Content Modifier - Barco HDR

[SteveLLamb]

Recent CPL created for Barco HDR Light Steering, DCI HDR doesn't fit, as per JZ: "Yes, the difference to other HDR standards is that we are going for PQ 300nits and Rec2020... I agree HDR1 is too generic".

For this project we opted for using the CTT Code: BarcoHDR
And extention:

<cpl-meta:ExtensionMetadata scope=http://www.dcimovies.com/schemas/2018/HDR-Metadata>
  <cpl-meta:Name>Image Encoding Parameters</cpl-meta:Name>
  <cpl-meta:PropertyList>
    <cpl-meta:Property>
      <cpl-meta:Name>EOTF</cpl-meta:Name>
      <cpl-meta:Value>ST 2084</cpl-meta:Value>
    </cpl-meta:Property>
    <cpl-meta:Property>
      <cpl-meta:Name>Color Processing</cpl-meta:Name>
      <cpl-meta:Value>Barco</cpl-meta:Value>
    </cpl-meta:Property>
  </cpl-meta:PropertyList>
</cpl-meta:ExtensionMetadata>

Open to other suggestions, perhaps HDR2. For discussion.

[palemieux]

The scope attribute should use a Barco-specified URI since this is presumably different than DCI HDR

[radford-for-smpte]

This is interesting, since it does use PQ and it is constrained to 300nit.

Interestingly, the other constraints it has that are specific to Barco HDR (limited APL, etc.) would not actually be relevant to an HDR1 capable system. They are however important to know for a Barco HDR system (so that it would presumably refuse to play DCPs lacking such constraints).

But related to what Pierre wrote, the Color Processing field should not be under the DCI namespace.

So in general, I think representing the content as precisely HDR1 is correct since it's a) technically indistinguishable, and b) would look 100% correct on an HDR1 system. But it should have non-DCI metadata indicating the Barco-specific constraints.

[SteveLLamb]

@jackers2024 please identify yourself in your profile with either a name or company, so we know who is contributing. We also require non-anonymous users for all PRs as well. Also please keep the discussions on issues to the current topic. While we don't mind the questions regarding the DCI HDR specifically, that should be it's own thread, as this was about Barco's HDR and relation to DCI HDR. I'll move this to a new topic.

[radford-for-smpte]

I'd like to sort this out ASAP. My (counter) proposal would be:

<ExtensionMetadataList>
  <ExtensionMetadata scope=http://www.dcimovies.com/schemas/2018/HDR-Metadata>
    <Name>Image Encoding Parameters</Name>
    <PropertyList>
      <Property>
        <Name>EOTF</Name>
        <Value>ST 2084</Value>
      <Property>
    </PropertyList>
  </ExtensionMetadata>
  <ExtensionMetadata scope=http://www.barco.com/schemas/2024/Light-Steering>
    <Name>Image Constraints</Name>
    <PropertyList>
      <Property>
        <Name>Light Budget</Name>
        <Value>true</Value>
      <Property>
    </PropertyList>
  </ExtensionMetadata>
</ExtensionMetadataList>

[palemieux]

@radford-for-smpte We need Barco folks to be involved in this discussion.

[radford-for-smpte]

Agreed. Are they excluded?

[palemieux]

Not AFAIK. Someone should alert them in case they missed this thread.

[SteveLLamb]

I'm good with those proposed extensions. They are not on this repo, I will reach out directly to them and report back.

[palemieux]

@SteveLLamb They need to join the repo and respond to this thread.

[jamiegau]

Hi Guys,
I have been meaning to bring this topic up at ISDCF but was giving it time to cook.

HDR is a development in Cinema that needs addressing.
It would appear, from general marketing I see about better options than typical 48nit experiences, that it's splintering into numerous proprietary versions.

Considering the DCP's purpose was to be a single master that can target all cinemas. This vision appears to be breaking down, somewhat, in the path we are currently taking with HDR. Deliverables will explode in number, generating a lot of costs.

I was of the understanding, DCI created the 300nit specification so that vendors could then develop projectors that met those requirements. Turning it back around to producing a single master that targeted specific min/max light levels, specific primaries and lookup table. Allowing HDR masters of generic nature to be produced, like we have with typical Cinema today, and vendors making projectors that met those requirements. Currently, HDR in cinema appears to be the complete opposite.

