Custom Resolution Utility (CRU)

[Alkerion]

(07-18-2026 02:52 PM)ToastyX Wrote:  

(07-18-2026 11:59 AM)Alkerion Wrote:  Hi,
I've installer CRU 1.5.3 on my W11 PC, I own a ASUS TUF GeForce RTX 5070 Ti OC Edition with a LG UltraGear 40WP95X-W (5120 x 2160 pixels, 39.70") monitor.
CRU doesn't detect it correctly, and resolution ara also missing.
Already tried reset-all.

Looks fine to me. Did you check the extension blocks? 5120x2160 is usually in the DisplayID extension block.

Thanks a lot, haven't seen that I can click on these lines to open sub-screens.

[Framerate]

For those of you using non-standard framerates like 47.952fps, 48fps, or anything else, I was wondering how you get audio to external speakers?

I’m using an AVR with external speakers and Google AI is adamant that AVRs, as well as HDMI splitters etc., will at minimum introduce microstutter or microjudder to custom framerates, if not convert them to 50fps or 60fps or create a black screen.

Meanwhile, end users are saying that it will work; however, they are probably going by eye test and I doubt they have ever tested it for microstutter or microjudder. Google AI also recommended “extend this display” at one point, but then at another said that it would also create these issues and possibly a lot of audio sync variance as well.

I want to interpolate mkv movies to 47.954fps and 48fps and output them at either 1x or 2x Hz without getting video artifacts or bad audio sync variance, and am wondering the best way to do it. I have an audio sync measurement device so I can correct for bad sync latency, but not for bad drift or variance. I need a steady number, even if that number is big.

[ToastyX]

(07-22-2026 07:15 AM)Framerate Wrote:  For those of you using non-standard framerates like 47.952fps, 48fps, or anything else, I was wondering how you get audio to external speakers?

I’m using an AVR with external speakers and Google AI is adamant that AVRs, as well as HDMI splitters etc., will at minimum introduce microstutter or microjudder to custom framerates, if not convert them to 50fps or 60fps or create a black screen.

Meanwhile, end users are saying that it will work; however, they are probably going by eye test and I doubt they have ever tested it for microstutter or microjudder. Google AI also recommended “extend this display” at one point, but then at another said that it would also create these issues and possibly a lot of audio sync variance as well.

I want to interpolate mkv movies to 47.954fps and 48fps and output them at either 1x or 2x Hz without getting video artifacts or bad audio sync variance, and am wondering the best way to do it. I have an audio sync measurement device so I can correct for bad sync latency, but not for bad drift or variance. I need a steady number, even if that number is big.

The video signal is a constant stream of data at a fixed rate, so there can't be stutter in the video signal itself, and sending audio to an AVR or splitting the signal wouldn't change that. Any visible stutter or judder would be on the software side. The biggest challenge is perfectly synchronizing the frame rate of the video player to the refresh rate of the display. If the video player and display both support variable refresh rate (FreeSync/G-SYNC), that would be the best solution.

[Framerate]

(07-23-2026 03:37 AM)ToastyX Wrote:  

(07-22-2026 07:15 AM)Framerate Wrote:  For those of you using non-standard framerates like 47.952fps, 48fps, or anything else, I was wondering how you get audio to external speakers?

I’m using an AVR with external speakers and Google AI is adamant that AVRs, as well as HDMI splitters etc., will at minimum introduce microstutter or microjudder to custom framerates, if not convert them to 50fps or 60fps or create a black screen.

Meanwhile, end users are saying that it will work; however, they are probably going by eye test and I doubt they have ever tested it for microstutter or microjudder. Google AI also recommended “extend this display” at one point, but then at another said that it would also create these issues and possibly a lot of audio sync variance as well.

I want to interpolate mkv movies to 47.954fps and 48fps and output them at either 1x or 2x Hz without getting video artifacts or bad audio sync variance, and am wondering the best way to do it. I have an audio sync measurement device so I can correct for bad sync latency, but not for bad drift or variance. I need a steady number, even if that number is big.

The video signal is a constant stream of data at a fixed rate, so there can't be stutter in the video signal itself, and sending audio to an AVR or splitting the signal wouldn't change that. Any visible stutter or judder would be on the software side. The biggest challenge is perfectly synchronizing the frame rate of the video player to the refresh rate of the display. If the video player and display both support variable refresh rate (FreeSync/G-SYNC), that would be the best solution.

Thanks for your help. My display is an LG C5 which has a framerate matching setting I believe, "true cinema." Anyway in PC mode it's supposed to not alter the incoming framerate anyway.

As far as Freesync/G-sync, want to avoid those because I think they introduce artifacts or microstutter/judder, at least they used to.

[beton247totchno]

FYI: successfully overclocked Iris XE (1195G7) eDP display from 60 Hz to 120 Hz (up to 130 Hz without artifacts, 333 MHz pixel clock). Latest drivers (Intel Arc panel 2024), W10. The trick is to not create any new entries, only modify the single entry. Refresh rate change is not available I suppose.

