HDR Discussion Thread

[puddy77]

I was just reading this HDR primer by Technicolor's Josh Limor from ICG Magazine.

One thing stood out to me in the HLG section:

This curve, which has been supported by the BBC and NHK, is a “relative transfer function” that is designed for HDR content. This ostensibly allows content to be sent without the need for metadata. (This point is disputed by some critics.) HLG is designed to allow for a greater use of code values with the assumption that most images may only be graded up to 1,000 or 2,000 nits. This assumption allows code values to be better distributed throughout the bits available in the signal.

Can anyone explain the bolded? Is he saying that HLG needs metadata? Or is it just a criticism that HLG might not be as good as other forms of HDR that use metadata?

[zmarty]

At NAB, ASC Looks at HDR and Preserving the Cinematographer’s Vision

http://www.studiodaily.com/2017/04/asc-panel-hdr/

[Penton-Man]

Robert, you’ve got competition….(today)…..
https://www.instagram.com/p/BTZiYnMjPDX/

[Penton-Man]

Originally Posted by puddy77

Can anyone explain the bolded? Is he saying that HLG needs metadata? Or is it just a criticism that HLG might not be as good as other forms of HDR that use metadata?

Puddy, I don't have time to read the article. For end to end HLG systems no metadata is needed. But for PQ -> HLG transcoding, metadata (mastering display max. luminance) or MaxCLL is needed for best result.

[Penton-Man]

Originally Posted by Richard Paul

The person who runs the Ultra4K twitter has posted several negative comments about HDR10+.

And yet, some folks who have real jobs (provided the capability to parse a JSON file containing metadata generated by Samsung's tone mapping software, thusly facilitating HDR10+ with HEVC) were smiling while at the show….

[Penton-Man]

Originally Posted by zmarty

At NAB, ASC Looks at HDR and Preserving the Cinematographer’s Vision
http://www.studiodaily.com/2017/04/asc-panel-hdr/

with regards to HDR acquisition, also for those who might have missed this free presentation during NAB, at the humble grass roots level, for Instant HDR shooters (who own or plan to own the FS5 camera), description of the upcoming HDR Picture Profile …..https://www.youtube.com/watch?v=S11OGKGaN6k#t=19m

[Penton-Man]

Originally Posted by zmarty

Cinematographer’s Vision

For the more advanced (when shooting with log), from the same presenation linked above, were some pearls about proper exposure, because good exposure is particularly important for HDR ….https://www.youtube.com/watch?v=S11OGKGaN6k#t=45m35s

[ray0414]

Originally Posted by puddy77

I was just reading this HDR primer by Technicolor's Josh Limor from ICG Magazine.

One thing stood out to me in the HLG section:

Can anyone explain the bolded? Is he saying that HLG needs metadata? Or is it just a criticism that HLG might not be as good as other forms of HDR that use metadata?

Some people are claiming that HLG is very easy to use for broadcast because it has no Metadata, others are saying it won't work how they think it will work. I've read this a couple places, that over in Europe some were skeptical about HLG and were looking for a more viable alternative.

[Richard Paul]

Originally Posted by puddy77

I was just reading this HDR primer by Technicolor's Josh Limor from ICG Magazine.

One thing stood out to me in the HLG section:

Can anyone explain the bolded? Is he saying that HLG needs metadata? Or is it just a criticism that HLG might not be as good as other forms of HDR that use metadata?

From what I have seen most of the criticism is coming from companies that are trying to sell proprietary HDR formats that could potentially be worth hundreds of millions of dollars. Simply having HLG as an option would cause them problems since broadcasters could use it and that could potentially force them to lower their royalties. HLG is a lose-lose situation for companies that sell proprietary HDR formats but a win-win situation for consumers. That doesn't mean that it doesn't have issues but there is marketing involved in this.

[StingingVelvet]

Originally Posted by ray0414

Picked up the ks8000? Now you have to rewatch all your hdr movies lol completely different look for sure! BTW, lowering the back light a few notches doesn't significantly lower the nits, I measured it with my meter on my ks9800.

It's a complete revelation and an entirely different experience, to the point where I would say people with KU6300 level televisions shouldn't even bother with UHD. They ain't really seeing it. However that's the one and only time I will say that because I don't want people feeling like I am shitting on their televisions.

