Mac mini: 4k @ 120hz?

[Lazylex]

Any thoughts on using something like this to emulate a Displayport device, forcing the Mac to output a full signal then adding on a DP to HDMI 2.1 Active Adapter on the output?

Headless DisplayPort 1.4 32G 8K EDID Emulator

Ideal zur Lösung von Problemen die bei Verwendung von DP-KVM-Umschaltern oder Video-Switchern auftreten können

www.ute.de

 

[frisbfreek]

Any thoughts on using something like this to emulate a Displayport device, forcing the Mac to output a full signal then adding on a DP to HDMI 2.1 Active Adapter on the output?

Headless DisplayPort 1.4 32G 8K EDID Emulator

Ideal zur Lösung von Problemen die bei Verwendung von DP-KVM-Umschaltern oder Video-Switchern auftreten können

www.ute.de

My google translation is rusty but from what I can tell, the only function this thing does is make the underlying display look like it's always on even if you turn it off, so that monitor settings aren't lost. If that's true, I don't think this will solve anything. Also, it doesn't support 4K120 either?

 

[Monstieur]

Any thoughts on using something like this to emulate a Displayport device, forcing the Mac to output a full signal then adding on a DP to HDMI 2.1 Active Adapter on the output?

Headless DisplayPort 1.4 32G 8K EDID Emulator

Ideal zur Lösung von Problemen die bei Verwendung von DP-KVM-Umschaltern oder Video-Switchern auftreten können

www.ute.de

macOS detects additional data from the descriptor identifying the display as HDMI and not the basic information you see in most EDID editors. DisplayPort to HDMI adapters passthrough the original EDID with some patches and perform signal conversion rather than emulating a full DispalyPort device. These display emulator devices also just clone the EDID. They have some presets with fixed EDIDs but those won't support 4K 120 Hz.

Maybe a programmable device similar to Dr HDMI 8K could do full emulation, but that effort is better spent developing a patch for macOS so you can use a simple DisplayPort to HDMI 2.1 adapter like Windows.

 

[Lazylex]

One additional item I noticed that I'm not sure if it's been mentioned here. When connecting the LG C1 to my MBP M1 Max, pressing the Green button on the remote 7 times shows the connection info as YCBCR444 8b TM. I can get this to RGB 8b by modifying the com.apple.windowserver.displays file's PixelEncoding and Range options, but more interesting is the switch from TM to 4L10.

If I use the Club3D HDMI 2.1 adapter (CAC-1586), it will show 8b 4L10.

TM = 18 GB/s (HDMI 2.0 spec)
4L10 = 4 lanes of 10GB/s

So it does recognize the bandwidth variation.

4k @ 120hz, 8bit, 4:4:4 = 32.08 Gbps uncompressed, so it should fit just fine.
4k @ 120hz, 10bit, 4:4:4 = 40.09, which is pushing it and most likely wont happen unless we can rely on DSC.

Still not sure why Apple does not open the ability for users to modify Chroma, Color Depth, etc, especially if they are trying to cater to creatives.

 

[Artagra]

My report to Apple - please let me know anything I've missed or got wrong!

MacOS incorrectly limiting pixel clock to 600mhz on Display Port 1.4 to HDMI 2.1 adapters, preventing 4k120

Description of the problem: Unable to display 4k120 on an external HDMI 2.1 display via a Displayport 1.4 to HDMI 2.1 adapter. It appears that MacOS is incorrectly limiting the pixel clock to 600mhz on such adapters, preventing 4k120.

Adapters such as the Realtek RTD2173 and the Synaptics VMM6100 allow the conversion of a Display Port 1.4 signal to HDMI 2.1, allowing the support of external 4k120 displays such as the LG OLED CX and others. Mac hardware and MacOS supports output of a Displayport 1.4 signal, however when one of these adapters are plugged, MacOS appears to detect the 600mhz Display Clock (which is still communicated across the adapter, even though it is no longer applicable. The same adapter works on a 2019 Macbook Pro 16” (Intel) in Windows via Bootcamp, but does not work in MacOS.

Expected result: 4k120 output

Actual result: 4k60 output

How to reproduce: Connect an LG CX (or other HDMI 2.1 display that supports 4k120) into a Mac that supports Display Port 1.4 over USB-C via a DP1.4 to HDMI2.1 adapter - output will be limited to 4k60, and will not be 4k120.


