer).

You can install Android 13 X86 (Gapps) on a USB flash drive through Rufus. And run as a LiveCD without installation. Will there be dithering/flickering on 6bit. Or is it a panel problem?

jordan

So is the Legion 7i Gen 8 considered comfortable at 240hz then? I'm curious about if the OS could still cause strain even without pwm/dithering

photon78s

jordan

So far so good (a few weeks) using the 240hz internal display and the uperfect montior at 144hz connected using usb-c /nvidia gpu. YMMV.

OS/drivers/software/accessibility settings could all matter. So much discussion of that on this forum already. And their is also the panel lottery situation as well.

I one point I tried using an usb capture card connected an e-ink android tablet as a DIY alternative to dedicated e-ink monitors for text only work but it was way too slow, bad contrast, etc. Those dedicated e-ink monitors also still have a long way to go but seems to work for some.

jordan

photon78s that's great! Maybe I'll consider trying one if my PC I'm building doesn't end up working for me.

photon78s

jordan

FYI, the keyboard backlight flickers on that laptop.

jordan

photon78s yeah I figured so. Assuming it's able to be turned off right ?

Edit: any idea why 240hz would disable dithering ? Wonder if this also applies to alternative laptops with similar specs

photon78s

jordan

I almost never turn on the keyboard backlight. Interestingly, the keyboard backlight on the T480s seems to not flicker or flicker at a very high frequency that I could not detect. Seems subjectively more comfortable to me as well. I think it's better to adjust the ambient light levels. As for the screen, perhaps at a higher refresh rate, it forces the dithering algorithm to change. I don't think these results are unique to this particular laptop.

Sunspark

T480S using the NV140FHM-N46 panel flickers at 990 Hz, so if you can't detect it that's good!

There are 2 other panels, it really depends what you have.. 1080, 1080 w/ multitouch and a 1440 panel.

jordan

When you test temporal dithering are you only testing on a white image? Or is it tested on different colors and different shades of the colors?

photon78s

jordan

All shades and gradients. If you test on white, you will most likely not be able to detect anything. Highly recommend people try themselves using that cheap microscope.

Also, ideally, one would test the pwm with a better oscilloscope and for the backlight led spectrum (people saying their is a definite "harshness" quality to the light). Perhaps something like this for optical spectrum?

https://www.amazon.com/dp/B081ZJMFBQ/ref=sspa_dk_detail_0?pd_rd_i=B081ZJMFBQ&s=industrial&sp_csd=d2lkZ2V0TmFtZT1zcF9kZXRhaWwy&th=1

https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=5276

https://www.youtube.com/watch?v=mNmvfSK-Dnw

jordan

photon78s wowww that sounds very promising!!! I was meaning to ask this earlier.

photon78s

jordan

Not to get hopes too high, I feel all of these monitors I've tested still have that "harshness" to the light and not anything to do with flicker/pixel inversion/dithering that is more straining than for example an old incandescent lamp.

jordan

photon78s gotcha. Between the 7i and t480s which one seems to be the most safe? I wonder if you would find anymore comfort with windows 7 for example on the t480s?

Worst case I'd convert a laptop to a transparent screen like someone did on reddit 🤔 just the color flicker is something I'd always want to avoid since it bothers me bad.

photon78s

jordan

I like the 7i better because of the high refresh rate. Also I get less discomfort with taller screens aspect ratios (16:10 not 16:9 on the t480s).

photon78s

jordan

I tested the 7i at native resolution for pixel inversion on https://pixelinversion.com/ made by forum member GregAtkinson. Pretty good with slight movement in some of the horizontal line tests. A different 7i even with the same panel model may be different.