win11 uhd hdr

As a totally new beginner (knowingly stepping into the deep end of the pool), I am attempting
to learn Davinci Resolve 17.4.6 editing of ProRes422hq source materials. I have a Dell medium
powered system (Alienware R11 w/upgrades) and an AW2521 monitor that does IMHO a very
reasonable job of displaying UHD HDR sample videos from YouTube. Beyond the obvious
first step of finding a current WinTel example of how to post process input from my camera,
I have been searching for a Win11(x64) utility that will allow my to display what has been
exported without uploading to YouTube first. Any pointers on either of these two issues
would be deeply appreciated.
Thanks in Advance,
New to all of this

As a totally new beginner (knowingly stepping into the deep end of the pool), I am attempting
to learn Davinci Resolve 17.4.6 editing of ProRes422hq source materials. I have a Dell medium
powered system (Alienware R11 w/upgrades) and an AW2521 monitor that does IMHO a very
reasonable job of displaying UHD HDR sample videos from YouTube. Beyond the obvious
first step of finding a current WinTel example of how to post process input from my camera,
I have been searching for a Win11(x64) utility that will allow my to display what has been
exported without uploading to YouTube first. Any pointers on either of these two issues
would be deeply appreciated.
Thanks in Advance,
New to all of this

Not within Resolve. Resolve in Windows cannot preview in HDR in the app video screen. You will need a BM hardware to feed an HDR monitor in order to view HDR from Resolve.

VLC Media Player I think supports HDR playback.

Does the AW2521 support HDR?

I think for review you should get the export off the computer entirely Take the OS, GPU driver and software out of the signal chain.

I use a calibrated Home Theater for review. It's a but clunky, but it works. The best option is below:

Yes, the aw2521 does support HDR (only up to 400 nits). I was going to add "max bit depth and render quality (similar to Premiere Pro)" for the preview window to the wishlist, but if the product team is fixated on forcing external monitors I guess I'm stuck.

Check if NobeDisplay mabe can do it or another OFX solution.

I think for review you should get the export off the computer entirely Take the OS, GPU driver and software out of the signal chain.

I use a calibrated Home Theater for review. It's a but clunky, but it works. The best option is below:

It's kinda funny that not a single person ever advocating for IO devices bothered once to run both the bad evil GPU signal and the holly Decklink signal back in into a capture card and compare them to back their claims for "Better quality". I did with color ramps. Can't find a difference.

Also this whole OS color management blah blah is exclusive to MacOS. There is no such thing on Windows.

There is a point for these devices, but certainly not do they provide magically better quality. However quality is even defined.

For the professionals I know, it is about accuracy and control and not necessarily "quality".

It all depends on what you are trying to achieve. If you are coloring and mastering for 2084 / 100 nits, 450 nits with 85% P3 coverage will put you in a precarious spot.

There is a point for these devices, but certainly not do they provide magically better quality.

I agree. That is not their point.

Accuracy in the signal is their point. It's not always guaranteed without one.

PeterJackson is 100% correct. Even if you use an io device, (which I do, the BM MiniMonitor 4K), what makes you think the io device itself is accurate, doesn't need to be calibrated?

When you use CalMan, a pattern generator can be used for the display. This is what the hired calibrators use. The disadvantage to that approach is that the GPU or IO device isn't included in the measurement. If you opt for the expensive version of the CalMan software, you have the option to use daVinci Resolve itself to be the pattern generator so that you have an end-to-end calibration that includes the GPU or IO.

MadVR can be used as a pattern generator with HCFR for the signal chain GPU-thru-display but not for IO. So now what? CalMan $$$$ is the only option for IO.

The bottom line here is that Resolve doesn't permit HDR without an IO even though it enables HDR metadata over HDMI with IO but not GPU, while other Windows softwares like the Edge Browser, Firefox, Chrome, VLC, WMP, MPC-HC (x64) HCFR/MadVR do.

I'm sorry, but I can't help believing this omission for Resolve is not unintended, but arbitrary and unfortunate for Windows users.

To answer the OP's original question, the most convenient player for HDR files (not YouTube) is Media Player Classic with the MadVR DirectShow filter installed. With that, when you open an HDR file it automatically switches the display into HDR (PQ-2084) mode and when you open an SDR file it opens in SDR without having to switch back and forth as with the Windows system setting.

Even if you use an io device, (which I do, the BM MiniMonitor 4K), what makes you think the io device itself is accurate, doesn't need to be calibrated?

wellll...... because it is.

an IO device just put out the digital signal itself, the display it is responsible of correctly transform that digital value in a correct amount of light/color/whatever....

from the IO device point of view, you give 1 from your hard disk, he give 1 to the display. (with the proper scaling when applicable).

if that was not the case, nothing could have ever worked.....

Okay, so I/O digital 1:1. How is that different than GPU pass through?

