Using the correct tags for delivery. Am I set up correctly?

Hi folks, apologies in advance for ANOTHER thread asking about tags/gamma/rec 709a etc

I'm hoping this is just a case of getting confirmation i'm set up correctly as for the most part everything is doing what I would expect.

So the setup I want to check consists of:

1. a MacBook Pro M3 Max with the display set to HDTV (709 - 1886)
2. a portable Rec709 monitor connected via an Ultra studio 3G monitor
3. Resolve set to use Mac display colour profiles for viewers
4. My timeline set to a Rec709 a colour space

With the above set this way my Resolve monitors match my external display, so I think I can safely say this part of my setup is golden, right?

Now I would like to export some sequences that need to be suitable for both broadcast and web, so I'm using tags to set my deliverables correctly. The settings are as follows:

1. The broadcast deliverable is encoded as ProRes 422HQ with a colour space tag of Rec709 and a gamma tag of 2.4
2. The web deliverable is encoded as H265 with a colour space tag of Rec709 and a gamma tag of 2.2

Once the above files are exported I can re-import them into resolve and everything looks the same and matches my sequence, so as far as I am concerned the file itself is encoded correctly

So the final part of the puzzle is the colour space and gamma tags which display as follows:

1. The broadcast deliverable has 1:2:1 tags and looks too dark when played in QT player
2. The web deliverable has 1:4:1 tags and looks exactly the same as what I see in Resolve when played in QT player.

So my question is, is the above behaviour as expected and am I setting the correct tags? I'd hate to think I'm doing something stupid that is messing up my deliverables at the final hurdle. As an additional (possibly stupid) extra question; do I even need to set gamma tags for the broadcast deliverables, or is this kind of meta data only really important for web based delivery?

Cheers
Rich

If you deliver for broadcast then all fine just custom tag file as 1-1-1 and leave it this way. Tag is really not that important for further delivery. What is important is grading to 2.4 gamma on correctly set display.
Web delivery will lok different on YT anyway, unless you grade for Mac users specifically and use Rec.709-A gamma on export.

If you deliver for broadcast then all fine just custom tag file as 1-1-1 and leave it this way. Tag is really not that important for further delivery. What is important is grading to 2.4 gamma on correctly set display.
Web delivery will lok different on YT anyway, unless you grade for Mac users specifically and use Rec.709-A gamma on export.

Hey Andrew, I was kind of hoping you might be the first one to respond

So to clarify, when exporting for broadcast I'm fine to set the export tags as 1-1-1 (Rec709-A?) and that is because the tags are just metadata designed for web and media players etc and not really relevant in a TV station? and using the 1-1-1 or Rec709-A tags will allow me to view everything the same way on my Mac. Have I understood all of that correctly?

And for web delivery (Vimeo,Youtube etc) The safest bet is 1-4-1 or gamma 2.2 tags as this gives something that should be pretty damn close for most situations?

So my final check list becomes...

1. Mac display set to HDTV (709 - 1886) or reference mode
2. a portable Rec709 monitor connected via an Ultra studio 3G monitor
3. Resolve set to use Mac display colour profiles for viewers
4. My timeline set to a Rec709-A colour space
5. The broadcast deliverable is encoded as ProRes 422HQ with a colour space tag of Rec709 and a gamma tag of Rec709-A
6. The web deliverable is encoded as H265 with a colour space tag of Rec709 and a gamma tag of 2.2

Have I understood this correctly?

Cheers
Rich

For broadcast just grade to 2.4 (definitely not Rec.709-A) on properly set monitor and done. File should be rather tagged 1-1-1 (Rec.709-A or scene preset).
If you want to review broadcast file on Mac screen then there are 2 ways:
- 2.4 graded file with 1-1-1 tag + new Mac screen set to BT.1886 reference preset
- for older screens 2.4 graded file needs 1-2-1 (+2.4 gamma tag), which can be done with AMCDX tool

For web there is no real solution if you want it global. You either target all users with eg. sRGB preset (this may be better than 2.2 one) or if it's for Mac users then you can export as Rec.709-A (so grade normal as 2.4 or 2.2 and on export set to Rec.709-A gamma).