Additional info:

- Nvidia had a very similar issue in one series of Windows drivers - https://www.nvidia.com/en-us/geforc...e-46192-game-ready-driver-feedback-thread-re/ “Realtek Displayport-to-HDMI 2.1 protocol converter clock limited to 600MHz pixel clock [3202060]”

- Multiple brands sell adapters using these chips, including Club3d and Cable Matters - https://www.cablematters.com/pc-1368-122-usb-c-to-8k-hdmi-adapter.aspx and https://www.amazon.ca/Club-3D-DisplayPort1-4-4K120Hz-CAC-1085/dp/B08BX49V5V

- Macrumors thread related to the issue: https://forums.macrumors.com/threads/mac-mini-4k-120hz.2267035/page-6

 

[joevt]

One additional item I noticed that I'm not sure if it's been mentioned here. When connecting the LG C1 to my MBP M1 Max, pressing the Green button on the remote 7 times shows the connection info as YCBCR444 8b TM. I can get this to RGB 8b by modifying the com.apple.windowserver.displays file's PixelEncoding and Range options, but more interesting is the switch from TM to 4L10.

If I use the Club3D HDMI 2.1 adapter (CAC-1586), it will show 8b 4L10.

TM = 18 GB/s (HDMI 2.0 spec)
4L10 = 4 lanes of 10GB/s

So it does recognize the bandwidth variation.

4k @ 120hz, 8bit, 4:4:4 = 32.08 Gbps uncompressed, so it should fit just fine.
4k @ 120hz, 10bit, 4:4:4 = 40.09, which is pushing it and most likely wont happen unless we can rely on DSC.

Still not sure why Apple does not open the ability for users to modify Chroma, Color Depth, etc, especially if they are trying to cater to creatives.

I guess TM is short for TMDS.
4L10 = 4 lanes 10 Gbps = 40 Gbps total using 16b/18b timing -> 35.56 Gbps of actual data.
4K 120Hz is 1188MHz (HDMI timing).
1188MHz * 24bpp = 28.5 Gbps.
1188MHz * 30bpp = 35.6 Gbps. Requires DSC - anything less than 30 bpp would be sufficient.
DisplayPort 1.4 limit is 25.92 Gbps so you need DSC or 4:2:2 10bpc or a different timing (CVT-RB2 timing is 1076MHz).

My report to Apple - please let me know anything I've missed or got wrong!

You should include some of the info from @Lazylex, specifically that the display can report the link rate and lanes even if Apple's crappy software cannot.
Would be interesting to get photos of that when it's working in Windows for each adapter.

 

[frisbfreek]

I guess TM is short for TMDS.
4L10 = 4 lanes 10 Gbps = 40 Gbps total using 16b/18b timing -> 35.56 Gbps of actual data.
4K 120Hz is 1188MHz (HDMI timing).
1188MHz * 24bpp = 28.5 Gbps.
1188MHz * 30bpp = 35.6 Gbps. Requires DSC - anything less than 30 bpp would be sufficient.
DisplayPort 1.4 limit is 25.92 Gbps so you need DSC or 4:2:2 10bpc or a different timing (CVT-RB2 timing is 1076MHz).


You should include some of the info from @Lazylex, specifically that the display can report the link rate and lanes even if Apple's crappy software cannot.
Would be interesting to get photos of that when it's working in Windows for each adapter.

My impression was CVT-RB2 timings were pretty common these days, and that’s how 4K120*24bpp works in Windows without DSC. Is that incorrect? Is CVT-RB2 not implemented on Macs?

I wonder if 30 bpp works with the adapter in Windows because the adapter is converting a DisplayPort signal with DSC to HDMI 4L10 without DSC. I hear the adapters are able to do both DSC conversion and passthrough.

So if 4K120*24bpp doesn’t work, it could be a CVT-RB2 or frequency issue. It could also be a DSC usage issue for 30bpp. Aside from Apple’s 6K XDR display, do we know if DSC is able to kick in automatically when needed?

 

[joevt]

My impression was CVT-RB2 timings were pretty common these days, and that’s how 4K120*24bpp works in Windows without DSC. Is that incorrect? Is CVT-RB2 not implemented on Macs?

I wonder if 30 bpp works with the adapter in Windows because the adapter is converting a DisplayPort signal with DSC to HDMI 4L10 without DSC. I hear the adapters are able to do both DSC conversion and passthrough.

So if 4K120*24bpp doesn’t work, it could be a CVT-RB2 or frequency issue. It could also be a DSC usage issue for 30bpp. Aside from Apple’s 6K XDR display, do we know if DSC is able to kick in automatically when needed?