[Framerate]

Is there any good place online to ask custom framerate questions that is more active?

I am trying to watch 47.952fps films on my LG C5 TV using a PC with madvr and mpc-HC player. I made a 95.904hz custom framerate with CRU. In Windows 11 display settings > advanced display settings, you can select it and it shows as "95.90," but in madvr data screen it says it is actually giving me 95.041xx. Are the windows display settings lying or is MadVR not connecting the right framerate?

A 47.952fps custom framerate I made gave me 48.02xx in madVR also. It's almost like it is rounding close to even numbers, but can't say why 47.952fps would round up to 48, while 95.90 would round down to 95.

Any way to get much closer to the exact value in madvr settings? This did not happen nearly to this extent when I tested last year with Windows 10 and a Sony Bravia 7. Now I have Windows 11 and an LC C5. Native (non-CRU) 23.976fps rounding error is also worse with this setup.

[ToastyX]

(07-30-2026 07:39 AM)Framerate Wrote:  Is there any good place online to ask custom framerate questions that is more active?

I am trying to watch 47.952fps films on my LG C5 TV using a PC with madvr and mpc-HC player. I made a 95.904hz custom framerate with CRU. In Windows 11 display settings > advanced display settings, you can select it and it shows as "95.90," but in madvr data screen it says it is actually giving me 95.041xx. Are the windows display settings lying or is MadVR not connecting the right framerate?

A 47.952fps custom framerate I made gave me 48.02xx in madVR also. It's almost like it is rounding close to even numbers, but can't say why 47.952fps would round up to 48, while 95.90 would round down to 95.

Any way to get much closer to the exact value in madvr settings? This did not happen nearly to this extent when I tested last year with Windows 10 and a Sony Bravia 7. Now I have Windows 11 and an LC C5. Native (non-CRU) 23.976fps rounding error is also worse with this setup.

AVS Forum might be better for that: https://www.avsforum.com/forums/

Hardware clocks can vary slightly (GPU's pixel clock vs. audio clock vs. system clock) which can affect synchronization, which is why I recommended variable refresh rate (FreeSync/G-SYNC). Otherwise you would have to tweak the timing parameters and refresh rate to get closer but never quite exact, and the process is a huge pain: https://forum.doom9.org/showthread.php?t=173571

The only other option would be to use a player that can adjust the audio to synchronize with the video instead of the other way around. That way the GPU's pixel clock becomes the source of truth: https://wiki.jriver.com/index.php/VideoClock

[joehannes]

Hey dude, thanks for the tool. I've had an annoying coil whine with my RTX 5070 and two new monitors, which is now almost gone.
Do you have a buymeacoffee.com account? I'd like to give you a tip.

[DeanOfMean]

Hi ToastyX,

I'm using CRU to get 3840x2160 @ 119.880 Hz, RGB 8 bpc Full over DisplayPort on an NVIDIA GTX 1080 Ti.

The monitor's original 120 Hz timing uses:
4000 x 2240 total, with a vertical back porch of 62, but it only runs as YCbCr 4:2:2 rather than RGB.

At 119.880 Hz:
- vertical BP 60 = RGB Full works
- vertical BP 61 = NVIDIA falls back to YCbCr 4:2:2

BP 60 appears stable so far, although I haven't tested it for very long.

Would you recommend leaving the blanking at the highest value that works, or reducing it a little further for margin, for example BP 59 or 58? Is there any practical stability benefit to leaving some margin below the RGB bandwidth limit if BP 60 seems to be working normally?

Also, if the timing is too close to that upper limit, would you expect actual instability/artifacts during use, or would the driver typically just reject RGB/fall back to YCbCr when the mode is set?

Finally, is there any realistic risk of damaging the monitor by experimenting with reduced blanking like this, considering the refresh rate is still well below the monitor's rated 144 Hz? Or would an unsupported timing normally just result in no signal, artifacts, or some other temporary display issue?

Thanks!

[ToastyX]

(Today 12:15 AM)DeanOfMean Wrote:  Would you recommend leaving the blanking at the highest value that works, or reducing it a little further for margin, for example BP 59 or 58? Is there any practical stability benefit to leaving some margin below the RGB bandwidth limit if BP 60 seems to be working normally?

Also, if the timing is too close to that upper limit, would you expect actual instability/artifacts during use, or would the driver typically just reject RGB/fall back to YCbCr when the mode is set?

Finally, is there any realistic risk of damaging the monitor by experimenting with reduced blanking like this, considering the refresh rate is still well below the monitor's rated 144 Hz? Or would an unsupported timing normally just result in no signal, artifacts, or some other temporary display issue?

The bandwidth limit would be based on the pixel clock, so the limit is exact and doesn't need a margin. The signal is streamed at a fixed rate, so there shouldn't be any instability. Reduced blanking normally doesn't cause problems unless it's lower than the monitor can handle, in which case you would see one of the temporary display issues you described.