For broadcast just grade to 2.4 (definitely not Rec.709-A) on properly set monitor and done. File should be rather tagged 1-1-1 (Rec.709-A or scene preset).
If you want to review broadcast file on Mac screen then there are 2 ways:
- 2.4 graded file with 1-1-1 tag + new Mac screen set to BT.1886 reference preset
- for older screens 2.4 graded file needs 1-2-1 (+2.4 gamma tag), which can be done with AMCDX tool

For web there is no real solution if you want it global. You either target all users with eg. sRGB preset (this may be better than 2.2 one) or if it's for Mac users then you can export as Rec.709-A (so grade normal as 2.4 or 2.2 and on export set to Rec.709-A gamma).

Sorry, maybe I've misunderstood you but when you say 'just grade to 2.4 (definitely not Rec.709-A)' do you mean my timeline setting should be gamma 2.4? I thought if my XDR screen is in reference mode (HDTV 1886 preset) I should have my timeline in Rec.709-A to display content correctly?
Or do you mean the timeline setting is Rec.709-A and the output setting is Rec.709 gamma 2.4? This gives me a match between Resolve and my external monitor and is what I presumed was correct

If it's ok I do have two other questions...

1. Is it at all dangerous to send out a ProRes file with 1-1-1 tags for broadcast? Is there a risk that using these tags and not 2.4 gamma tags (1-2-1) could cause the file to be incorrectly interpreted? Or are these tags irrelevant in the broadcast world?

1. I've seen mention that this kind of setup only really works if you don't use colour management in Resolve due to the clash between Apple's colour management and BlackMagic's. If I want to have a colour managed workflow, does this mean there is no accurate way to see a correct image in the Resolve GUI displays and I would always need to use an external display? Or are there compromises I could make to get something at least close to accurate?

Your output (project level color management setting) should be set to 709 gamma 2.4 to match your grading monitor but then you manually change the metadata tagging on the delivery page to tag as 1-1-1 (which is Rec709-A). In my workflow, I use a much larger gamut for the timeline and typically a log gamma because I like the feel of the controls in log. In your case, if all your media is already Rec709 then using a Rec709 gamma 2.4 timeline would likely be fine.

Timeline can be anything (typically colorists use log option). Output and your reference screen settings are what counts.
File should be graded to 2.4 and exported with 1-1-1 tag for broadcast. This is correct way. Don't use custom tags for broadcast as they do nothing (actually can cause rejection in some cases).
If you have one stick to external preview for broadcast grading.
If I remember well ref mode BT.1886 should work fine with Resolve preview. There was a thread about it.

Your output (project level color management setting) should be set to 709 gamma 2.4 to match your grading monitor but then you manually change the metadata tagging on the delivery page to tag as 1-1-1 (which is Rec709-A). In my workflow, I use a much larger gamut for the timeline and typically a log gamma because I like the feel of the controls in log. In your case, if all your media is already Rec709 then using a Rec709 gamma 2.4 timeline would likely be fine.

Hi Steve, thanks for the response

So in this particular scenario where I am working in a non colour managed project I am as you suggested setting my output colour space to Rec.709 2.4 as I am delivering SDR content and I have my timeline set as Rec.709-A as I am editing on a new MacBook with an XDR screen that has an HDTV 709-1886 reference setting. As I understand it, this reference setting combined with Rec.709-A counters Apple's gamma implementation (often referred to as the gamma bug) and allows me to have my external calibrated monitor, my Resolve viewers and my QT exports (If I export using 1-1-1 flags) all looking the same. This is particularly handy for me as I regularly need to work on the road and don't always have access to my calibrated monitor.

I think the above is a safe way of working in a non-colour managed workflow, although I'd love to get confirmation from those more knowledgeable in this area that I'm not doing something stupid?

Timeline can be anything (typically colorists use log option). Output and your reference screen settings are what counts.
File should be graded to 2.4 and exported with 1-1-1 tag for broadcast. This is correct way. Don't use custom tags for broadcast as they do nothing (actually can cause rejection in some cases).
If you have one stick to external preview for broadcast grading.
If I remember well ref mode BT.1886 should work fine with Resolve preview. There was a thread about it.

Thanks Andrew,

When you say custom tags, what do you mean by that? 1-1-1 is not a custom tag?