CVT-RB2 is just a method to calculate how much time/pixels is spent during the horizontal and vertical blanking period. It can be arbitrary but probably not zero - there needs to room/time for audio and for other info and for the circuitry of the display between lines and frames. A display can include and timing it wants in its EDIDs. It can use an arbitrary timing and it can use predefined timings. The CTA spec has many timings defined for HDMI. Install edid-decode. It can dump the contents of an EDID and it can also list those predefined timings and it can calculate timings.
edid-decode -h
edid-decode --list-established-timings --list-dmts --list-vics --list-hdmi-vics
edid-decode --vic 118 --vic 120
edid-decode --cvt w=3840,h=2160,fps=120,rb=2

I've seen DSC work in Catalina with Apple's USB-C to HDMI adapter and CalDigit SOHO dock (both of those are USB-C hubs with two lanes of DisplayPort 1.4 so they require DSC to get 4K60 10bpc). They convert the DSC input to non-DSC output. I think Big Sur disabled DSC in some cases. I'm not sure about Monterey.

 

[Artagra]

Thanks for the comments I would expect that 4k144hz monitors that are working via DP1.4 are using DSC, eg LG UltraGear 27GP950-B and Gigabyte M28U.

 

[YoYoMa]

Quick one for you guys - has anyone at least identified how to get 4k HDR 444 60hz to the CX OLED? I know it can be done without HDR, but anyone at least get it done with HDR?

 

[desert0616]

Quick one for you guys - has anyone at least identified how to get 4k HDR 444 60hz to the CX OLED? I know it can be done without HDR, but anyone at least get it done with HDR?

As far as I know, the answer is no.

You cannot achieve anything beyond HDMI 2.0 under MacOS if your monitor/TV is HDMI, even if you are using DP adapters.

 

[Artagra]

So, talking to some Mac System Admins, they recommend we all file similar feedback and reference to each other, as a way of getting it on Apple's radar.

So anyone who would actually use this and cares about this issue, consider submitting a "feedback" via Feedback Assistant or the web portal, and reference my feedback - FB9734589. Then, post your feedback number here.

In order to do so, you will probably need to be enrolled in the Beta program (but don't need to be running the Beta OS). I have confirmed this problem applies on the current Monterey Beta.

Here's a link with how to file feedback:

https://www.imore.com/how-submit-feedback-betas

I feel like this could be such a tiny thing for Apple to fix if it is just the clock issue, and would really remove any complaints around the HDMI port only being 2.0

 

[Toss4]

Done! Really great summary of the issue! Would probably be best to post the exact same thing on Reddit and other forums as well to get even more people to send in feedback.

 

[Cristianku]

Have you seen this ?

perl script parses ioreg output, can output JSON, dumps M1 Mac display timings.

perl script parses ioreg output, can output JSON, dumps M1 Mac display timings. - ioreg.pl

gist.github.com

 

[kasakka]

So, talking to some Mac System Admins, they recommend we all file similar feedback and reference to each other, as a way of getting it on Apple's radar.

So anyone who would actually use this and cares about this issue, consider submitting a "feedback" via Feedback Assistant or the web portal, and reference my feedback - FB9734589. Then, post your feedback number here.

In order to do so, you will probably need to be enrolled in the Beta program (but don't need to be running the Beta OS). I have confirmed this problem applies on the current Monterey Beta.

Here's a link with how to file feedback:

https://www.imore.com/how-submit-feedback-betas

I feel like this could be such a tiny thing for Apple to fix if it is just the clock issue, and would really remove any complaints around the HDMI port only being 2.0

Filed FB9791601

 

[Artagra]

Filed FB9791601

Sweet, referenced from mine.

 

[SerpentEagle]

So, talking to some Mac System Admins, they recommend we all file similar feedback and reference to each other, as a way of getting it on Apple's radar.

So anyone who would actually use this and cares about this issue, consider submitting a "feedback" via Feedback Assistant or the web portal, and reference my feedback - FB9734589. Then, post your feedback number here.

In order to do so, you will probably need to be enrolled in the Beta program (but don't need to be running the Beta OS). I have confirmed this problem applies on the current Monterey Beta.

Here's a link with how to file feedback:

https://www.imore.com/how-submit-feedback-betas

I feel like this could be such a tiny thing for Apple to fix if it is just the clock issue, and would really remove any complaints around the HDMI port only being 2.0

Feedback submitted referencing your submission. FB9795062

The more submissions they get the better our chances of getting this fixed. Please help out everyone!

 

[SerpentEagle]

I have also created a post on Reddit calling on others to file bug reports as well.

https://www.reddit.com/r/mac/comments/r8inmn

 

[Artagra]

Thanks that's great! I've also posted in the Mac Sysadmins slack, as it's applicable to some of them

 

[mfractal]

came here with same issue. Glad it's being discussed and hope resolution will follow.

