MacOS respects true color displays preventing dither + save monitor

Intet

I recently found a workflow that lets you be reasonably certain in acquiring a true color screen:
https://ledstrain.org/d/3841-how-to-search-for-a-true-6-bit-8-bit-10-bit-monitor

Based on that principle I bought the AOC Q25G4SR
Which uses the panel SN245ES01-1

I am delighted to report I see not dither within MacOS in this true 8-bit color screen.
It seems all you have to do is make sure the screen advertises no FRC capability over it's connection.

I will do more testing under different circumstances and using it longer (I have it for 5 hours now), but so far it feels very comfortable and on all the testing pictures I usually use to see temporal dither I see none.
I am testing with up to 920 fps recordings through a microscope.

This monitor uses a high quality oxide panel instead of A-Si which prevents current leakage flicker.
So far I see only extremely minimal pixel inversion flickers (among the best I've seen) and of course true DC dimming (no PWM flicker).
Very fine grain matte finish
RGB Vertical Stripe

Relevant linkt to the specific monitor I am using:
https://aoc.com/ch-de/gaming/products/monitors/q25g4sr
https://www.displayspecifications.com/de/model/a6cc4304
https://www.panelook.com/SN245ES01-1_HKC_24.5_CELL_parameter_65938.html

I will update when I have it for longer, but it looks very promising so far.

WhisperingWind

Intet

The presence of temporal dithering seems to depend heavily on the hardware. On a true 8‑bit monitor, my 2018 Mac mini shows no temporal dithering by default; however, when outputting from an M1 MacBook Pro (without using StillColor) over HDMI to the same monitor, temporal dithering does appear.

This thread thread has a lot of information about temporal dithering on Apple Silicon Macs.

Donux

Intet They ship to reviewers best units. And in the marketing content, they do not promise true 8 bit, or flicker free. They will pick whatever panel they like to fulfill the specs, https://aoc.com/us/gaming/products/monitors/q25g4sr

As for HP Omen 27q, they are at least mentioning, its true 8 bit plus FRC 2 bit. Also at least they are mentioning its flicker free. So when their sourcing department has a goal to find new suppliers because current one can not supply right panels at the right time, they will at least inspect the requirements from new suppliers. Or at least, that is how it works in real world I assume, nobody is running around with spectometers or microscopes, correct me if I am wrong.

AshX

WhisperingWind The presence of temporal dithering seems to depend heavily on the hardware. On a true 8‑bit monitor, my 2018 Mac mini shows no temporal dithering by default; however, when outputting from an M1 MacBook Pro (without using StillColor) over HDMI to the same monitor, temporal dithering does appear.

I’ve discussed with aiaf the difference between Intel Macs and Apple Silicon as it relates to resolution and bit-depth over different connection times. Many different Intel Macs had reduced resolutions and color bit depth when driven at specific resolutions.

The 2018 13” MacBook Pro shows this on the tech specs page:

Simultaneously supports full native resolution on the built-in display at millions of colors and:

One display with 5120-by-2880 resolution at 60Hz at over a billion colors

Up to two displays with 4096-by-2304 resolution at 60Hz at millions of colors

Up to two displays with 3840-by-2160 resolution at 60Hz at over a billion colors

Interestingly, the 2019 21.5” 4K iMac I used for years is cited as being capable of this on the internal screen itself:

Simultaneously supports full native resolution on the built-in display at millions of colors (21.5-inch) or 1 billion colors (21.5-inch 4K)

SwitchResX actually allowed me to toggle between these two settings on the internal iMac screen, which made it tolerable for a while. I wasn’t sure what it was toggling, but now I’m convinced it was disabling the panel’s FRC, not the AMD GPU’s dithering. So it was shifting between 8-bit and 8-bit+FRC like an external display.

Where things get interesting is this post we have an M1 MacBook Air that is usable on an external display and an M1 Mac Mini that is not, even with Stillcolor and BetterDisplay. Same OS.

The only difference is the MBA uses the old Retina bezel screen and the tech specs cite millions of colors on the internal panel. The mini makes no mention of any of this. It also doesn’t have a TCON.

I’ve looked at Apple patents and I think the simplest answer is the newer Macs lack the ability to output anything less than 10-bit. I think the OS will try to override into 10-bit and only through brute force limiting the bandwidth using HDMI or a weaker connection is it possible to avoid this. And maybe for some people it is tolerable enough that having what remains of dithering isn’t a deal breaker.