Resolve, QT and external preview should look 'the same' for 2.4 grading if you use BT.1886 reference mode for your screen. This should not depend on the timeline setting. If it doesn't your Rec.709-A timeline setting workaround sounds fine (if then preview is the same).
I'm not sure sure bout latest Resolve as BM keeps changing things behind the scenes way too often

Timeline can be anything (typically colorists use log option). Output and your reference screen settings are what counts.
File should be graded to 2.4 and exported with 1-1-1 tag for broadcast. This is correct way. Don't use custom tags for broadcast as they do nothing (actually can cause rejection in some cases).
If you have one stick to external preview for broadcast grading.
If I remember well ref mode BT.1886 should work fine with Resolve preview. There was a thread about it.

Thanks Andrew,

When you say custom tags, what do you mean by that? 1-1-1 is not a custom tag?

No, 1-1-1 is an "official" tag for HD SDR file.
Custom tag is eg. 1-2-1 or actually anything else than 1-1-1 for standard HD Rec.709 grade.
Also if you use BT.1886 reference mode for your screen then tag should be 1-1-1.

Urghhhh! Just when I thought I had a handle on this I realise I haven't after all

I just noticed my display settings weren't in fact set to HDTV video after all, I actually had it in the HDR (P3 ST-2084) reference preset

So here's the weird bit... everything looks correct like this! But if I switch to the BT.1886 preset resolve suddenly looks a lot darker than my calibrated monitor, which makes no sense to me as this should be the correct setting and not HDR mode

Is this a possible Mac OS bug? I'm genuinely baffled now

What is your output color space set to in your project settings?

What is your output color space set to in your project settings?

Rec.709 2.4. That's never changed


Sent from my iPhone using Tapatalk

Urghhhh! Just when I thought I had a handle on this I realise I haven't after all

I just noticed my display settings weren't in fact set to HDTV video after all, I actually had it in the HDR (P3 ST-2084) reference preset

So here's the weird bit... everything looks correct like this! But if I switch to the BT.1886 preset resolve suddenly looks a lot darker than my calibrated monitor, which makes no sense to me as this should be the correct setting and not HDR mode

Is this a possible Mac OS bug? I'm genuinely baffled now

HDR screen preset with SDR Resolve project may be auto switching to SDR mode and then gamma is BT.1886, so this is why it's correct. Not so sure about it though.

What is your output color space set to in your project settings?

Rec.709 2.4. That's never changed


Sent from my iPhone using Tapatalk

First- if you export eg. ProRes file tagged as 1-1-1 and open in QT how does it look then compared to reference screen (with BT.1886 reference mode for Mac screen)? If this looks fine then do below.

Copy BT.1886 ref mode, go inside copied one and disable gamma boost. Check then Resolve preview vs. ref screen. In this case Resolve may look fine, but QT preview won't If this is the case then Resolve is messing with Apple color sync.

What is your output color space set to in your project settings?

Rec.709 2.4. That's never changed


Sent from my iPhone using Tapatalk

First- if you export eg. ProRes file tagged as 1-1-1 and open in QT how does it look then compared to reference screen (with BT.1886 reference mode for Mac screen)? If this looks fine then do below.

Copy BT.1886 ref mode, go inside copied one and disable gamma boost. Check then Resolve preview vs. ref screen. In this case Resolve may look fine, but QT preview won't If this is the case then Resolve is messing with Apple color sync.

Ahhhhhhhhhh, give that man a prize!

Thank you Andrew, you were spot on! I tweaked the display profile to disable the gamma boost and what would you know? Resolve and the reference monitor now match and the QT file has boosted gamma

I'm currently on the Resolve studio 19 beta so could that be why Resolve is clashing with Apple colour sync?

EDIT:

Interestingly it also fixes a long term issue I've had where using the iPad in reference mode it didn't match Resolve. It looks like the iPad was always right but Resolve on the MacBook was too dark. Very interesting

No, it's Resolve related. Way how BM done it (looks like they compensate preview to 2.4 gamma, where they should leave it for color sync to do it through this "gamma boost" setting).
I would rather preferred for it to work with standard ref mode, so then it's all as it should be.

You mean QT file is brighter? This is what I would expect.

No, it's Resolve related. Way how BM done it (looks like they compensate preview to 2.4 gamma, where they should leave it for color sync to do it through this "gamma boost" setting).
I would rather preferred for it to work with standard ref mode, so then it's all as it should be.

You mean QT file is brighter? This is what I would expect.

Yes that's exactly what I mean, the QT file is brighter

So other than QuickTime displaying the wrong gamma when 'gamma boost' is switched off, can you think of any other downsides?