Due to this, it is extremely important that a vendor has been able to achieve the HDR-DCI-specification. Indicating, this line in the sand has been met and its position crystallised. It's not too high. And vendors are now encouraged to meet, but not necessarily exceed it (to keep costs in a reasonable affordable band).

Barco have been indicating to me that the light steering projector meets the DCI-HDR-Spec. Which would be fantastic, as it's a major milestone in making all this a reality. However, I see there is resistance from studios on this. I have been patiently waiting to hear from the studios on a official position on this issue. (Last ISDCF, the topic was side stepped.) But it is taking considerable time.

The thread above does, based on my assumptions (As suggested that the tag "Light Budget" be applied to Barco masters), appear to indicate that, because the Barco product cannot do 300nit full screen, and is only a percentage of the screen at 300nit, that it does not fully pass the line in the sand.

I would ask for clarity on this issue, as it's an important development in the industry and will significantly change the narrative and efforts behind HDR in cinema. I understand this is a extremely sensitive issue as it impacts on market forces. However, we do need movement on this issue in cinema. It's taking too long. The longer it takes, the longer it may take to get back into a 1 HDR grade for all projectors world. (Could make the Interop -> SMPTE DCP transition look fast and easy comparatively.)

My position on this is that the use of light budget approach is only common sense considering what we are trying to do here.
And the DCI-HDR-Spec should be updated to encompass this method. It's been in use in domestic monitors from the beginning. (Due to power limitation and physical limits in the technology, which also applies here.) And as the average brightness even on a HDR cinema presentation is still around the 48nit level, leaving the extra headroom for specular highlights and bright light sources, in general, a budget for those small percentages of the overall image means the budget is rarely challenged. I am only a back seat observer, so my opinion does not hold a lot of weight. I am not siting in grading rooms with HDR content on a daily bases. The studios would be the best position to make this call.

I hope to be at the ISDCF meeting in a few days to hear how this is all going.
James

[jackers2024]

Amen. That’s all I can say.

[radford-for-smpte]

James, Thanks for your input on this. A few comments on your post: appear to indicate that, because the Barco product cannot do 300nit full screen, and is only a percentage of the screen at 300nit, that it does not fully pass the line in the sand. While that might be the case, the question on the table (specifically in the github thread) is about the nature of the content rather than the capabilities of the device. My comments and suggestions are strictly about how to identify and annotate light-steering DCPs. From a content perspective, a light-steering DCP is 100% equivalent to the HDR DCP defined in DCI's HDR Addendum. It uses the EOTF defined in ST 2084 (PQ), it utilizes the DCI HDR color-volume, it carries luminance values from 0.005-300 nits, etc. In addition to that, the content happens to have a constraint of sorts on APL on a per-frame basis. As you point out, this is nothing new in the HDR world as consumer devices have always had a variance between peak and full-screen capabilities. The degree to which the Barco LS projector deviates from the HDR Reference Display is also an interesting topic but, IMHO, one that's divorced from the discussion of content metadata.