The only way to prove this is to do a comparison with the M1 MBA and the M1/M2 Touchbar MBPs and the older Intel Macs. My guess is the internal screen being limited to millions of colors is somehow allowing for an output option within the display pipeline to push 8-bit, with the right monitor and connection.

This is all conjecture but this topic has had me thinking about this again.

Intet

My mac is a M3 Mac Studio also connected over hdmi.

Still color/ better display did not work with my old monitor which advertises 8 + 2 bit frc. I guess it could still be on? I will check when I am home again.

WhisperingWind

Intet

Still color/ better display did not work with my old monitor which advertises 8 + 2 bit frc.

In this case, you need to switch to 8-bit output; otherwise you’ll get a 10-bit signal from the Mac, which will trigger the monitor’s FRC. That will result in temporal dithering. StillColor won’t work here, since it doesn’t let you select the output signal’s bit depth. Only BetterDisplay can do this.

Using BetterDisplay, you need to switch to 8‑bit and then disable dithering (two settings). At least on my M1 MBP, this works.

asus389

Intet What version of Mac OS and what cable are you using specifically?

lukeb

Intet My mac is a M3 Mac Studio also connected over hdmi.

@AshX interesting that the mac here is an M3 studio. I'm sure you mentioned previously that you found it the most tolerable in store. It is the only model I haven't tested myself. Might be something different with it.

Intet

AshX Yes, like I said still color/better display were not able to disable dither on my old monitor which was 8 + 2 bit frc.

But if you get a true 8-bit monitor that does report no frc capability, it will not force gpu dither.

What you are talking about different modes supporting different colors is really just cable bandwith limitations. More colors need more bandwith, so on high resolution and framerate it is sometimes only possible to output the image with resuced color depth to the screen. You can see the same limitations when you look i to graphics card documentation.

AshX

Intet What you are talking about different modes supporting different colors is really just cable bandwith limitations. More colors need more bandwith, so on high resolution and framerate it is sometimes only possible to output the image with resuced color depth to the screen. You can see the same limitations when you look i to graphics card documentation.

Interesting. So a few follow-up questions:

Are you saying you found that Apple Silicon Macs will bypass the commands from Stillcolor/BetterDisplay and push 10-bit if an external display with 8-bit+FRC capabilities is detected?

Is Apple Silicon’s GPU applying dither or is it just that the bit depth output is reading as 10-bit so the external display is applying its own FRC?
Does the type of connection (Thunderbolt, USB-C, HDMI, Display Port, VGA) have an impact in your experience when it comes to Apple Silicon Macs, or only Intel/AMD Macs?
What are your thoughts on anecdotal reports that the M4 chips, especially the MacBooks, are passing along additional processing to external displays that previous generations (M1-M3) did not?
Curious if you found any differences between versions of MacOS.

I’m trying to drill down whether MacOS and possibly the TCON (relevant on MacBooks, not the Mac mini or Mac Studio) is re-enabling GPU dithering despite Stillcolor/BetterDisplay telling it not to.

If I’m understanding what you’re saying correctly, you’ve found that the issue is more likely 10-bit being mistakenly pushed to monitors that have a capability to use FRC to process it? So in that case GPU dithering would not be occurring because it is disabled, but the external display’s own FRC is activating? So if you’re running a true 8-bit external display then MacOS’s only recourse is to use GPU dithering, which is effectively disabled. Hence a “fake 8-bit” output even though the OS is technically coded for 10-bit.

Or is it possible to actually deliver a true 8-bit output, whether through BetterDisplay or reducing the bandwidth, as was possible on certain Intel Macs at specific configurations? But this to me sounds more like we are just bottlenecking the output to an artificial 8-bit.

I know this is to some degree semantics, but I think the distinction is relevant. I believe MacOS shifted over to 10-bit output prior to Apple Silicon, but it’s interesting that since MacOS Big Sur the possibility of outputting in true 8-bit seems to have been removed entirely (I’m not speaking about BetterDisplay).

If I’m making any sense then this is sort of another way to explain why no matter what, MacBook internal panels are going to use FRC. Every Apple Silicon Mac internal display is 8-bit+FRC.

What confuses me is that on the M1 MBA and M1/M2 13” MBP BetterDisplay says the internal screen is 8-bit, backed up by Apple’s “millions of colors” spec sheet description.