Ok, at least this makes sense. This is good news

If QT displays it wrongly then anything color sync controlled will as well (Safari, etc.)
In this scenario Resolve is wrong and I hope BM will fix it, so you can leave default BT.1886 ref mode and all will be correct.
Interesting that HDR mode works fine. Maybe this is actually a solution as Mac will (I assume switch) to BT.1886 for any SDR source even if you are in HDR preset. And Resolve preview is correct as well as any HDR source will if you decide to play it

This is very interesting. If I may ask, when you had gamma boost enabled for your display, did the file played in QuickTime look the same as the file played out to your reference monitor, but the display in Resolve was too dark?

Yes, that's the problem. Resolve preview seems to apply 1.96 to 2.4 conversion twice (by itself and by color sync).

This is very interesting. If I may ask, when you had gamma boost enabled for your display, did the file played in QuickTime look the same as the file played out to your reference monitor, but the display in Resolve was too dark?

Yes it did, that's exactly what was happening

Yes, that's the problem. Resolve preview seems to apply 1.96 to 2.4 conversion twice (by itself and by color sync).

So it appears BlackMagic have optimised Resolve for non-XDR displays, ie without the Apple Gamma boost

Surely a simple fix from BlackMagic would be to have the gamma boost in Resolve as a switch that can be turned on and off dependant on whether you have an XDR display or not

Yep, or it can even work behind the scenes by checking if BT.1886 ref mode is selected or not.

You can also try to disable "use system color profiles" in the Resolve main settings and then it may actually work fine with standard BT.1886 ref mode.

Where is that menu Andrew?

In Resolve main settings.
Not sure if this will properly limit gamut for the screen. Try saturated colors so then it's easy to tell.

You can also try to disable "use system color profiles" in the Resolve main settings and then it may actually work fine with standard BT.1886 ref mode.

Andrew, you're a genius! That was it!

Resolve displays, reference display and Quicktime are all now matching

Thank you!

Now by changing gamma boost you can for example get 2.2 gamma reference screen out of Mac display. You need 1.12 boost for this.
I assume gamut is properly limited to Rec.709? Colors saturation is correct ?

Now by changing gamma boost you can for example get 2.2 gamma reference screen out of Mac display. You need 1.12 boost for this.
Is gamut properly converted to Rec.709? Colors saturation is correct ?

Oh, yes of course! So I could have two display profiles on my Mac, one for broadcast and one for web delivery

Yes the gamut appears correct. Initially it was wrong and the Resolve viewers looked over saturated, but following a reboot everything was fine

We need to put this together as a guide


Sent from my iPhone using Tapatalk

We need to put this together as a guide

So you have left the gamma boost ON in the XDR color profile and turned off the 'Use Mac display profile for viewers' option? And your XDR color profile is set to BT.1886 - or did you modify the BT.1886 to a Gamma 2.2 with gamma boost to 1.12?

Sorry, trying to understand where you landed with all of these changes, lol.

So now things should look the same in Resolve as in QuickTime as in your reference monitor (I forget, is your reference monitor gamma 2.2 - or at least does it have a gamma 2.2 mode so that if you switch your XDR to a BT.1886 with gamma 2.2, gamma boost 1.12, does it look the same as your reference monitor?).

A summery of your final configuration would be very helpful to others. Cheers.

We need to put this together as a guide

So you have left the gamma boost ON in the XDR color profile and turned off the 'Use Mac display profile for viewers' option? And your XDR color profile is set to BT.1886 - or did you modify the BT.1886 to a Gamma 2.2 with gamma boost to 1.12?

Sorry, trying to understand where you landed with all of these changes, lol.

So now things should look the same in Resolve as in QuickTime as in your reference monitor (I forget, is your reference monitor gamma 2.2 - or at least does it have a gamma 2.2 mode so that if you switch your XDR to a BT.1886 with gamma 2.2, gamma boost 1.12, does it look the same as your reference monitor?).

A summary of your final configuration would be very helpful to others. Cheers.

Hey Steve,

The Mac display profile is untouched HDTV 709-1886
The 'Mac display profiles for viewers' setting in Resolve is turned off
Output colourspace is set to Rec.709 2.4 in Resolve
The reference display is also Rec.709 2.4
Everything matches!