Basically this, I don't observe any difference between Mini Monitor 4K and NVIDIA 3090 GPU color wise or gamma in SDR or HDR. I don't see how the 1's and 0's would be different with the OS and GPU in the signal chain if no ICC color profiles are in play in Windows Color Management, no color profiles, just the default PnP generic monitor, and the GPU color management set to override to the reference mode. No color adjustments possible in OS or GPU.

Because the OS can interfere.

Operating system doesn't interfere when no ICC profiles are used.

Do as you like then. It doesn't matter anyway on the internet.

DaVinci Resolve 18 public beta has released.
Still do Windows users need Decklink to monitor HDR videos, or is there any update for the weakness?
What's New page says "Laptop HDR Monitoring", but is it only for Mac?
I would appreciate knowing about it.

GPU doesn't interfere if color accuracy mode is set to reference and adjustments are set to be made by the player, (e.g. Resolve).

Do as you like then. It doesn't matter anyway on the internet.

Not so fast Uli.

Your profile says you have an M1 MacBook Pro, isn't that right? According to some sources, it's possible to set the profile for it to 10 bit and Mac profile and view HDR on the MacBook screen. Is that true?

Yeah. Here it is. The link.
https://daejeonchronicles.com/2021/09/2 ... ve-viewer/

Is this true?

from the IO device point of view, you give 1 from your hard disk, he give 1 to the display. (with the proper scaling when applicable).

if that was not the case, nothing could have ever worked.....

Time was, when hardware had some analog components, production tolerances, temperature, drift and aging.

Yes, it works on the Mac.

Yes, it works on the Mac.

Uh huh. And that beautiful Mac has an operating system, no? And a wonderful integrated GPU no? And no I/O box was required to display HDR on the viewer.

That pisses me off and the grudge goes back to Resolve 12.0b, the first version with HDR. Resolve since gained the ability to enable HDR metadata over HDMI but on the PC platform that only applies when a decklink card is used. But if you have a Mac, you don't need the decklink, you can enable the HDR right on the viewing window.

Even if you assign an ICC profile in Windows it doesn't change anything. Windows has no system wide color management. Every application has to do it's own color management. And Resolve can't use ICC at all.

The only thing that ever can cause a difference is if the ICC contains a VCGT an this is loaded into the GPU.

Windows does not do any color management.

The GPU drivers don't do any color management.

The GPU can load a gamma table if you instruct it to so.

The only practical way to mess up on windows is to load a VCGT. And I have no idea how one could ever do this by accident.

Outputs are bit identical. There is no additional accuracy gained by an IO card.

If you think there is, conduct actuall messurements and post them here. All this warm fuzzy gut feeling talk is a waste of life time. If you have a point, provide any sort of evidence.

When you use CalMan, a pattern generator can be used for the display. This is what the hired calibrators use. The disadvantage to that approach is that the GPU or IO device isn't included in the measurement. If you opt for the expensive version of the CalMan software, you have the option to use daVinci Resolve itself to be the pattern generator so that you have an end-to-end calibration that includes the GPU or IO.

DisplayCal provides exactly the same functionality. Just select its Resolve preset and it will wait for a Calman connection coming from Resolve.

Until recently, Apple didn't do any better in an affordable display. AFAIK, they are still not telling the makers of probes the spectral parameters of their screens and light sources. So, what about drift of electronic components?

But anyway, as long as you want some kind of HDR, who cares? True, standard-compliant HDR screens are unaffordable and nobody has them in their home or in their hands anyway.

AFAIK, they are still not telling the makers of probes the spectral parameters of their screens and light sources. So, what about drift of electronic components?

What could Apple tell probe vendors that they can't measure themselfs?

There is drift in electronic components. But only screens and probes matter in this context as they are the only thing that does AD/DA conversion.

I haven't tried Resolve 18 with 10 bit viewers yet, but unless that's buggy that even lesses the case for dedicated IO hardware.

There is nothing wrong with IO hardware. It has plenty of ports and special connectivity no consumer GPU offers. This is what you buy them for. Not for a more accurate signal.

Since all pro users own these devices. What is preventing you guys from doing a simple loop back test and capture the signal from the IO box and the viewer GPU and compare them? It's like people don't want to give up a religion they are following for so long.

There is actually a good case for IO hardware. Smooth 60fps playback on a 60fps screen. Resolve still can do it properly on Windows. Admittedly I haven't tried 18 yet.

This non sense blah blah of Windows, drivers and GPU somehow messing with the signal. It doesn't happen and doesn't exist on Windows. Pure FUD: https://en.m.wikipedia.org/wiki/Fear,_u ... _and_doubt

Even the most expensive screen needs calibration. The fact that they store their LUT internally doesn't make them better in any way.

And even though I might sound harsh at times, I really appreciate everyone on these forums. I'm just not into religions and those religions beeing force fed every new member of these forums.