My 2019 iMac’s spec sheet specifically references “millions or billions of colors” and SwitchResX confirmed that allowing me to actually toggle between the two (the screen would go black and then turn back on when you switched modes - this is what makes me think that the post-Retina panel technology implemented with Liquid Retina is locked to 8-bit+FRC on the hardware level - OR MacOS Big Sur has locked it). If these laptops are capable of 10-bit like the later M2 MBA, why not just say “billions”?

So are those devices actually using FRC? I’ve not seen conclusive proof they are. We know they’re utilizing GPU dithering, but that can be disabled.

WhisperingWind

Intet

Now my finding is that in a M3 Mac Studio it is not possible to disable temporal dither/frc if the screen reports capability of it.

EDID (the “handshake” info a monitor sends to your computer) doesn’t have a field that says “this monitor uses FRC.” FRC is an internal panel trick, so it isn’t reported. Because of that, a 6‑bit+FRC panel tells the computer it’s 8‑bit, and an 8‑bit+FRC panel tells the computer it’s 10‑bit. To your Mac/PC, a true 10‑bit monitor and an 8‑bit+FRC monitor look the same. Software can’t tell them apart from EDID.

If you select 8‑bit in BetterDisplay, does the M3 still send a 10‑bit signal to an 8‑bit+FRC monitor?

If you connect a true 8‑bit monitor, does the M3 send an 8‑bit signal?

In both cases, is there GPU‑side temporal dithering (assuming it isn’t disabled in StillColor/BetterDisplay)? I’m not referring to FRC, since that’s handled internally by the monitor and isn’t visible to the Mac.

asus389

AshX

AshX Does the type of connection (Thunderbolt, USB-C, HDMI, Display Port, VGA) have an impact in your experience when it comes to Apple Silicon Macs, or only Intel/AMD Macs?

I’m curious about this also. It’s possible to have a cable/resolution combo that won’t support 10 bit color @ 60fps. The monitor the OP is using advertises 300hz refresh. I wonder if it’s just giving up on 10 bit because 10 bit @ 300 hz is not possible using the bandwidth available in the current generation of cables?

Intet

This is the first time I hear of distinctions in terms of dither between the different M1 type products. To my knowledge from other posters, all M1 type products were using temporal dithering (M1/M2/M3/M4), but for a while you were able to use Stillcolor or BetterDisplay to disable it. But since then compatibility was broken with subsequent MacOS updates, while it still works on Intel Macs. This info might be wrong, I am just telling you what my information was to this date.

Now my finding is that in a M3 Mac Studio it is not possible to disable temporal dither/frc if the screen reports capability of it. My old monitor has 8 + 2 bit capability, where two more bits areachieved via temporal dither/frc. The monitor can run in native color mode (8-bit) and I had it successfully doing so on Windows and Linux. But whatever I did on my Mac Studio it would always force a 10-bit (max possible) output to the display. But with my new Monitor that only supports native 8-bit it outputs 8-bit on my Mac Studio to this display. The MacStudio does not inject its own temporal dither into the display out via its gpu, it just automaticially asks for temporal dither if the screen it is attached to reports it has that ability.

It is possible this changed after the M3. But I assume from your writings that more likely the screens in the M-line MacBooks just all support temporal dither.

There is no "pseudo 8-bit" to my knowledge if 8-bit was not supported at all, the image could not be displayed. 8-bit and 10-bit color outputs are different formats which are not backwards compatible. Relevant here but besides the main point: Theoreticially some screens upscale all 8-bit signals to 10-bit, often via FRC.

jordan

Intet what if you change the edid to not advertise 8+2? I would also be curious if the anker 563 display link hub prevents forcing 10bit on external monitor because someone on reddit said its the only way to get their mac to output to their eink monitor without flickering which I assume is dithering. That hub doesnt need stillcolor either.

AshX

Intet products were using temporal dithering (M1/M2/M3/M4), but for a while you were able to use Stillcolor or BetterDisplay to disable it.

They all apply GPU dithering - every chip generation. But Stillcolor is confirmed via terminal to disable it. M3 was the latest released when Stillcolor was released, so most of the capture card confirmation is from M1-M3 chips. To my knowledge no one has tested an M4 or M5 Mac via the capture card method outlined in the Stillcolor thread.

I should have been more clear that there is an assumption Stillcolor isn’t working on M4 and M5 because it doesn’t seem to alleviate symptoms. I think it’s likely Stillcolor still disables dithering, but there is a difference in the internal panels themselves or additional processing is being applied. Again, no one has tested this - we are just going by symptoms.

