hpst @JTL I am clearly out of my depth but in @deepflame's situation he said a software setting on the GPU was set to "10-bit mode"...and even though the laptop's panel wasn't 10-bit this fooled the apps requesting it into believing they didn't need to dither to get that 10-bits so it sounds like the apps trusted the gpu and not the display?

I don't have the same software & hardware to test in front of me so I don't know what his setup is doing.

hpst Is that just obviously stupid and I am too ignorant to how everything comes together for this to even be possible? What is it in linux for example that says what quality to display the graphical DE in and triggers dithering to get it? Is it the gpu driver? Xorg? As far as I know laptop panels are "dumb" and aren't making that choice internally like some desktop monitors can right?

No. @deepflame's experiment is the first I've heard of this, but I'd need the same hardware in front of me in order to understand what it's doing and if I can port it to be more universal.

  • hpst replied to this.

        JTL I believe it is a Thinkpad W500? If you don't need one in awesome condition that's under 100usd on eBay. If nobody else would pitch in a few bucks toward it, I'd be willing to just order it for you myself if you believe it could at least rule this in or out as an option since it/along with your work/ are the only actionable ideas I have seen. I can eat beans for a while if it could even possibly lead to a breakthrough.

        • JTL replied to this.

            hpst I want to get the capture card stuff under control first which requires some free time from other duties to write the image capture/comparison program (easy, don't worry)

            Thanks though. Means a lot to me.

                JTL Sorry for being pushy or overeager sounding. I'm desperate as you know so I tend to get intense and a burst of "hope energy" when I think there is a possibility in the darkness. If you know anyone else who would be good to approach for this particular angle please let me know. I am still interested in contributing to the card as well since any solution is a godsend...and I am surely missing something about this idea as things are never as easy as toilet logic makes them sound.

                  KM

                  KM Some cards probably ignore it and always dither.

                  s/some/almost all

                  If they respected the EDID. I think for the most part we wouldn't have this issue.

                  • KM likes this.

                      If the EDID was respected and reliable then the FireGL/Firepro cards would not have the option to enable 10bit color mode. If it was reliable then the card would just rely on the info it got from the screen, right?

                      I think enabling this "10-bit color" mode just tells the graphics card what to do when 10bit color is requested. Should it dither or not.

                      So I am not sure if fooling the system to believe it has a 10bit display attached solves the dithering issue.

                      For me as a software engineer the better approach would be to see when the problem occurred.
                      My idea would be the following:

                      • get a setup that does not dither on an old Linux box
                      • update the graphics driver/kernel so long until dithering starts.
                      • check the graphics driver source and look for DITHERING constants or other code changes that might have affected this.
                      • reverting the code change, compile the driver/kernel module and see if it helped.

                      Think this would help us to see what really is going on. Again, I think the issue is color depth (temporal dithering)

                      The same we could do for Chrome as well. As people have pointed out old versions of Chrome were working well. Maybe we can find old versions of Chromium (best portable versions) somewhere that we can test and see which version introduced the different rendering method that causes strain.

                      As Chromium is also open source we could check for code changes here as well. I would assume they could be massive but it is work a try. In the case of Chrome I think we are talking about an updated version of harfbuzz for font rendering here ( https://www.freedesktop.org/wiki/Software/HarfBuzz/ ).

                        AgentX20 I'm OK with my Dell 2407/2410 or similar CCFL screens

                        I heard that the 2410 has PWM and the 2407 does not. Can you confirm that? Are you ok with PWM on the 2410 If it has?

                            deepflame If the EDID was respected and reliable then the FireGL/Firepro cards would not have the option to enable 10bit color mode. If it was reliable then the card would just rely on the info it got from the screen, right?

                            In theory.

                            A good friend of mine had one of those off-brand Korean TV's that claimed by the EDID it supported 1920x1080 resolution, but from memory it could only scale 1280x720 to a 1920x1080 image.

                                deepflame Would you be willing to work on this? My primary need is desktop linux but I am not a power user or programmer. I have a desperate and urgent need for a solution because of my life situation and will lose a once in a lifetime chance at getting a life back under me if I can't use a damn computer anymore. At this point I don't even care what distro. Is there anything we could do to help? If it's a funding issue I know I would put in some and hope others can too. Personally I think we should crowdfund this in general so it's someone's primary focus. We have a few very smart people here but there isn't any full time work being done on this.