So as Andrew suggested we could easily have a 2.2 gamma version which would require three things for matched content

1. Tweaking the 'apply system gamma boost' on the Mac 709-1886 colour profile to 1.12
2. change output colourspace in Resolve to Rec.709 2.2
3. Set a gamma of 2.2 on the reference monitor

I'll need to test the 2.2 workflow before confirming it definitely works but I'm shooting on location today so won't be able to check until this evening

It's all very interesting though as up until now I'd understood the 'use mac display profiles for viewers' as an essential setting when trying to view accurate output on the Mac, but it appears that isn't the case with an XDR screen. Is that Resolve setting potentially doing the same thing as the system gamma boost on the XDR reference profiles?

It's all very interesting though as up until now I'd understood the 'use mac display profiles for viewers' as an essential setting when trying to view accurate output on the Mac, but it appears that isn't the case with an XDR screen.

It was until Apple added reference modes.

Thanks for the detailed reply.

I'm using gamma 2.2 w/ gamma boost 1.12 on my MacBook and I have a reference monitor calibrated to Rec709 gamma 2.2. I've disabled the 'Use Mac display profile for viewers'.

Set my output color space to Rec709 gamma 2.2 and when I export a color managed clip tagging it as 1-1-1 it still is slightly off between what I saw in Resolve and what I see in QuickTime.

More experimenting is required.

Add - What's interesting is that if I bring the QT export back into Resolve and plunk it on the timeline (making sure Resolve knows that the QT is 1-1-1 (Rec709 gamma 709-a) then it looks identical in Resolve and in QuickTime with the above settings.

However, when I look at the original graded clip the gamma is slightly lower in Resolve. When I think about it, this is probably to be expected. There will be a slight shift because the output color space was actually gamma 2.2 but I'm tagging it as if it were gamma 1.96, I guess. I'll need to think about this last point a bit but I think the bottom line is that when I export gamma 2.2 tagged 1-1-1 there is a tradeoff I"m making such that this file should playback closer to what I see in Resolve across most players. Think of this last statement as notes to myself until I can do a bit more testing. Cheers.

Thanks for the detailed reply.

I'm using gamma 2.2 w/ gamma boost 1.12 on my MacBook and I have a reference monitor calibrated to Rec709 gamma 2.2. I've disabled the 'Use Mac display profile for viewers'.

Set my output color space to Rec709 gamma 2.2 and when I export a color managed clip tagging it as 1-1-1 it still is slightly off between what I saw in Resolve and what I see in QuickTime.

More experimenting is required.

That's interesting Steve. How off is it? Does the reference monitor match Resolve or QuickTime... Or neither?


Sent from my iPhone using Tapatalk

I added some detail to my previous post.

I added some detail to my previous post.

Yes, thinking about it the export should be tagged 1-4-1, not 1-1-1 no?


Sent from my iPhone using Tapatalk

Thanks for the detailed reply.

I'm using gamma 2.2 w/ gamma boost 1.12 on my MacBook and I have a reference monitor calibrated to Rec709 gamma 2.2. I've disabled the 'Use Mac display profile for viewers'.

Set my output color space to Rec709 gamma 2.2 and when I export a color managed clip tagging it as 1-1-1 it still is slightly off between what I saw in Resolve and what I see in QuickTime.

More experimenting is required.

That's interesting Steve. How off is it? Does the reference monitor match Resolve or QuickTime... Or neither?


Sent from my iPhone using Tapatalk

If I remember well it won’t match ( I forgotten about this 2.2 special case).
Resolve 2.2 gamma preset doesn’t encode files with 1.96 gamma, which is required for ref mode with boost to operate properly.

Apple color sync doesn’t understand 1-4-1 tag unfortunately ( unless it changed).

Solution may be use of reference mode which is not based on BT.1886 preset, but pure 2.2 gamma one. Not sure if 2.2 in Resolve is perfect curve or it has some linear bit as well. If so match won’t be 100%, but quite close.

Thanks for the detailed reply.

I'm using gamma 2.2 w/ gamma boost 1.12 on my MacBook and I have a reference monitor calibrated to Rec709 gamma 2.2. I've disabled the 'Use Mac display profile for viewers'.

Set my output color space to Rec709 gamma 2.2 and when I export a color managed clip tagging it as 1-1-1 it still is slightly off between what I saw in Resolve and what I see in QuickTime.

More experimenting is required.

