Monitor Tests Forum

(08-17-2017 06:37 AM)ToastyX Wrote: [ -> ]

(08-16-2017 06:12 PM)Chrismartin76 Wrote: [ -> ]Thanks!
Under Device Manager, I have Intel HD Graphics 630 and NVIDIA GeForce GTX 1050. Both are currently enabled. I have Windows 10. The first detailed resolution is the same resolution (2560x1404) but 59.002 Hz. I also installed the HDMI extension block and the first detailed resolution there is 2560x1440 @ 59.950 Hz. When I installed the extension block, I kept the default settings because I don't understand what exactly the extension block does.

I just changed the refresh rate on both to 60.000 Hz and did a quick test. The monitor was OK after waking up. I'll post again if the problem recurs, but please let me know if I should do anything about the switchable graphics. Thanks!

I asked mainly because I was trying to find out what driver was responsible. This sounds like an Intel driver bug. I'm not familiar with Intel GPUs, so I don't know if there's a workaround. You shouldn't need to define the same resolution twice. Just having it as the first detailed resolution in the main window should be enough.

Usually the default extension block already includes HDMI support, but CRU can't read the extension blocks with Intel GPUs. If the default doesn't work, I was going to suggest importing hdmi-audio.dat, but it seems you're already trying a custom extension block.

OK I removed the second instance of the detailed resolution. The screenshots below are my settings on the main dialog and the block dialog. I don't want HDMI audio support at the moment, so I used hdmi.dat rather than hdmi-audio.dat.

Any news on forcing Freesync over DVI on 144hz monitors? I can convert to single link HDMI and get 60hz freesync

(08-18-2017 07:19 AM)Prefix Wrote: [ -> ]Any news on forcing Freesync over DVI on 144hz monitors? I can convert to single link HDMI and get 60hz freesync

That's not up to me. The FreeSync data block in CRU is provided mainly for editing HDMI FreeSync ranges. It wasn't intended to force FreeSync with DVI monitors. That's just a quirk with AMD's driver that happens to work for a few monitors.

(08-17-2017 03:54 PM)Chrismartin76 Wrote: [ -> ]OK I removed the second instance of the detailed resolution. The screenshots below are my settings on the main dialog and the block dialog. I don't want HDMI audio support at the moment, so I used hdmi.dat rather than hdmi-audio.dat.

That looks fine. That should be enough to get 2560x1440 @ 60 Hz with HDMI. I noticed in your screen shot that the monitor list is missing the "(active)" status next to the *. I don't know if this is related to your problem or if it's just an Intel driver limitation.

(08-17-2017 10:08 AM)paulobmarcos Wrote: [ -> ]I have a GTX 770. I know it is not the best video card, but can it actually be the culprit?

The problem is the graphics driver, not the video card. Assuming the test is accurate, the driver seems to be offering both refresh rates but is mixing them up sometimes. You'll have to ask NVIDIA why it behaves that way. It might be a driver bug.

Also, the GTX 770 doesn't have HDMI 2.0, so you'll only be able to get 3840x2160 @ 60 Hz with YCbCr 4:2:0. That's not related to the 59/60 Hz problem, but YCbCr 4:2:0 resolutions can only be defined in a specific way, so you won't be able to define custom detailed resolutions or delete TV resolutions to see if that works around the problem.

(08-18-2017 09:58 AM)ToastyX Wrote: [ -> ]

(08-17-2017 03:54 PM)Chrismartin76 Wrote: [ -> ]OK I removed the second instance of the detailed resolution. The screenshots below are my settings on the main dialog and the block dialog. I don't want HDMI audio support at the moment, so I used hdmi.dat rather than hdmi-audio.dat.

That looks fine. That should be enough to get 2560x1440 @ 60 Hz with HDMI. I noticed in your screen shot that the monitor list is missing the "(active)" status next to the *. I don't know if this is related to your problem or if it's just an Intel driver limitation.

Thanks. I happened to take the screenshots when I was working at a coffee shop so I wasn't connected to the bigger monitor.

(08-18-2017 10:31 AM)paulobmarcos Wrote: [ -> ]Would an adapter like the Club3D Displayport 1.2 to HDMI 2.0 be better for pc use? Or a card with an HDMI 2.0 port and a certified HDMI 2.0 cable are enough for YCbCr 4:4:4 4K 60hz?

Either will work.