I theorized that the gray color flicker was behind the symptoms because it is worse on M4 MacBook Airs, which because they lack the same aggressive PWM as the Pros, were tested by more users here.

Intet But since then compatibility was broken with subsequent MacOS updates, while it still works on Intel Macs. This info might be wrong, I am just telling you what my information was to this date.

Yeah, we don’t know if Sequoia and Tahoe stop Stillcolor from working. No one has tested this. Stillcolor was originally tested on Ventura and Sonoma

Intet The monitor can run in native color mode (8-bit) and I had it successfully doing so on Windows and Linux. But whatever I did on my Mac Studio it would always force a 10-bit (max possible) output to the display.

So even with BetterDisplay disabling GPU dithering and selecting 8-bit output, FRC was still enabled by your monitor on Mac only?

Intet But with my new Monitor that only supports native 8-bit it outputs 8-bit on my Mac Studio to this display. The MacStudio does not inject its own temporal dither into the display out via its gpu, it just automaticially asks for temporal dither if the screen it is attached to reports it has that ability.

I’m not completely following you. By temporal dither you’re referring to monitor FRC, correct? I’m assuming you’re still disabling GPU dithering and sending an 8-bit signal via BetterDisplay?

Intet But I assume from your writings that more likely the screens in the M-line MacBooks just all support temporal dither.

Every screen is 8-bit+FRC on the MacBooks and Studio Display. I’m not sure what the M1 MBA and M1/M2 13” MBP are using because they have the older Retina screens and are marketed as being capable of “millions of colors” internally but also P3 wide color.

This is where I’m trying to draw a conclusion based on older Intel Macs with the same specs as to whether they’re achieving P3 wide color using solely GPU dithering, or whether the internal panel is also using FRC

Intet here is no "pseudo 8-bit" to my knowledge if 8-bit was not supported at all, the image could not be displayed. 8-bit and 10-bit color outputs are different formats which are not backwards compatible. Relevant here but besides the main point: Theoreticially some screens upscale all 8-bit signals to 10-bit, often via FRC.

This is what I’m basically trying to understand: does BetterDisplay actually deliver an 8-bit signal when selected, or is MacOS just always output at 10-bit and if the monitor is only 8-bit capable, then it’ll just get bottlenecked down to 8-bit as long as Apple Silicon GPU dithering is disabled?

I’m assuming BetterDisplay only gives you an 8-bit output option for your true 8-bit external display? But if it does offer both 10 and 8 bit output options within BetterDisplay, you could test to see if leaving it on 10-bit makes a difference.

Sorry for all the questions, but I actually think the answers are relevant in regard to the proper protocol we should direct users to follow to get a clean signal from a Mac.

It also kind of would confirm (or at least heavily imply) that this is why the user who posted here a few months ago that they had a clean signal from an M1 MacBook Air but when using an M1 Mac Mini on the same version of MacOS with the same cable and same BetterDisplay output options on the same monitor, they had symptoms.

The reason seems to be that the older Retina “millions of colors” laptops actually have an 8-bit output option whereas the Mac Mini only has 10-bit. It would be an example of a difference in the display pipeline likely carried over from the TCON component and firmware from the old Retina Intel era technology.

This in theory would not mean that you can’t achieve 8-bit on the new Macs, but rather that it will only work on a true 8-bit external display and possibly only on specific chips and OS versions.

asus389

The older 13 inch intel machines, at least the 2017 13 inch that I have, have the following behavior:

The panel is 8 bit and so long as the image assets are sRGB (millions of colors) there is little to no dithering and symptoms. If you load an image that uses wide color explicitly, such as a P3 encoded image, it will turn on dithering (presumably via the GPU) for that area of the screen which that image is displayed. This also applies to apps that use metal apis to do hardware acceleration via the GPU such as recent versions of browsers.

Starting with Big Sur/Apple Silicon (Big Sur was the first OS to support AS and shipped with the M1 machines) Apple started to make more use of Metal and P3 image assets and encouraging developers to do so, so dithering became more pervasive so people started having more issues overall, even on older machines, once they updated the Os.

