hpst
As usual, we can only guess, and try to find a needle in a haystack. I've been also thinking about refresh rates, frequencies, etc. For example I have some ear issues where particular frequencies tend to cause me physical pain (probably due to many years with headphones, etc.). There seems to be a relation between eye strain, sinus / jaw issues in some cases, both of them have also effect on ears. The pressure I can feel from staring at some laptops resembles sinus pain. So maybe our eyes are sensitive to specific frequencies, the same as ears, and these frequencies might have changed in recent years - due to dithering, new technology, more powerful gear, etc. Since we have different sight we may be sensitive to different frequencies - that would explain why someone can't use PWM screen with low refresh rate, while it doesn't bother other people. Just a thought, one of options.

hpst But since laptop panels are "dumb" and require the gpu to tell them to dither

No. I believe in the vast majority of setups the GPU is the one that does the dithering since the panel is unable to. The driver and/or VBIOS of the card "controls" what the GPU does.

    hpst why would it be happening on a true 8-bit panel?

    Because GPU companies and driver developers don't know how to implement a proper dithering configuration it seems?

    JTL We are on the same page I think...I may have just expressed it wrongly. The laptop panel doesn't have built in dithering instructions or blue light/contrast/brightness whatever etc like a desktop monitor can so I called that a "dumb panel". As I understand it the panel's pixels have to "dither" to get the "super awesome vibrant" colors and thats what I meant by saying the gpu is making the panel dither. I understood the gpu is making it happen.

    reaganry The GTX 9xx/10xx look horrid even in a BIOS setup screen. I doubt it helps with those that seem awful no matter what.

      JTL Nearly every BIOS screen I have seen is on 100% brightness with no ability to reduce it so I can't stand looking at them just due to that...it's blindingly bright and hurts on anything but a super old dim CCFL. A 2014 Thinkpad I got for testing did surprisingly allow brightness controls to work in the BIOS but an update ruined that.

      • JTL replied to this.

        hpst I was just pointing out how even outside of a running OS with just a greyscale BIOS screen and the newer Nvidia cards, it's dithering.

        • hpst replied to this.

          JTL

          What does that mean? Why in the world would a text only, pre OS utility be dithered? Does that mean its just hardwired into the gpu and whenever a computer is on its dithering regardless of what is displayed? I had some hope it was dithering because instructions in the OS/drivers were telling it to which would mean it might be possible to turn off. If its in firmware before the OS and drivers even get loaded how could we even affect it?

          11 days later

          deepflame Teradici (a company that makes display over Ethernet extenders that use image compression which dithering interferes with) claims that the 10-bit color option on certain AMD graphics cards can disable dithering. So that's another datapoint.

          https://help.teradici.com/s/article/1569

            JTL

            This could be huge as an easy DIY test to see if dithering is or isn't an obvious problem. Do you have any more info on what gpus? There isn't any detail. I see there are linux instructions there which was a nice surprise as I thought "yeah it probably only works on Windows". If this would work on AMD Ryzen/Vega laptops etc it would be an easy test as there are a couple at Best Buy I could try in store if this Catalyst Center thing is normally installed...but they have locked down internet so no installing new stuff there.

            Do you know if you can enable 10-bit monitor support, thus turning dithering off, on a Linux system with an AMD GPU/APU somehow? I could also live boot linux in a shop and try that as well. Worst case I would buy an AMD laptop and just return it if it didn't work...but I guess I need to know what GPUs/APUs this works for first. Does Nvidia have a similar process since there are a lot more Nvidia models out there to test?

            • JTL replied to this.

              hpst Don't know. I suspect this could only work for FIreGL graphics but I could be wrong.

              hpst Does Nvidia have a similar process since there are a lot more Nvidia models out there to test?

              If they did then I would be able to use my 1070 sitting on a box in a shelf 🙁

              • hpst replied to this.

                hpst I do plan to test this after I get the same/similar laptop as @deepflame after the capture card. Just a few more things to sort out with my desktop before I order it, nothing to do with you.

                I'll write you a summery email tomorrow.

                  hpst Also Linux no longer uses the Catalyst drivers so the instructions they provide don't work anymore.

                  JTL I guess I need to try to get to Best Buy soon and see if this Catalyst Control Center is installed on any of their AMD laptops and try these instructions. Does the 32bit registry entry mean these are old instructions? There are no dates on the article/entry.

                  • JTL replied to this.

                    hpst Does the 32bit registry entry mean these are old instructions?

                    No, it means how many bits are allocated to store the value. Nothing to do with 32/64 bit software or hardware.

                    • hpst replied to this.

                      JTL Well if Catalyst isn't used anymore as you said its likely this is old news. I was hoping I could either with a swtich or config file do the 10-bit enabling on a modern system and test it. Excited for nothing I guess since a 2008 year Thinkpad being usable isn't really helpful to me even if it works on that.

                      • JTL replied to this.

                        hpst Catalyst still exists on Windows, but I don't think 10-bit works on anything except the "professional" cards (FireGL/FirePro) Some laptops do have them but they aren't your average laptops.

                        JTL I have two of these Lenovo W500 now and one has gone bad. I think the reason was that I installed the Eizo drivers that came with the ColorEdge display I got. Now it even dithers on boot which the other good one does not.

                        I already tried to remove the battery and to reset the CMOS but I am not sure if it really helped. It may be that the Windows software made it bad again after boot.

                        Will test this again when time permits. ( I learned the hard way that color calibration has a big impact on dithering. this just shows me that colors and color depth play a big role in this. )

                        Just to let you know that there may be some things to look out for.

                        • JTL replied to this.

                          deepflame Hmm. When I eventually get my capture card I'll try to snag a W500 and do some of my own objective tests.

                          a month later

                          andc i suspect that notebookcheck pwm reviews are highly untrustworthy.

                          i have read several reviews that dont match.

                          Also my Xperia XZ 2, in notebookcheck, says it has pwm until 28%. But i check that it has also beyond that point almost until 100% with my canon 550D at 1/4000. I also use a 1.4 aperture lens, so i can i have light to check it.
                          Above the 28% brightness what happens is that the phone seems to go to a higher pwm frequency.

                          Until 28% is 2khz, above that seems to double it.

                          I am waiting for my oscilloscope to arrive to check it properly.

                          So maybe some displays, change PWM frequency depending on brightness.

                          • AGI replied to this.
                          • AGI likes this.
                            dev