Hi to all, i'm new and i'm here because i've downloaded CRU 1.30 for try to enable Freesync on my AOC U3277PQU monitor. I've done several test enabling freesync range etc, but none: Amd driver always write "freesync unsupported". Have i forget something or it's impossible ? AOC create another monitor, called AOC U3277PWQU and that's with freesync enabled, so i think is only a firmware question. I've got a Radeon 290x and DP cable

(08-16-2017 12:16 AM)ToastyX Wrote: [ -> ]

(08-15-2017 10:30 AM)sevs Wrote: [ -> ]Hi. Simple question: do you need two crossfire bridges to be able to use resolutions that are exceeding 300/330mhz pixel clock even though you have disabled crossfire in the driver settings? Or what else could be the reason to why i'm unable to set 1920*1440 at 160hz interlaced? 158hz (fieldrate 79hz) works fine.

Using Windows 7 x64, latest amd drivers, latest pixel clock patcher (maybe it's not compatible with the latest drivers (yet)?), dual 280x, two monitors connected, one to each gpu via dvi-bnc cables (VGA/RGBHV) (so no edid coming from monitors). Monitors are Eizo F931 and Eizo T960. I want to use 2048*1536i at 160hz on the F931 and try to see if 3840*2880i at 75hz or 80hz works on both of them.

What do you mean when you say you're unable to set the resolution? What happens? You only need two CrossFire bridges if CrossFire is enabled; otherwise, some games will not load correctly, but the resolution should still be available. This is separate from the 330 MHz dual-link DVI limit. VGA supports up to 400 MHz without the patch. You shouldn't need the patch for 2048x1536i @ 160 Hz, and I don't know if the video card can even output 3840x2880i @ 75 Hz because the pixel clock would be almost 600 MHz.

Thanks for your reply.
Anything with a pixel clock of more than 330MHz doesn’t show up in the List All Modes window or anywhere else in Windows. They do show up in the Radeon settings though.

I had a Sapphire HD7970 before, I did test out resolutions with pixel clocks of up to 680MHz on that card, so i would think the XFX 280X would be the same as they run on the same GPU chip.
I mostly tested out progressive resolutions though as the AMD driver calculated interlaced resolutions as having double the pixel clock than what they actually did have (i sold the card back in 2016).

I have been testing even more now, and it’s clear that there is a limit of exactly 330MHz on the DVI-I port.

I was thinking maybe the XFX 280x for some odd reason has an onboard active dvi-d to vga transcoder isntead of an internal analog RAMDAC but that sounds utterly weird and stupid, but who knows. Patching the driver should still fix the 330mhz dvi-d limit anyway, though?
Keep in mind that i’m adding custom resolutions inside of the Radeon settings, and not using CRU, as i still have issues with CRU, becuase of the (pseudo-) GPU scaling that is always enabled. Trying to enable and then disable GPU scaling after using CRU triggers the looping glitch again.

The XFX 280Xs also have dual DVI-I ports physically, but only the lower one (the one closest to the GPU PCB) outputs analog signals according to my testing. That one port is black and says “VGA” below it, in addition to a DVI sign. The other one is red, and only has the DVI sign below it, and doesn’t output VGA signals when i plug a CRT into it. Weird.

(08-16-2017 12:16 AM)ToastyX Wrote: [ -> ]

(08-15-2017 10:30 AM)sevs Wrote: [ -> ]I changed the monitor drivers in device manager from non-pnp to pnp, otherwise i would get crashes where the screens would continuously turn off and on and the only thing you could do was to hold power button/cut power. I think i resolved this issue now, but i'll test to make sure later.

This is an AMD driver bug that affects monitors without an EDID. The driver isn't really crashing. It's getting stuck in a loop trying to detect a monitor without an EDID. Running restart.exe can break the loop if you can manage to run it, but opening Radeon Settings seems to trigger the problem again. Are you saying using the "Generic PnP Monitor" driver works around the problem?

This is correct. In fact, i could sometimes see the Windows screen resolution control panel loop between “Display device on: VGA” and “Generic Non-PnP Monitor”, with a minimal but annoying flicker between each switch. Other times the monitors would loop between two different resolutions which meant they would switch off briefly between a resolution change, so you could only see what’s on screen for a split second every time. On second thought, it’s more like a glitch than a crash.