AshX

asus389 The older 13 inch intel machines, at least the 2017 13 inch that I have, have the following behavior:

The panel is 8 bit and so long as the image assets are sRGB (millions of colors) there is little to no dithering and symptoms. If you load an image that uses wide color explicitly, such as a P3 encoded image, it will turn on dithering (presumably via the GPU) for that area of the screen which that image is displayed. This also applies to apps that use metal apis to do hardware acceleration via the GPU such as recent versions of browsers.

This is very interesting because it’s technically the same screen as the M1/M2 13-inch Touchbar Pros. The specs on Apple’s site are identical except the 2022 M2 has True Tone. The only other difference is the 2017 13-inch has different limitations on driving external displays depending on resolution. The 2022 13-inch has no such color specifications just that it drives up to 6K at 60Hz.

I think we can infer from this that the GPU is wholly responsible for driving external displays on these Retina laptops, and the Intel GPU is the bottleneck.

I’m also assuming that it would only be the GPU performing dithering on both the 2017 and 2022. This would explain why Liquid Retina on the MacBook Airs are still uncomfortable for people - the panel is likely 8-bit+FRC. But I don’t know if we can actually confirm this.

asus389 Starting with Big Sur/Apple Silicon (Big Sur was the first OS to support AS and shipped with the M1 machines) Apple started to make more use of Metal and P3 image assets and encouraging developers to do so, so dithering became more pervasive so people started having more issues overall, even on older machines, once they updated the Os.

Metal 2 was the last to contain large amounts of code from the Intel Mac era, which would be MacOS Monterey. Ventura introduced Metal 3 and from what I can tell, removed a lot of the frameworks from the Intel Macs. I wonder if this makes Ventura and onward more likely to apply processing that could engage dither…although that is allegedly only supposed to occur in gaming.

asus389 I’m curious about this also. It’s possible to have a cable/resolution combo that won’t support 10 bit color @ 60fps. The monitor the OP is using advertises 300hz refresh. I wonder if it’s just giving up on 10 bit because 10 bit @ 300 hz is not possible using the bandwidth available in the current generation of cables?

I didn’t know what monitor he was using. That makes sense.

This is sort of where I think we’re likely to be at: MacOS will try to output at 10-bit no matter what. And it seems like there is no way to force a true 8-bit output, but rather only ways to bottleneck the signal so external displays do not apply their own FRC processing.

The even bigger meta question I have is whether MacOS will continually try to re-enable dithering depending on the programs being used - or even the OS itself. The one point of evidence I think that backs this up is that BetterDisplay seems to be superior to Stillcolor because it is intermittently sending the command to disable dithering whereas Stillcolor only seems to do it when the application first disables it. So MacOS seems to still be trying to turn it on. This makes me wonder if Sequoia - and more likely Tahoe - are relying on this so much more compared to earlier versions and that’s why it seems like Stillcolor “doesn’t work.” Or maybe the chipset and configuration does matter like Intel/AMD, and M1/M2 were more likely to be bottlenecked, whereas M4 even at the base configuration is not. The only way to test is capture card or better yet microscope.

AshX

lukeb interesting that the mac here is an M3 studio. I'm sure you mentioned previously that you found it the most tolerable in store. It is the only model I haven't tested myself. Might be something different with it.

Yeah, I clocked that as well. The difference was the Mac Studio I tested was an M3 Ultra and connected to the Pro Display XDR, so higher specs. I asked on the Mac Studio subreddit if it was possible the M3 Ultra was influencing the quality of the external display, and people thought I was crazy. But I asked the Mac guy at Micro Center later on and he thought that theory had merit.

The problem is to test that will run you almost $10K for that setup.

asus389 The reason given for this is that MacOS doesn’t trust 3rd party monitors to do FRC consistently, so for the sake of consistency it’s done on the GPU side essentially bypassing the external monitors own FRC implementation but fulfilling the “10 bit capable” promise.

This sounds plausible.

asus389 With respect to their own built in displays, Apple supposedly trusts their own FRC and TCON so it sends a 10 bit signal to the built in display and let’s the panel use FRC with the TCON to simulate 10 bit color. This is in contrast to 3rd party external displays where for consistency sake they do the dithering on the GPU side and send an 8 bit signal with temporal dithering.

Except the Macs are also using GPU dithering and FRC. The question then becomes where one begins and ends. It seems highly redundant to dither at the GPU level and then also apply FRC. But it is conceivable they’re using different techniques that work in tandem.

Come to think of it…has anyone actually found any information in Apple patent documents specifically mentioning panel FRC?

asus389 Chat GPT is known to be wrong about things at times so idk how accurate this is. Although “not trusting 3rd party FRC and doing it themselves for consistency sake on the GPU/OS side” would be on brand for Apple. It’s a bit counter intuitive, however.