That's interesting Steve. How off is it? Does the reference monitor match Resolve or QuickTime... Or neither?


Sent from my iPhone using Tapatalk

If I remember well it won’t match ( I forgotten about this 2.2 special case).
Resolve 2.2 gamma preset doesn’t encode files with 1.96 gamma, which is required for ref mode with boost to operate properly.

Apple color sync doesn’t understand 1-4-1 tag unfortunately ( unless it changed).

Solution may be use of reference mode which is not based on BT.1886 preset, but pure 2.2 gamma one. Not sure if 2.2 in Resolve is perfect curve or it has some linear bit as well. If so match won’t be 100%, but quite close.

So ultimately this is all just a nice to have now. We can get the Resolve display and a reference monitor to match for both 2.2 and 2.4 gamma and providing the tags are correct on export (1-1-1 for 2.4 and 1-4-1 for 2.2) then if you import the video into your project and it matches the timeline then you can be sure everything is golden right?


Sent from my iPhone using Tapatalk

2.4 works.
2.2 is made up and there are no proper specs for it as far as I know.
1-4-1 won’t work as color sync doesn’t understand it.
Resolve needs to export files for 2.2 gamma with 1.96 encoding in order for this to work with Bt.1886 +1.12 boost. Not sure why it doesn’t use 1.96 encoding for 2.2 gamma ( is it because it’s made up value)?

Doesn’t 1.12 boost definitely work?

When I think about it, this is probably to be expected. There will be a slight shift because the output color space was actually gamma 2.2 but I'm tagging it as if it were gamma 1.96, I guess. I'll need to think about this last point a bit but I think the bottom line is that when I export gamma 2.2 tagged 1-1-1 there is a tradeoff I"m making such that this file should playback closer to what I see in Resolve across most players. Think of this last statement as notes to myself until I can do a bit more testing. Cheers.

You are tagging 1.96, because file actually is (should be) 1.96 encoded.
You are getting 2.2 final decoding gamma due to 1.12 boost.
BT.1886 spec expects 1.96 encoded file, then decodes it to final preview gamma which takes boost into account. 1.96*1.12 gives about 2.2 final preview gamma.
Problem is elsewhere. Resolve 2.2 preset seems to not exporting files with 1.96 gamma, but actual 2.2, which is not what we need in order to work with BT.1886 spec.
Whole confusion is based on fact that Rec.709 encoding/decoding is not 1:1. Decoding uses stronger gamma, so it’s always final preview stage which dictates what you’re getting: 2.2 or 2.4 or maybe just 1.96 ( for 1:1 pipe) which is Apple way without ref mode.

As of today 2.2 is not really a correct value, so it shouldn’t be used for serious projects anyway. For web you rather should use sRGB, but crappy YT won’t preserve it, so it will all get messed up.

If you are pedantic and have new Apple screen you should use ref mode and then can specify boost which works for you ( to get final 2.4 or 2.2 gamma for brighter watching conditions).

If you have reference screen then set project as for 2.4 work, but monitor it at 2.2. Exported file will have 1.96 encoding anyway, then ref mode with 1.12 boost will give you match to reference screen. Actually Resolve preview should also be correct because file will be 1.96 encoded and displayed with 1.12 boost. You just need to have “Use display profiles…” turned off to stop Resolve messing with the preview. Should work.

To simplify everything (only for Macs with reference modes).

Always use 2.4 gamma export preset in Resolve.
Disable “Use display profiles…” in Resolve settings.

Then use default BT.1886 ref mode to grade for 2.4 gamma ( set reference screen to 2.4 as well if you have it).

If you want to grade to 2.2 create custom ref mode based on BT.1886 with 1.12 gamma boost ( and set reference screen to 2.2 if you have it). Needs validation to be 100% sure.

Thanks Andrew for all of your suggestions. I'm working in gamma 2.2 so I'll have to sort through this. Here is an article that talks about using the gamma 2.2 method to split the difference to get the closes between different platforms. Nothing is perfect.

https://www.thepostprocess.com/2020/07/17/color-on-mac-displays-from-davinci-resolve-to-the-internet-with-quicktime-tags/

If you want something custom for general web audience then you can grade to 2.1 gamma, which will be 2.4 workflow with BT.1886 ref mode boost at 1.07 (no reference screen).
PCs will decode it with sRGB/2.2 and Mac with 1.96, so 2.1 will be in the middle.