(08-16-2017 12:16 AM)ToastyX Wrote: [ -> ]

(08-15-2017 10:30 AM)sevs Wrote: [ -> ]I am using the custom resolution tool in the amd drivers because using cru turned on gpu scaling with no option to turn it off, and the custom resolution would not be displayed in the windows list all modes window, even though the khz rate etc on the monitor matched the resolution i had set in CRU. So basically the monitor reported for example 1920*1440 at 150hz interlaced, and the Windows CP would say 1600*1200 at 60hz progressive. The picture would be unsharp and obviously upscaled.

AMD's driver has (or had) an option to either use the EDID or to manually set the maximum resolution and refresh rate for VGA monitors. You might need to enable that option if you're using CRU since it creates EDID overrides. I don't know if that option is still available with the latest driver because they got rid of Radeon Additional Settings. The last version with Radeon Additional Settings is 17.7.1.

Also keep in mind when adding an interlaced resolution with CRU, you need to make sure the vertical resolution is halved, so 1920x1440i would be 1920x720 and 3840x2880i would be 3840x1440. CRU automatically does this for you if you enter the resolution first before enabling the interlaced checkbox. I don't know if the vertical resolution needs to be halved with AMD's custom resolutions.

The custom interlaced resolutions I’ve got working in Radeon Settings had input the correct values for the final resolution, but half the refresh rate (the field rate if you will). I could probably try to see what happens if i input the values like you explain it though.

(08-16-2017 12:16 AM)ToastyX Wrote: [ -> ]

(08-15-2017 10:30 AM)sevs Wrote: [ -> ]Also i have another issue where enabling Crossfire enables the output on my lower GPU instead of the upper one. Why does AMD radeon control panel identify my left monitor that is connected to the upper GPU as number 2 and the right monitor connected to my lower GPU as number 1, while the windows CP shows the opposite? Is it possible to swap primary and secondary GPU? I know i couldj ust swap the cables but my top GPU has a waterblock on it so i'd prefer for it to be the primary GPU. It wasn't like this before. This happened after reinstalling Windows and drivers.

I don't use CrossFire, so I'm not sure how this works. If the driver doesn't provide a way to set the primary GPU, then I don't know if you can change it. I would assume the GPU that shows the BIOS/UEFI screen would be the primary GPU, which is usually the first PCIe x16 slot by default. The order of the monitors in the Windows display settings shouldn't matter.

I partially resolved this issue by uninstalling the bottom GPU in device manager, then doing a scan for new hardware changes, every time i want to enable Crossfire. The problem still comes back after every reboot though, but that’s okay.

It was super weird because the UEFI would show up on the primary GPU, Windows bootanim on the primary GPU, then when the PC booted into windows the first monitor turned off and the second turned on, and password screen and everything else showed up on the second one.

Hello, first of all i am new to this forum but i have already used CRU often, the tool was very helpful so i greatly appreciate the work you put in.

I have a question (although my problem is already solved, i don't understand something) which i hope someone may shed some light upon.

I own a samsung cf791 monitor (21:9, 3440 x 1440 , 100hz, VA, that supports freesync). There is a wide reported issue (it seems 100% of users have this problem) that the monitor starts to flicker when freesync is turned on (more so if the framerates are low (roughly lower then 60fps, and therefore the monitor is running at lower hz because of freesync)
The freesync range used is 48-100hz. It works (tested with games and techdemos like gsync demo) and LFC also works.

Now i have read a lot, exchanged opinions and experiences with other users and tried a lot, and it seems at least so far i have found and fixed the problem, or at least the problem has disappeared: Someone posted that the pixel clock and h rate was wrong:

please correct me if i am wrong but as far as i read it pixel clock = v total * h total * refresh rate but when i openened cru and looked at the settings for the cf791 the pixel clock rate was too fast (it was set to 550 mhz and the h rate was too fast too)
but if we take the formula: 3600 * 1481 * 100 = 533, 160 mhz (lower then the pixel clock rate that was set) and the standard h rate was too high too ( i corrected it to the correct one)

After i did these corrections the flicker in freesync disappeared, now my question is were these standard settings really wrong and is it likely that these slightly too high pixel clock and h rate caused this flicker ?, and if so, how can it be that wrong values/settings are used for this monitor and whose fault is this (the radeon driver or samsung firmware)?

This is a short video i uploaded on youtube that shows the flicker that happened before i changed the pixel clock and h rate values, it is a short video of ROTTR running on 3440 * 1440p on a rx 480 with freesync enabled, although it was recorded only with a 30 fps camera, the flickering that one can see in the video is about what you see with your eyes if you sit in front of it
cf791 flicker
After changed the values this flicker is completly gone ( and freesync is still working)