si_edgey So true, and I think we need to start thinking of this condition as though it were a disability, which is just as valid as something like epilepsy and should be treated as such by developers.

Agreed. We know that dithering causes flicker, which the patents we've found suggest is 15hz. If I switched this monitor to 24hz now it would floor me within seconds, I dread to think what 15hz would be like.

si_edgey That's interesting, but I'm certain that it is always either a) the gpu driver that is responsible or b) some kind of scan-out hardware on the output if you're using a laptop. When I got through to the graphics tech guys at Microsoft they said:

"Everyone I’ve asked says that there’s nothing Windows does at a software level for temporal dithering so that would imply the problem is at the hardware level." (this was when using the MS basic display adapter and still having problems)

I'm starting to believe that most devices post-W10 have dithering 'baked in' at the hardware level. This is why we're finding it so hard to use devices. In fact one test I would suggest everybody does on a new device is; Connect it to a known good monitor, enter the BIOS and go through the menu options or just stare at it for 15 minutes. It should become clear if something is 'off' with the output in that timeframe. So we have two issues, really, the Driver used by the OS and the actual output of the device itself.

I'm going to give 20H2 another try today, after some googling there should be a way for my 2012 Intel driver to install. Hope it works.

        diop Agreed. We know that dithering causes flicker, which the patents we've found suggest is 15hz. If I switched this monitor to 24hz now it would floor me within seconds, I dread to think what 15hz would be like.

        I must've missed that - could you link me to the thread discussing this?

        diop I'm starting to believe that most devices post-W10 have dithering 'baked in' at the hardware level

        I sincerely hope that this isn't the case, but I fear you might be right..

        • diop replied to this.

              si_edgey I must've missed that - could you link me to the thread discussing this?

              Sure. https://patents.google.com/patent/US20100238193

              However, depending on the algorithm used to vary an applied dither pattern from frame to frame, temporal dithering can cause the undesirable visible artifact known as “flicker.” Flicker results when dithering produces a sequence of pixels that are displayed at the same location on a display screen with periodically varying intensity, especially where the frequency at which the intensity varies is in a range to which the eye is very sensitive. The human eye is very sensitive to flicker that occurs at about 15 Hz, and more generally is sensitive to flicker in the range from about 4 Hz to 30 Hz (with increasing sensitivity from 4 Hz up to 15 Hz and decreasing sensitivity from 15 Hz up to 30 Hz). If the pixels displayed at the same screen location (with a frame rate of 60 Hz) have a repeating sequence of intensities (within a limited intensity range) that repeats every four frames due to dithering, a viewer will likely perceive annoying 15 Hz flicker, especially where each frame contains a set of identical pixels of this type that are displayed contiguously in a large region of the display screen. However, if the pixels displayed at the same screen location (with a frame rate of 60 Hz) have a repeating sequence of intensities (in the same intensity range) that repeats every sixteen frames, a viewer will be much less likely to perceive as flicker the resulting 3.75 Hz flicker.

              In order to reduce the flicker caused by temporal dithering a repeating sequence of dither bits with a sufficiently long period of repetition can be used. However, until the present invention, dithering had not been implemented in a programmable manner that allows the user to vary both spatial and temporal dither parameters so as to reduce artifacts caused by the dither process itself or the interaction of dithering with other processes, such as pixel inversion.

                Haha you can never be sure in this game. But it works 100% for me - 6 months of use with the old Intel driver for many hours per day and no headaches or migraines.
                If I switch to a driver from 2018 I feel dizzy and sick and switch to a 2020 driver and I can't focus on anything - if I used it for 30 minutes I'm sure it would trigger a migraine. But for some reason the 2016 driver is like looking at paper for me. All of which implies that my issue is caused entirely by the Intel driver when using Windows v2004.

                So you have proved that with the same OS 2004 you can use only intel drivers from 2017 and not more recent one. That is nice, at least you can use a recent OS and you can play some games again 🙂
                Now it would be interesting to see if you can use that driver with other OS version like 1809/1903 (i am using these 2 right now)!

                Agreed. We know that dithering causes flicker, which the patents we've found suggest is 15hz. If I switched this monitor to 24hz now it would floor me within seconds, I dread to think what 15hz would be like.

                If the dithering cause flickering at 15hz we should know it from reviews like notebookcheck?! They use an oscilloscope to measuring PWM. Is it enough to find also this type of flickering right?!

                  Lauda89 If the dithering cause flickering at 15hz we should know it from reviews like notebookcheck?! They use an oscilloscope to measuring PWM. Is it enough to find also this type of flickering right?!

                  It is a patent so isn't a verified scientific article, we can only trust that it is the right information. Also 15hz may be a worst case scenario.

                  When notebookcheck test the devices for flicker/PWM, we don't necessarily know if they're running an OS on the machine at the time. The panel simply has to be switched on to measure PWM (AFAIK). Intel's testing came back inconclusive with good/bad drivers. It could be that the tests out there at the moment are only looking for 'on/off' strobing and not something more sophisticated like dithering patterns.

                    Lauda89 If the dithering cause flickering at 15hz we should know it from reviews like notebookcheck?! They use an oscilloscope to measuring PWM. Is it enough to find also this type of flickering right?!

