Custom Resolution Utility (CRU)

[hannes69]

(02-26-2025 06:01 AM)ToastyX Wrote:  The HDMI output shouldn't be able to corrupt the EDID. Did you unplug the projector's power or move it at any point? Maybe it was already corrupted for a while and cached in the projector's memory.

My projector has a permanent flashing blue power led when turned off, so after using the projector I pull the power plug... Once every day...
The projector got some 'treatment' from me several times (opening, cleaning, clean color wheel, new lamp, remove dust particels from DLP chip,...) and I have used a self made RS232 cable to access the service menu. Yes, maybe there got something wrong...
But interesting that it worked with my old PC but immediately not with my new PC.

(02-26-2025 06:01 AM)ToastyX Wrote:  Are you able to read the EDID with EDWriter? If it's corrupted, it will offer to fix the data. Then you can try writing the EDID and hope it's not write protected

Didn't know that you have written a great tool beside CRU utility
EDWriter connected, read the data, offered a fix, applied the fix and finally could write to the EEPROM. Puh, lucky in this regard, no write protection I know from the inner parts of the projector, that the EDID chip is not directly accessible, so that would have meant another effort, doable but like always with this projector - effort.
After that everything went back to normal, the projector is recognized as HDMI and all resolutions are available
Many thanks for the link to another great tool made by you, helped me!
In the meantime I checked my old PC and found the EDID data of my projector in Windows registry. I compared the raw data with the new corrected data from EDWriter and used an online EDID decoder. Seems to me that the only difference I could detect was the checksum. The corrected EDID has a valid checksum and the old one on my old PC has an invalid checksum.
So my old PC maybe knew the working setting, then there was corruption and it still worked by ignoring the invalid checksum. But my new PC got the invalid checksum directly from the start and maybe didn't like that? How can a checksum be wrong while at the same time all data being correct, is this possible?

(02-26-2025 06:01 AM)ToastyX Wrote:  DVI is always full range RGB. To get an HDMI signal, add a CTA-861 extension block and an HDMI data block.

Good to know, thanks.

Beside that: The whole video timings topic seems to be very complicated especially in HTPC and in combination with audio clocks.
In my new PC another strange thing is happening: After a fresh boot or reboot the refresh rate of a display is way off to the calculated one from pixel clock and horizontal/vertical totals. Both madVR and vsynctester.com report the same and totally off refresh rate. E.g. a perfect 100.00000 Hz refresh rate (for 25fps and 50fps movies) leads to 99.81 Hz. Not acceptable for a HTPC. But: systematically after sleep mode and wake up in Windows 10, the refresh rate is more or less like the calculated one. Simply found out by accident. Made my tests, nothing worked as intended, got away some time and PC was in sleep mode, came back, woke up PC and changed nothing, and magically everything OK. And now found out that it is systematically like that. So from a clean boot wrong things happen and from a sleep wake up (where many people have big problems) the right things happen... Strange world. Maybe it has something to do with clocks, but I have no clue what to do here. The Ryzen 8600g is my first iGPU, before I always had discrete GPUs. I don't trust the iGPU totally, the graphics outputs are part of the mainboard and the GPU is within the CPU, so maybe the clocks are spread and not 'unified' on one device like on a discrete GPU, who knows...

[ToastyX]

(02-27-2025 03:27 AM)hannes69 Wrote:  In the meantime I checked my old PC and found the EDID data of my projector in Windows registry. I compared the raw data with the new corrected data from EDWriter and used an online EDID decoder. Seems to me that the only difference I could detect was the checksum. The corrected EDID has a valid checksum and the old one on my old PC has an invalid checksum.
So my old PC maybe knew the working setting, then there was corruption and it still worked by ignoring the invalid checksum. But my new PC got the invalid checksum directly from the start and maybe didn't like that? How can a checksum be wrong while at the same time all data being correct, is this possible?

Are you sure it's only the checksum? The EDID in the registry is not used for caching, and I don't see how it could be in the registry if the checksum was wrong because it should have been rejected unless you were using a driver that didn't reject an invalid checksum. The EDID is read every time the projector is connected and every time Windows boots. Maybe static electricity happened to flip a bit somewhere when you connected it to the new computer.