I'm also aware that Resolve Studio is incredibly low price and BM needs to make money. That's surely in my mind when buying BM hardware.

Since all pro users own these devices. What is preventing you guys from doing a simple loop back test and capture the signal from the IO box and the viewer GPU and compare them? It's like people don't want to give up a religion they are following for so long.

There is actually a good case for IO hardware. Smooth 60fps playback on a 60fps screen. Resolve still can do it properly on Windows. Admittedly I haven't tried 18 yet.

I would do the loopback but the Mini Monitor 4K is output only. It can do 4K playback at only up to 30 fps. 60 fps is possible at 1080p. But this limitation is the reason I would to like to engage the app viewer, since the 3090 will smoothly playback 8K, even 12K at 60 fps to the screen's native 7680x4320 resolution. Resolve just won't permit enabling HDR metadata over HDMI except for its own decklink. AFAIK, there is no technical reason this should not be done. The reasons given so far are as you say, religion.

from the IO device point of view, you give 1 from your hard disk, he give 1 to the display. (with the proper scaling when applicable).

if that was not the case, nothing could have ever worked.....

Time was, when hardware had some analog components, production tolerances, temperature, drift and aging.

in analog days, sure, in digital days, no.

But displays still have analog components and ageing of the light source. Why would you need to warm them up before calibration?
As I already wrote, I/O devices or calibrated screens don't matter for 99.9 % of your audience on the internet, because their devices are not calibrated either. But it might be helpful to know that you sit somewhere in the middle of non-calibration instead of one of the extremes …

But displays still have analog components and ageing of the light source. Why would you need to warm them up before calibration?
As I already wrote, I/O devices or calibrated screens don't matter for 99.9 % of your audience on the internet, because their devices are not calibrated either. But it might be helpful to know that you sit somewhere in the middle of non-calibration instead of one of the extremes …

That's what profile verification is for in common software. Quick check if there is a larger delta, if so create new calibration. Often done once a month. At least in print. Also the probe should be left on screen during warmup.

Calibrating a screen for web delivery is still important. Particular as better screens these days have wide gammut. How would you grade on these without a 3D LUT constraining them to Rec709 gammut? Same for gamma. How would grade on a 2.2 screen if delivery is supposed to target 2.4. A 3D LUT takes care of this. I usually use two. One for 2.2. and one for 2.4 gamma.

If that 3D LUT lives in Resolve, an expensive pro screen or external LUT box doesn't matter in terms of accuracy. But does in convince. I certainly don't want to reach so some external hardware just to switch a LUT. And externals LUT solution affect the entire screen, while doing it in software affects just the viewer/app in question. That is not only more correct but also allows for comparisons between different software / players / color management solutions.

I tend to watch YT with MPV and I have a gamma 2.4/2.2 toggle there. There seems to be little consent. Some YT content seems to target 2.4, some 2.2.

And many apps don’t respect flags (if there are any).

I guess the sad YT reality is:

Grade for Rec709 gammut while keeping saturation defensive to keep skin tones somewhat acceptable on non-managed P3 displays.

Tag Rec709 (Scene) so YT doesn't gamma convert.

Target gamma 2.2 for anything that isn't likely to be primarily be viewed on TVs.

Firefox, Chrome, Edge all don't seem to color manage YT/video on Windows even if color management is enabled explicitly. Firefox and Chrome based browsers use different gamma. So that end is pretty fckd up.

I use color managed MPV.NET to judge streaming sites. Exactly matches Resolve viewers. May put my config on github at some point.

Very interesting debate. It reminds me a lot of the spurious reasons the pro audio industry give for Windows needs X hardware, it's not technically possible like on a mac etc, yet it's often guff or based on pre-conceived assumptions from 20 years ago. I even sent Dolby clear evidence the reasons they gave for not having a Windows version was nonsense. The conversation ended when they saw it was all nonsense what they were saying.

Resolve just won't permit enabling HDR metadata over HDMI except for its own decklink. AFAIK, there is no technical reason this should not be done. The reasons given so far are as you say, religion.

There is likely no technical reason. Just a financial one. It keeps Resolve Free... free and paid version $300.
BM is a hardware company, just like Apple. They sell their software at low price in order to sell hardware. Just like Apple.

Resolve just won't permit enabling HDR metadata over HDMI except for its own decklink. AFAIK, there is no technical reason this should not be done. The reasons given so far are as you say, religion.

There is likely no technical reason. Just a financial one. It keeps Resolve Free... free and paid version $300.
BM is a hardware company, just like Apple. They sell their software at low price in order to sell hardware. Just like Apple.

That's fine. Luckily they don't make only IO boxes, but all sorts of good gear.

Seems a little OS biased from Blackmagic that a Windows user must buy their hardware to do what a Mac user can do without, even though both OS are capable of doing the same thing natively without? You see this time and again in the apple-centric media industry I find.