                    I think it's almost impossible to see temporal dithering with their setup. The typical sensors (Notebookcheck uses a Thorlabs one) pick up the light from multiple pixels at least plus the backlight in-between. And the difference in light output caused by temporal dithering in this scenario is super small. Maybe on a big TV screen with low pixel density you could pick something up. In a spectogram overview perhaps, as small spikes at 15/30 Hz.
                    They certainly don't test for that. They look for strong flicker but not super tiny ripple. You wouldn't see this on a typical oscilloscope screenshot. The line would appear as flat.

                    Edit: Now that I think about it, maybe we could change our oscilloscope setup from the old thread to something that's more likely to pick up such small flicker. The sensor would need to be covered so that it only picks up a small fraction of the display and if we then greatly increased the resistance (which on one hand lowers the response time, which is not important for low frequencies like 15 Hz, but on the other hand amplifies the voltage) and have an oscilloscope capable of switching to AC mode, we might be able to see something. Haven't tried that yet, not sure if I find the time.
                    https://ledstrain.org/d/312-homemade-oscilloscope-to-detect-pwm-diy-guide

                        KM Just a thought, but has anyone actually spoken to notebookcheck about developing a test for temporal dithering? Of course, it'll be easier when you can intercept the signal going to the display rather than having to test with a spectrogram on the laptop display itself, but maybe there's a way..

                          Sunspark It’s in the original post actually! But for reference:

                          Intel Date: 17/06/2016
                          Intel Driver: 21.20.16.4475

                          Also see the OP in the discussion I started named Temporal Dithering - My Solution for full details of the journey to finding this working setup. 🙃

                              I completely forgot about this thread:
                              https://www.eevblog.com/forum/testgear/beginner-needs-to-record-dignals-from-a-display-backlit/?all
                              Posts #22 and #25 - the user tells us about finding 15 Hz temporal dithering in his spectrum analysis. He also speaks of some graphics drivers enabling temporal dithering when you change gamma or color values. I bet this happens on Android, too. I know it happened to me back on Windows XP whenever I changed the gamma values (immediate eye strain), so he probably knows what he's talking about. It must be an ages-old basic driver function to ensure colors are reproduced correctly when gamma/color or other adjustments are enabled. Maybe some drivers enable this by default? With no user setting to turn it off. I also remember this eye strain mode got activated when I chose to rotate the screen. Another non-default value.

                                si_edgey So dependent on the hardware.. 4475 doesn't exist for the broadwell generation, but 4474 on the 20.* branch does. It sucks. I didn't try force-installing the 21.* version, though I probably could. 5166 is latest, but it's just a lame security update and meh. On the HD 6000, the Windows catalog wants to serve up 4531 which is ok, and as an optional, 4624 which I find to be the best one I have seen to date on a HD 6000 in W10. It's disappointing that there never will be driver updates for this GPU other than in the Linux world (slowly).

                                KM I believe the findings. I often use the 4624 driver in YCbCr mode, and it does not use dithering in that mode but RGB does.

                                    This type of driver doesn't work with i7 right that has UHD 630?

                                    si_edgey I got bored and force-installed 21.20.16.4475 into Broadwell as just Intel HD Graphics (no model number). It works and it's better than the 20.*.4474 for Broadwell. It's not absolutely perfect (may not be possible) but it's definitely usable. Now I am contemplating looking up the driver version of what the work thinkpad used to have and seeing if I can locate that in the catalog and force installing that as well for comparison.

                                    Edit: Force-install, while works, comes with sacrifices.. one gives up DirectML and OpenGL.. not worth it.

                                      si_edgey I’m trying to recreate the exact same configuration you have and I’m not able to find this intel driver anywhere (21.20.16.4475). The drivers on Dell’s site are all newer. The only matching version I found elsewhere was for the HD 620 or a WHQL version which I have no idea what that means. Do you mind posting a link to download this driver?

                                      P.S: This is my first post here but I’ve been following your temporal dithering thread for a while. I just received my XPS 9560 FHD from eBay today and am on my way to testing your exact setup.

                                          davidgrey Hi @davidgrey, really hoping it works for you. Here's a link to the Intel driver:

                                          https://drivers.softpedia.com/get/GRAPHICS-BOARD/INTEL/Intel-HD-Graphics-620-Driver-2120164475-for-Windows-10-64-bit.shtml

                                          Extract the cab file contents to a folder and then in Device Manager on the Intel display adapter, right click and update the driver, then browse for the driver on your computer and tell it where the extracted cab contents are. It'll pick up the driver from there and install it for you.

                                          I'm happy to help you with anything at all - try to set everything up like in the latest update of the original post of the thread you're talking about and email me on my username @yahoo.com if you need any extra help at all. Fingers crossed for you for a good result. 🤞

                                              si_edgey Thank you so much! I’ll keep you posted on my progress! I’ve been pushing through on my Mackbook pro 2015 for a few years and just recently it became unbearable. Really hoping this will work for me.

                                                @si_edgey i noticed you are using Windows 10 Home. I installed the May 2020 version but it’s the Windows Pro version and I’m not sure I can change to Home since it defaults to the Pro when I install it. Is there any significant difference between the two?

                                                  davidgrey There is no difference, all editions are the same OS just different additional features are available or not. Stick with the Pro you have already.

                                                      dev