bqqq3rcv

Dynabook N156BGA-EA3; HP 255 G5 BOE06A5; E6520 LTN56HT01; Lenovo E440 LP140WH8-TPD1
I wonder what you can do with those data :-).

https://www.panelook.com/modelsearch.php?keyword=N156BGA-EA3
https://www.panelook.com/PT156WHM-N10__15.6__overview_42420.html
LTN56HT01 ??
https://www.panelook.com/modelsearch.php?keyword=LP140WH8-TPD1
at first glance, all these panels are 6bit
26 x 26 x 26=262144 (262K colors)
at a minimum, this means that such panels are not capable of displaying 8-bit color without tricks like dithering 🙁
and we return again to the questions: does dither affect strain? how to disable dithering guaranteed? What samples can be used to verify that dithering is disabled?

I continued my experiment at Windows 7 Classic :
(6bit+frc HP 27es HDMI) in combination with ATI HD4850 + reg tweak "HDMI_DisableDither"=dword:00000001 give banding on the sample http://www.lagom.nl/lcd-test/gradient.php
(6bit+frc HP 27es DVI) in combination with ATI HD4850 + reg tweak "TMDS_DisableDither"=dword:00000001 NOT give banding on the sample 🙁 WTF ?

bqqq3rcv

in the right part Ditherig is on (banding visible):

By "Dithering is on" you mean:
dithering.exe launched with the "temporal" option is selected ?
or Dithering.exe launched with the "disable all dithering function" option is selected ?
Do you feel a difference in eye strain?

        glvn

        glvn or Dithering.exe launched with the "disable all dithering function" option is selected ?

        Yes.

        glvn Do you feel a difference in eye strain?

        I think it is better with ditherig on than without it (but only when using intel iGPUs). I still struggle to find out exactly what my is main problem causing eyestrain. I very much appreciate your ideas about panels and their properties in this respect. I also tried intel igpu+6bit-frc external display over hdmi and I find out that when ditherig is on (all dithering disabled) that bands when watching gradient are not so much visible as compared to laptop panels. Thank you for pointing that out.

        • glvn replied to this.

                bqqq3rcv

                I also tried intel igpu+6bit-frc external display over hdmi and I find out that when ditherig is on (all dithering disabled) that bands when watching gradient are not so much visible as compared to laptop panels.

                What full model name was the external display?
                It would be interesting to see the results of dithering.exe + igpu on a real 8-bit panel (like Gigabyte M32Q)...

                Enabling color management (.icm color profile) also affects the smoothness of gradients - it worsens them. I have long come across an assumption that enabling the color profile automatically turns on dithering on ATI/AMD gpu drivers. In any case, it is better to check whether you are using color management, as this may further affect the smoothness of the gradients.

                    glvn What full model name was the external display?

                    The Viewsonic I mentioned earlier (besides that I have a few old CCFL displays but these are probably all 6bit)

                    glvn Enabling color management (.icm color profile) also affects the smoothness of gradients - it worsens them

                    I will try to find out how to disable colour management.

                    I also tried to override EDID (by changing info about supported bit depth) but had no success so far (i.e. it had no visible effect).

                    • glvn replied to this.

                          bqqq3rcv

                          I also tried to override EDID (by changing info about supported bit depth)

                          what software do you use to read EDID? Perhaps I misunderstand the information from the edid, but I still can’t determine the bit depth of the panel from the edid information 🙁

                              5 days later

                              glvn I used this one: https://www.deltacast.tv/products/free-software/e-edid-editor . It is necessary to use email address for download, so I used temporary email address for that. It has GUI that allows you to change values. There is an option to view changes in HEX format (in red). I used the changed value to add a key to registry (I copied the original value and edited it). There is a website from which I found out what has to be changed in registry: https://migueltek.com/tool/simple-edid-editor-for-win-11/ ;

                              So, We just need to write the modified EDID Block 0 information to "0" in the following path (0 means override EDID Base Block), value type is "REG_BINARY", length is 128 bytes.HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY\%MONITOR_INSTANCE_PATH%\Device Parameters\EDID_OVERRIDE

                                  I also made a discovery about gradient picture available at lagom. It actually can be zoomed. So when I tried ryzen 5600g + 6 bit-frc CCFL panel in win10 with zoomed gradient and ditherig I could see changes to the picture across all ditherig settings. Now I can confirm it works with integrated vega gpu.