I've just created a bug report and referenced the original one as instructed, did my part i guess. Now we wait

 

[repoman27]

I'm all for filing feedback with Apple regarding this issue, but as far as I can tell, the underlying issue here is that macOS is just completely oblivious to HDMI 2.1 Fixed Rate Link (FRL). There is simply no support at the OS or driver level for FRL either natively or via DisplayPort to HDMI protocol converters. When macOS sees an EDID with FRL options, it just ignores them. Since HDMI 2.1 devices are still required to support TMDS for backwards compatibility, macOS will simply find those timings in the EDID and handle them accordingly. TMDS links are limited to a 6 Gbit/s signaling rate or 600 Mchar/s per the HDMI specification. That means 4K UHD at up to 60 Hz, either RGB / YCbCr 4:4:4 8 bpc (SDR), or YCbCr 4:2:2 10 bpc (HDR) is the most it can do.

 

[SerpentEagle]

I'm all for filing feedback with Apple regarding this issue, but as far as I can tell, the underlying issue here is that macOS is just completely oblivious to HDMI 2.1 Fixed Rate Link (FRL). There is simply no support at the OS or driver level for FRL either natively or via DisplayPort to HDMI protocol converters. When macOS sees an EDID with FRL options, it just ignores them. Since HDMI 2.1 devices are still required to support TMDS for backwards compatibility, macOS will simply find those timings in the EDID and handle them accordingly. TMDS links are limited to a 6 Gbit/s signaling rate or 600 Mchar/s per the HDMI specification. That means 4K UHD at up to 60 Hz, either RGB / YCbCr 4:4:4 8 bpc (SDR), or YCbCr 4:2:2 10 bpc (HDR) is the most it can do.

Whatever the case, if it is a software limitation, then it should be fixed. It has been nearly FOUR years since HDMI 2.1 was released. We cannot hold Apple to such mediocre standards.

 

[Arvine]

I can also join the effort of filing feedback to Apple. I have both the Club3D DP 1.4 to HDMI 2.1 adapter as well as using HDMI directly between AMD 6800 xt and a Samsung Q800 8K TV. Both support HDMI 2.1. My experience is that the limit of only 4K@60 applies to both DP and HDMI using MacOS, not specifically to DisplayPort or any adapter.

 

[PieroPontra]

I can also join the effort of filing feedback to Apple. I have both the Club3D DP 1.4 to HDMI 2.1 adapter as well as using HDMI directly between AMD 6800 xt and a Samsung Q800 8K TV. Both support HDMI 2.1. My experience is that the limit of only 4K@60 applies to both DP and HDMI using MacOS, not specifically to DisplayPort or any adapter.

Hello,

I'm actually considering to buy a 55" 8K QN700A (Only available in UE I think, basically a smaller QN800A) and I'm using a 6800XT too (in an hackintosh). I use both Win 20% and macOS 80% of the time. In Win I should get 8K 60 4:4:4 easily, will I get 4k 60hz 4:4:4 on macOS? If that's the case, tv's AI Upscaler will do a good job presenting a 4K image on an 8K panel? How do you judge text sharpness in macOS on an 8K panel? Is it doable waiting for Apple to overcome present limitations? Thanks for all the advices you'll end up giving to me.

 

[repoman27]

Whatever the case, if it is a software limitation, then it should be fixed. It has been nearly FOUR years since HDMI 2.1 was released. We cannot hold Apple to such mediocre standards.

Once again, I'm all for asking Apple to introduce support for HDMI 2.1 FRL in macOS. Just bear in mind that what you're asking is for Apple to dedicate not insignificant development resources to writing and testing a driver to support third-party hardware.

The HDMI 2.1 specification is just a document. Hardware is developed based on that spec, and then drivers can be written to support that hardware. Apple's GPUs do not support HDMI natively. They do sell AMD Navi 21 based GPUs, but the native HDMI functionality is not exposed on Apple's MPX modules. Apple is also clearly on the outs with NVIDIA, so it's unlikely they'll be releasing updated drivers for GA102 based GPUs.

That leaves DisplayPort to HDMI 2.1 protocol adapters. The dongles currently on the market use chips from either Synaptics or Realtek. While Apple does have working relationships with those companies, a functional driver generally requires input from both parties, and neither is Apple's preferred vendor for display protocol converters. Apple uses MegaChips (now part of Kinetic Technologies) for PCONs in their own hardware these days, but also have a lengthy history of working with Parade as well. Parade only launched their PS195 and PS196 DisplayPort 2.0 to HDMI 2.1 protocol converters in August, so I could easily see Apple including HDMI 2.1 FRL support in a macOS release around the same time as adapters based on those chips start showing up in the market.