Can anyone who’s actually tested these combinations comment on the accuracy of this?

Grok was giving me links to fake scientific studies about flicker and I had to call it out two times before it stopped and copped to making it up. So I don’t even know what to think as far as accuracy.

asus389

So interestingly ChatGPT thinks the following of how MacOS decides what signal to send to an external monitor and if it enables GPU side dithering. This is apparently informed by what the EDID tells the host as far as the monitor capabilities.

6 bit + FRC - 8 bit signal sent from Mac OS. no GPU dithering, panel uses internal FRC to approximate 8 bit color.

8 bit (True 8 bit) - 8 bit signal sent from Mac OS. No GPU dithering, panel does not use FRC. Panel displays 8 bit color natively.

10 bit (8 bit + FRC) - 8 bit signal sent from Mac OS with temporal dithering done at the GPU level. Panel uses 8 bit mode, but dithering on the GPU side simulates 10 bit color. The reason given for this is that MacOS doesn’t trust 3rd party monitors to do FRC consistently, so for the sake of consistency it’s done on the GPU side essentially bypassing the external monitors own FRC implementation but fulfilling the “10 bit capable” promise.

10 bit (True 10 bit) - No GPU dithering, panel does not use FRC. Panel displays 10 bit color natively.

With respect to their own built in displays, Apple supposedly trusts their own FRC and TCON so it sends a 10 bit signal to the built in display and let’s the panel use FRC with the TCON to simulate 10 bit color. This is in contrast to 3rd party external displays where for consistency sake they do the dithering on the GPU side and send an 8 bit signal with temporal dithering.

I asked for references and it vaguely pointed to example code or code in the Darwin display drivers and references to WWDC sessions and documents.

Chat GPT is known to be wrong about things at times so idk how accurate this is. Although “not trusting 3rd party FRC and doing it themselves for consistency sake on the GPU/OS side” would be on brand for Apple. It’s a bit counter intuitive, however.

Can anyone who’s actually tested these combinations comment on the accuracy of this?

WhisperingWind

asus389

10 bit (8 bit + FRC) - 8 bit signal sent from Mac OS with temporal dithering done at the GPU level. Panel uses 8 bit mode, but dithering on the GPU side simulates 10 bit color. The reason given for this is that MacOS doesn’t trust 3rd party monitors to do FRC consistently, so for the sake of consistency it’s done on the GPU side essentially bypassing the external monitors own FRC implementation but fulfilling the “10 bit capable” promise.

10 bit (True 10 bit) - No GPU dithering, panel does not use FRC. Panel displays 10 bit color natively.

As I mentioned above, macOS cannot obtain information from the monitor about whether it uses FRC.

But the GPU driver developer knows that certain iMac/MacBook models use a display with FRC, so the model number is hard-coded in the driver code, and GPU-side temporal dithering is disabled specifically for that model.

However, that same developer knows nothing about external third-party monitors, therefore the behavior should be the same for both 10-bit (8-bit + FRC) and 10-bit (true 10-bit) external third-party monitors, since macOS cannot see the difference between them; to it, they are both 10-bit. Therefore, ChatGPT is mistaken—the behavior in both cases should be the same. Yet in its explanation, in one case GPU-side temporal dithering is present, and in the other it is absent.

Perhaps ChatGPT is confusing the displays used in Apple devices with external third-party monitors (or maybe I misunderstood the post).

JTL

WhisperingWind Perhaps ChatGPT is confusing the displays used in Apple devices with external third-party monitors (or maybe I misunderstood the post).

I think you understood at least this part correctly. I was going to raise the same discrepancy.

AshX

WhisperingWind But the GPU driver developer knows that certain iMac/MacBook models use a display with FRC, so the model number is hard-coded in the driver code, and GPU-side temporal dithering is disabled specifically for that model.

Do you have a source for this? I’ve never seen this confirmed.

What would be the benefit of relying on FRC? Saving GPU resources? I suppose that would have made more sense on Intel Macs which were more bottlenecked depending on the configuration.

asus389

WhisperingWind

Yeah IDK, apparently ChatGPT thinks that the OS can use the EDID CTA-861 Extension Block to tease out what the monitor supports with respect to 8/10 bit etc. I'm not familiar enough with this to know if its accurate.