(02-27-2025 03:27 AM)hannes69 Wrote:  Beside that: The whole video timings topic seems to be very complicated especially in HTPC and in combination with audio clocks.
In my new PC another strange thing is happening: After a fresh boot or reboot the refresh rate of a display is way off to the calculated one from pixel clock and horizontal/vertical totals. Both madVR and vsynctester.com report the same and totally off refresh rate. E.g. a perfect 100.00000 Hz refresh rate (for 25fps and 50fps movies) leads to 99.81 Hz. Not acceptable for a HTPC. But: systematically after sleep mode and wake up in Windows 10, the refresh rate is more or less like the calculated one. Simply found out by accident.

I'm not sure what would cause that. Are there multiple versions of 100 Hz in the EDID? Like if there is one as a detailed resolution and another as a TV resolution, it might be switching between the two.

[ToastyX]

(02-26-2025 10:58 AM)juanme555 Wrote:  

(02-26-2025 06:01 AM)ToastyX Wrote:  Is there a CTA-861 extension block with an HDMI data block? Is the max TMDS clock set?

how do i create one and what do i put in it?

What do you mean? Is it there or not? If not, you can just add the extension block and HDMI data block manually.

[juanme555]

(03-02-2025 03:05 AM)ToastyX Wrote:  

(02-26-2025 10:58 AM)juanme555 Wrote:  

(02-26-2025 06:01 AM)ToastyX Wrote:  Is there a CTA-861 extension block with an HDMI data block? Is the max TMDS clock set?

how do i create one and what do i put in it?

What do you mean? Is it there or not? If not, you can just add the extension block and HDMI data block manually.

there was no data block, i added one CTA 861 to no good, i tried adding both hdmi extension blocks and nothing gave me a signal, i gave up its all good, i'll get by with this disgusting 1080p lcd for 2 days max, tomorrow or the day after that im getting a gpu and im choosing between vega56/64 and GTX 1070/1080/1080Ti , i want interlaced resolutions on windows 10.

[hannes69]

(03-02-2025 03:04 AM)ToastyX Wrote:  Are you sure it's only the checksum? [...] Maybe static electricity happened to flip a bit somewhere when you connected it to the new computer.

Shame on me, you're right. I took another look and indeed - 2 bits in a row flipped And really annoying - it was a 'wrong' manufacturing week. Really. The only difference was, that the projector was now produced 3 weeks later than before. And for that I am granted with a 640x480 resolution? Hmmm...
I could only check the first 128 bytes in registry, the rest (extension block) is in nirvana for me. At least it seems that the extension block isn't stored in Windows.

(03-02-2025 03:04 AM)ToastyX Wrote:  I'm not sure what would cause that. Are there multiple versions of 100 Hz in the EDID? Like if there is one as a detailed resolution and another as a TV resolution, it might be switching between the two.

Made further tests.
To rule out your theory I made a custom resolution of 80.00000Hz which should be nowhere in EDID. Same problem. And interesting, all refresh rates are low by a fixed amount, 1880 ppm too low. Means a framedrop every 11 seconds for 50fps video.
This is the case when using a DP within USB-C connection between PC and monitor.

I also tried a HDMI connection between PC and monitor.
The refresh rates are low as well, but now not by a fixed amount, but dependent on refresh rate in the range of 200 ppm - 500 ppm too low. And - in this case the sleep - wakeup procedure doesn't change anything.

Last test: HDMI connection between PC and my projector. Refresh rates are as calculated, stable, independent of sleep-wakeup.

In all three cases/tests the current used timings are shown correctly in AMD software (meaning the ones I want and have chosen).

So there are differences dependent on device (today's monitor, 17 years old projector) and used GPU output (USB-C vs. HDMI on same display device).

My theory is still that it has to do with the iGPU architecture from the Ryzen 8600g and faulty clock design.

At least the real life problem is not too big for me. For the home cinema with the projector the refresh rates are accurate like described, and when watching videos on the monitor like youtube then you're always in no man's land with different framerates. When really watching longer videos on the monitor, I can still use the sleep-wakeup trick.