…

-Mike

________________________________ From: James Gardiner ***@***.***> Sent: Monday, September 30, 2024 4:38 PM To: ISDCF/registries ***@***.***> Cc: Radford, Mike ***@***.***>; Mention ***@***.***> Subject: Re: [ISDCF/registries] Content Modifier - Barco HDR (Issue #941) This Message is From an External Sender Caution: Do not click links or open attachments unless you recognize the sender and know the content is safe. Hi Guys, I have been meaning to bring this topic up at ISDCF but was giving it time to cook. HDR is a development in Cinema that needs addressing. It would appear, from general marketing I see about better options than typical 48nit experiences, that it's splintering into numerous proprietary versions. Considering the DCP's purpose was to be a single master that can target all cinemas. This vision appears to be breaking down, somewhat, in the path we are currently taking with HDR. Deliverables will explode in number, generating a lot of costs. I was of the understanding, DCI created the 300nit specification so that vendors could then develop projectors that met those requirements. Turning it back around to producing a single master that targeted specific min/max light levels, specific primaries and lookup table. Allowing HDR masters of generic nature to be produced, like we have with typical Cinema today, and vendors making projectors that met those requirements. Currently, HDR in cinema appears to be the complete opposite. Due to this, it is extremely important that a vendor has been able to achieve the HDR-DCI-specification. Indicating, this line in the sand has been met and its position crystallised. It's not too high. And vendors are now encouraged to meet, but not necessarily exceed it (to keep costs in a reasonable affordable band). Barco have been indicating to me that the light steering projector meets the DCI-HDR-Spec. Which would be fantastic, as it's a major milestone in making all this a reality. However, I see there is resistance from studios on this. I have been patiently waiting to hear from the studios on a official position on this issue. (Last ISDCF, the topic was side stepped.) But it is taking considerable time. The thread above does, based on my assumptions (As suggested that the tag "Light Budget" be applied to Barco masters), appear to indicate that, because the Barco product cannot do 300nit full screen, and is only a percentage of the screen at 300nit, that it does not fully pass the line in the sand. I would ask for clarity on this issue, as it's an important development in the industry and will significantly change the narrative and efforts behind HDR in cinema. I understand this is a extremely sensitive issue as it impacts on market forces. However, we do need movement on this issue in cinema. It's taking too long. The longer it takes, the longer it may take to get back into a 1 HDR grade for all projectors world. (Could make the Interop -> SMPTE DCP transition look fast and easy comparatively.) My position on this is that the use of light budget approach is only common sense considering what we are trying to do here. And the DCI-HDR-Spec should be updated to encompass this method. It's been in use in domestic monitors from the beginning. (Due to power limitation and physical limits in the technology, which also applies here.) And as the average brightness even on a HDR cinema presentation is still around the 48nit level, leaving the extra headroom for specular highlights and bright light sources, in general, a budget for those small percentages of the overall image means the budget is rarely challenged. I am only a back seat observer, so my opinion does not hold a lot of weight. I am not siting in grading rooms with HDR content on a daily bases. The studios would be the best position to make this call. I hope to be at the ISDCF meeting in a few days to hear how this is all going. James — Reply to this email directly, view it on GitHub<#941 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ACVYF6HO5XZGGTQPH5SPQ2TZZHHHFAVCNFSM6AAAAABM2EXJAKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGOBUGQ2TCMBXG4>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[jamiegau]

Hi Mike,
Yes I understand these are two different topics. But from my standpoint they "could" be related. So, I was just looking for clarity.

If the content needs a metadata reference that describes it as Barco light steering, it suggests that the Barco implementation is not considered to fulfill the generic specification of the DCI's HDR Addendum.
Is the tag an informative requirement or a restrictive playback tag indicating only BarcoLSP should only ever utilise this content. It was not made clear.

Is a projection system expected to treat the DCP differently than if the metadata was not present?

Your comment
`From a content perspective, a light-steering DCP is 100% equivalent to the HDR DCP defined in DCI's HDR Addendum. It uses the EOTF defined in ST 2084 (PQ), it utilizes the DCI HDR color-volume, it carries luminance values from 0.005-300 nits, etc.'
suggests it is informative, but then again, I know these issues can be very nuanced.

So the crux of my comments is, do DCI related parties consider a master created against a BarcoLSP, acceptable as a new form of generic HDR-DCP creation? A master that can then be shown on any vendor projector, if capable of the specifications mentioned in your paragraph (ie. the DCI's HDR Addendum?)

Or is this still too early to make that commitment? Do other vendors need to meet these specifications, likely with different technology pathways, resulting in a deviance of producer's intent not yet understood?

[BertrandRuwet]

Hello Guys,
This is Bertrand from Barco (PM Cinema servers)

Thanks for your comments and suggestions.
We had a round of discussions and our output and suggestions are the following;

• we confirm and agree to use "BarcoHDR" in the CTT.
• We suggest to use the name "Lightsteering" or simply "LS" instead of "Light Budget" in the secondary descriptor

Have a good day everyone.

[SteveLLamb]

Thanks Bertrand, I'll put in "Lightsteering", and we can continue the discussion about the values as needed via the PR when I put it in.

Locking this issue for further discussion to keep it on topic, as Mike says, is only about metadata identification. We can continue the related topics elsewhere.

[palemieux]

Barco also needs to allocate (and publish) a URL for the scope attribute, e.g., https://www.barco.com/schemas/2024/Light-Steering

[radford-for-smpte]

@BertrandRuwet here is a sample CPL I've made just as an example. Please let me know if this appears consistent with your expectations.

cpl.xml.zip