I think 10bit viewers in Resolve 18 is already a good sign that HDR support might be on the agenda

Just to put a final note on this (from me, not the community).
beta 18.5 appears to be using the native 10 bit support via Win11 and the monitor.
My current job flow is to render the YouTube out as a file, but not allow DR to attempt an upload. I check
for acceptable quality by using CNX Player (available from the M$ store) to play the file. If I'm happy
I then upload the video. I find YouTube's current process interesting in that they take the HDR file and clone an SDR file of it so that it can be viewed by users who may not have the appropriate hardware, the
selection of what they send is automagic.
Thanks for the rigorous debate. As an aside I calibrate my native monitors monthly using Calibrite Video Checker.
Al Edlund

It's kinda funny that not a single person ever advocating for IO devices bothered once to run both the bad evil GPU signal and the holly Decklink signal back in into a capture card and compare them to back their claims for "Better quality". I did with color ramps. Can't find a difference.

You are right very few people have compared it here is one I found

Okay, so I/O digital 1:1. How is that different than GPU pass through?

Basically this, I don't observe any difference between Mini Monitor 4K and NVIDIA 3090 GPU color wise or gamma in SDR or HDR. I don't see how the 1's and 0's would be different with the OS and GPU in the signal chain if no ICC color profiles are in play in Windows Color Management, no color profiles, just the default PnP generic monitor, and the GPU color management set to override to the reference mode. No color adjustments possible in OS or GPU.

Same 0s and 1s (or even better ones as without any conversions straight from GPU), but dedicated card makes things easier. Just don't forget about calibration/validation.
There are small nuances and to have all working at 100% you need proper support from software side and I don't think Resolve (specially on Widows) is there yet.
There are tools which do support it- eg. Scratch, Flame.

This non sense blah blah of Windows, drivers and GPU somehow messing with the signal. It doesn't happen and doesn't exist on Windows. Pure FUD: https://en.m.wikipedia.org/wiki/Fear,_u ... _and_doubt

It's a story from 30 years ago (when it was actually true to some extent) which has been repeated all over internet, by many 'experts', who do not much except repeating it
It's 2022 and today's GPUs can easily replace dedicated IO cards. It's just a matter of software side to use them properly. Please don't generalize it- if Resolve can't do something it doesn't mean it can't be done at all.

But displays still have analog components and ageing of the light source. Why would you need to warm them up before calibration?
As I already wrote, I/O devices or calibrated screens don't matter for 99.9 % of your audience on the internet, because their devices are not calibrated either. But it might be helpful to know that you sit somewhere in the middle of non-calibration instead of one of the extremes …

That's what profile verification is for in common software. Quick check if there is a larger delta, if so create new calibration. Often done once a month. At least in print. Also the probe should be left on screen during warmup.

Calibrating a screen for web delivery is still important. Particular as better screens these days have wide gammut. How would you grade on these without a 3D LUT constraining them to Rec709 gammut? Same for gamma. How would grade on a 2.2 screen if delivery is supposed to target 2.4. A 3D LUT takes care of this. I usually use two. One for 2.2. and one for 2.4 gamma.

If that 3D LUT lives in Resolve, an expensive pro screen or external LUT box doesn't matter in terms of accuracy. But does in convince. I certainly don't want to reach so some external hardware just to switch a LUT. And externals LUT solution affect the entire screen, while doing it in software affects just the viewer/app in question. That is not only more correct but also allows for comparisons between different software / players / color management solutions.

I tend to watch YT with MPV and I have a gamma 2.4/2.2 toggle there. There seems to be little consent. Some YT content seems to target 2.4, some 2.2.

There is tinny hope Vimeo will start passing through properly sRGB graded/tagged videos (without overwriting tags to 1-1-1), so this will give correct preview on Macs by default and also in some cases (depending on browser/player) on PCs.
Took them few months just to take it as feature request, so don't put big hope

Seems a little OS biased from Blackmagic that a Windows user must buy their hardware to do what a Mac user can do without, even though both OS are capable of doing the same thing natively without? You see this time and again in the apple-centric media industry I find.

Windows color management is non-existent I think, so BM can't do much about this. They would have to do it differently than on OSX, but of course they sell hardware for monitoring, so it's all matter of business decisions.

It's kinda funny that not a single person ever advocating for IO devices bothered once to run both the bad evil GPU signal and the holly Decklink signal back in into a capture card and compare them to back their claims for "Better quality". I did with color ramps. Can't find a difference.

You are right very few people have compared it here is one I found

It's a lot of talking. Real story would be a proper calibration etc. and then validation between both ways.
This is not it at al. It's just - look different NLEs/players and monitors look different. We already know this.

You are right, this was the least comparison (without calibration) which I could find on youtube.