                                  • glvn replied to this.

                                    bqqq3rcv

                                    So when I tried ryzen 5600g + 6 bit-frc CCFL panel in win10 with zoomed gradient and ditherig I could see changes to the picture across all ditherig settings.

                                    Can I ask for more details about the computer configuration?

                                    1. motherboard model name
                                    2. monitor connection type VGA/DVI/HDMI/DP ?
                                    3. windows 10 version/build
                                    4. GPU driver version
                                    5. ditherig.exe version (2,1 ??)
                                      and the most important question: is there a difference in eye strain from different settings ditherig.exe ?
                                        4 days later

                                        bqqq3rcv This is big security risk installing these apps. Maybe its better if someone shares how to use database csv in this app and how to manually apply needed changes.

                                            9 days later

                                            glvn 1. ASUS PRIME A320M-A; 2. DVI; 3. 19045.3086; 4. 23.2.2?; 5. 2.1.

                                            Whenever I used ditherig my eyestrain lessened (even though I haven't spent enough time with this computer to verify this). Nevertheless, I can confirm that picture displayed changed with various settings of ditherig.

                                              Donux Thank you for your security warning. I actually only used this program to read the data and do the change manually (by making a registry file).

                                                glvn This software can read you current EDID data. After launch, I enabled the hex view, and only after then I made changes to the opened EDID. Whenever a change was made, I could see that changed field turned red (at least two turn red, the changed value and checksum). After that I found the EDID in windows registry, exported it to a reg file and made a new key with overridden EDID (I actually edited the EDID saved by the changed values from the program in red) in the reg file and then I added it to registry and then restarted windows. I believe that overriding EDID to higher colour depth value will cause a black screen. I actually tried to lower it from 8bit to 6bit. I still don't know if this was a waste of time.

                                                  2 months later

                                                  I am using Linux and a Sony TV as the display. APU is Ryzen 5600G, so the driver in use is amdgpu. There are some Linux desktop applications that make use of GPU acceleration and cause instant eye strain. Let's assume that temporal dithering is the cause, and that the driver enables temporal dithering for certain GPU-accelerated applications. I am wondering if modifying the display's EDID information in the right way could make the driver stop doing that.

                                                  This is the EDID as seen in xrandr --props, in hexadecimal bytes:

                                                  00 ff ff ff ff ff ff 00 4d d9 04 7e 01 01 01 01
01 1c 01 03 80 5f 36 78 0a 0d c9 a0 57 47 98 27
12 48 4c 21 08 00 81 80 a9 c0 71 4f b3 00 01 01
01 01 01 01 01 01 08 e8 00 30 f2 70 5a 80 b0 58
8a 00 b8 17 32 00 00 1e 02 3a 80 18 71 38 2d 40
58 2c 45 00 b8 17 32 00 00 1e 00 00 00 fc 00 53
4f 4e 59 20 54 56 20 20 2a 30 30 0a 00 00 00 fd
00 17 3e 0e 88 3c 00 0a 20 20 20 20 20 20 01 6a

02 03 59 f0 5b 61 60 5d 5e 5f 62 1f 10 14 05 13
04 20 22 3c 3e 12 16 03 07 11 15 02 06 01 65 66
2c 0d 7f 07 15 07 50 3d 07 bc 57 06 00 83 0f 00
00 6e 03 0c 00 20 00 b8 3c 2f 00 80 01 02 03 04
67 d8 5d c4 01 78 80 01 e2 00 cb e3 05 ff 01 e5
0f 03 00 00 06 e3 06 0d 01 01 1d 00 72 51 d0 1e
20 6e 28 55 00 b8 17 32 00 00 1e 00 00 00 00 00
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 93

                                                  Can you help me find out which bytes to modify? I could then try to load the modified EDID as a kernel parameter and test if the changes fixed the problem. This is all assuming that xrandr displays correct EDID data, and that passed EDID kernel parameters are actually respected by the driver.

                                                  • JTL replied to this.

                                                    KM Can you help me find out which bytes to modify? I could then try to load the modified EDID as a kernel parameter and test if the changes fixed the problem. This is all assuming that xrandr displays correct EDID data, and that passed EDID kernel parameters are actually respected by the driver.

                                                    I have reason to believe in some situations (not sure about amdgpu driver) that custom EDID binaries are only loaded after initialization of the display output, which may use the existing EDID received from the sink device.

                                                      dev