                                    hpst Make my university courses and other work go away and I'll do it 😛

                                    • hpst replied to this.

                                        JTL Oh I hope you didn't think that was a dig at you in any way! You are one of the smart people I was referencing...and I knew you didn't have the time or energy to work on this full time. I wish I had the knowledge and skills or knew someone we could crowdfund payment to so it's their job to solve this. I hate having theories and hopes with no real way to act on them directly and it seems any attempt to get outsiders to help, like the intel thread, get shut down or time out. It took forever to get them to even interact...then they did and did a lot of testing but not the one thing most likely to be our problem and said basically "cant reproduce not a problem".

                                        • JTL replied to this.

                                            hpst Oh I hope you didn't think that was a dig at you in any way

                                            I didn't, don't worry.

                                              JTL Hmm, not sure if I understand. I meant that if the EDID was respected then there was no need for the driver to offer the 10bit mode. It is offered because the EDID info is not reliable and the card cannot tell if it is really a 10bit screen that it connected. Hence there is an option to turn 10bit mode on explicitly.

                                              So yeah, your example shows that the EDID is not reliable. Guess this is what you wanted to say?

                                                hpst Would you be willing to work on this?

                                                Well, I think it would be interesting to find out. Not sure if I have too much time though and I do not want to feel the pressure of someone whose life depends on this in my back. Hope you understand me correctly here. I am glad to help but I would assume that there are other possibilities as well.

                                                E.g. currently I work with Windows 7. If I would have to do development work I would rather work with a Linux machine (easier to use Git, Ruby, Node, ...). So my idea is to use my Windows machine to connect through SSH to a Linux machine (or my Mac) and do development stuff there.

                                                Hmm, ok guess your problem would also be the code editor. Well, I use VIM in the terminal. So this would not be an issue for me. However you could also use SMB or something to expose the Linux files over the network and you could use Sublime Text or whatever you prefer on Windows then. Or you can run a Linux virtual machine with VMware or Virtual Box.

                                                • hpst replied to this.

                                                    deepflame I guess I should have said I stongly WANT linux rather than NEED it. I simply need a browser and text editor to work. For me at least no modern OS and hardware is strain free so if dithering is our problem and it can be shut off then I'd just as soon do that in linux and keep using it since it removes other issues I don't want to deal with like walled gardens, expensive hardware, spyware OS nonsense etc.

                                                    I had only asked because you mentioned being a SWE and brought linux up so I assumed it was your preference. About the pressure...right now we don't have any organized effort toward this so anything done is a positive if nothing else to rule out a theory. I mentioned funding because it could motivate someone to work in a more dedicted manner where they aren't squeezing it in between al the other demands on their time.

                                                      deepflame I heard that the 2410 has PWM and the 2407 does not. Can you confirm that? Are you ok with PWM on the 2410 If it has?

                                                      I'm pretty sure the 2407s DO have PWM and that it kicks in below around 50%. I run them at the top end of the brightness curve and use Nvidia software to reduce the brightness. To be honest I don't think the PWM on these has bothered me anyway. I think my problems lie elsewhere.

                                                      I've not used the 2410 extensively (its more of a backup in case one of my two 2407s fail), but when I've used it it's been A-OK - even on the end of laptop that gives me trouble if I use its built-in screen.

                                                          AgentX20 I've not used the 2410 extensively (its more of a backup in case one of my two 2407s fail), but when I've used it it's been A-OK - even on the end of laptop that gives me trouble if I use its built-in screen.

                                                          I wonder why this is? IF dithering is the issue does that mean somehow certain monitors are not allowing it/deciding to ignore it? I have seen several people claim laptops were ok but when plugged in to external displays they had strain...this is the first time I think I have seen someone say the laptop display is bad but the external was ok. So frustrating and so many variable...connection types, protocols etc.

                                                              hpst I wonder why this is? IF dithering is the issue does that mean somehow certain monitors are not allowing it/deciding to ignore it?

                                                              If we knew the answer this I think we'd be well towards solving this dilemma.

                                                              I think some screens are bad, some video cards are bad, and some software is bad. Put them together in various combinations and who knows what you'll get.

                                                                pro tip. if you turn the color all the way down.. no FRC/ color temporal dithering. #changemymind

                                                                  has anyone here actually accomplished true 8-bit with FRC turned off?

                                                                  dev