I wanted to mention the problem here, because I thought it could be interesting for you and some people here, I haven't read about this in the net, so something new

Ok, the EDID problem wasn't that new, there are several reports of people complaining about a spontaneous generic non-pnp monitor / 640x480 max resolution. Now I learned the background of that For me it's strange though that this can happen, the EEPROM should be electrically more robust. There was no lightning strike to my PC...

[Monkeyking]

Hello ToastyX,

Someone found out the registry key to disable DSC on the driver level, i'm not sure what else it could break. (https://www.reddit.com/r/OLED_Gaming/com...and_other/)

Code:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0000

"EnableTiledDisplay"=dword:00000000

It definitely allowed me to get DSR/DLDSR back on my S95D, with CRU edits to remove 4096 res and MPOs.



Edit : This disables RTX HDR and NVAPI HDR :/

[Case]

(03-03-2025 08:48 AM)Monkeyking Wrote:  Someone found out the registry key to disable DSC on the driver level, i'm not sure what else it could break.

Code:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Class\{4d36e968-e325-11ce-bfc1-08002be10318}\0000

"EnableTiledDisplay"=dword:00000000

This is really cool. For the first time, I have MPO's available in Game mode on my monitor, and DLDSR is indeed back, while 4k 144Hz mode is now unavailable, just 120Hz (so as expected). I will definitely be testing this to see if it breaks anything.

[juanme555]

Okay now, based lord Toasty , i have now bought a GTX 1080 Ti and im using the same HDMI to VGA adapter i was using with intel igpu that was capable of doing , but now on nvidia drivers im on drivers 572.60 and interlaced resolutions are not supported at all! when i try to create one on the control panel it just says "this resolution isnt compatible with your display" or smth like that.

[p14t1pu516]

(01-03-2023 05:56 AM)ToastyX Wrote:  

(01-02-2023 10:07 PM)p14t1pu516 Wrote:  hi toastyx, happy new year. It worked exporting and installing the .exe file. Everything works well. When pc starts, and the motherboard logo appears, it blinks for a moment, I guess that's when it changes to dvi. it's okay? thanks for all

Changes don't take effect until Windows starts and the graphics driver loads. If it works then there's no problem.

Hello Toastyx, i hope you are well.
Since the last time I wrote I had no problems.
The other day my SSD stopped working and I had to install Windows on an HDD I had on hand. The problem is that when using cru, deleting the extension block and making sure 1366x768 @ 60 Hz be the first detailed resolution, using the .exe export method (this is the way I found it to work), If I use restart, it works, but when I restart the PC, gets stuck during startup and never start.

To make matters worse, when starting in safe mode, when I want to use reset all the system gives me an error and says that it cannot find the file, so I have no choice but to format it with each attempt.

Could you help me? Is there another way to solve the problem? thank you so much for your time.
Below is a summary of the original problem:

------
Hello, I bought a new pc: i3 8100, 8gb ram, integrated intel graphics 630. When I connect to the TV, the image is not clear, the photos are dark, the borders of letters and the mouse cursor are contoured...poor quality, the real photos look like photos of video games...etc. .it's very noticeable when connecting via VGA, looks perfect.
I've been having this problem for days, I did everything, Format, uninstall and install drivers, try different TVs, cables... everything.
Tv led lce24xh12 hd sanyo -resolución hd (1366 x 768), and windows recommend that resolution.
I tested the tv on two other tvs noblex 32 and 46 and have the same problem.
------

[ToastyX]

(03-05-2025 03:19 AM)p14t1pu516 Wrote:  Hello Toastyx, i hope you are well.
Since the last time I wrote I had no problems.
The other day my SSD stopped working and I had to install Windows on an HDD I had on hand. The problem is that when using cru, deleting the extension block and making sure 1366x768 @ 60 Hz be the first detailed resolution, using the .exe export method (this is the way I found it to work), If I use restart, it works, but when I restart the PC, gets stuck during startup and never start.

To make matters worse, when starting in safe mode, when I want to use reset all the system gives me an error and says that it cannot find the file, so I have no choice but to format it with each attempt.

That error message is not coming from reset-all.exe. What is the exact error message? You shouldn't need to reformat. The exported .exe file can also uninstall the EDID override.

If the PC hangs on boot after using CRU, that can only be a driver bug. Have you installed the latest graphics driver for your Intel GPU?