[Penton-Man]

Originally Posted by ray0414

Some people are claiming that HLG is very easy to use for broadcast because it has no Metadata, others are saying it won't work how they think it will work. I've read this a couple places, that over in Europe some were skeptical about HLG and were looking for a more viable alternative.

Viability (ready feasibility) is actually one of HLG’s attributes; on the other hand, critics tend to condemn it as being possibly weak in terms of lack of future-proof-ness. but that notion is based upon assumption/speculation as to if and when of things to come.

If not profoundly PQ-biased to begin with, somebody you’ve read is in need of some in-service training wrt the production workflow of HLG, as the SMPTE Post NAB event for all groups (students, non-members and members) has now sold out since the heads-up...https://forum.blu-ray.com/showthread...s#post13501127

[Penton-Man]

Originally Posted by Penton-Man

Puddy, I don't have time to read the article. For end to end HLG systems no metadata is needed. But for PQ -> HLG transcoding, metadata (mastering display max. luminance) or MaxCLL is needed for best result.

Perhaps for better understanding, the thinking depicted in flow chart form….

[HeatEquation]

Any more news about HDR10+ coming out of NAB?

[Penton-Man]

Question:
Do we all see (with our human visual system) vis-à-vis constant or non constant luminance?….as described 4 years ago back in 2013 when we were discussing the parameters of BT. 2020.

[Penton-Man]

Originally Posted by HeatEquation

Any more news about HDR10+ coming out of NAB?

There was some speculation as to who would be next to adopt HDR10+ (following that of Amazon Web Services - https://aws.amazon.com/digital-media/aws-nab-2017/ )

[Penton-Man]

https://forum.blu-ray.com/showthread...h#post13434815

Originally Posted by Penton-Man

the thinking...

For a limited time only....http://channel.nationalgeographic.com/genius/

[Penton-Man]

Originally Posted by Penton-Man

Question:
Do we all see (with our human visual system) vis-à-vis constant or non constant luminance?….as described 4 years ago back in 2013 when we were discussing the parameters of BT. 2020.

Answer:
Constant luminance

[Penton-Man]

Originally Posted by Richard Paul

From what I have seen most of the criticism is coming from companies that are trying to sell proprietary HDR formats that could potentially be worth hundreds of millions of dollars....

One example of an HDR format agnostic company that doesn't criticize different approaches….http://www.colorfront.com/index.php?...spage=Approval

[DanBa]

An opinion on "Dolby Vision vs HDR10Plus":
https://twitter.com/JohnnyFocal/stat...25949048041476
http://www.imdb.com/name/nm0860338/?ref_=ttfc_fc_cr3

"HDR10+ doesn't have the same level of picture analyse in the prep stage as DV. All the clever bits are in a @dolby patent."
"HDR10+ is like some trying to reinvent the wheel but it cant be round. Its a very inelegant solution."
"But still better than vanilla HDR10 as you say."


Whatever is best.
As the industry is unable to reach a consensus, consumers should require that HDR TV shall support all HDR formats.
Thus, it doesn’t matter what HDR format a given program uses; a TV that implements all formats can accommodate the content and display it to its best advantage.

At the end of the day, it is content that matters most, not content format.
No more early-adopter industry-driven HDR TV incompatible with other HDR content, but consumer-driven universal HDR-compliant TV able to play any HDR content!

[Penton-Man]

Originally Posted by DanBa

An opinion on "Dolby Vision vs HDR10Plus":

"HDR10+ doesn't have the same level of picture analyse in the prep stage as DV. All the clever bits are in a @dolby patent."
"HDR10+ is like some trying to reinvent the wheel but it cant be round. Its a very inelegant solution."
"But still better than vanilla HDR10 as you say."

I find the bolded to be somewhat of a cavalier comment coming from someone who did not serve on the respective SMPTE engineering project standardizing various dynamic metadata approaches. In terms of ‘elegance’, yes, as a total solution DV is technically superior to HDR10+, but the later is an open solution, so apples to oranges.

I don’t see HDR10+ (and it was on display at NAB) as trying to “reinvent the wheel”. The problem/challenge to giving consumers the best HDR picture is due to manufacturing limitations from the get-go, i.e. different mastering and target displays with differing peak luminances which are controlled by power limiting circuits.

Way back when, prior to publication, Samsung strove (see A.4, p. 27 of the Study Group report on HDR, referenced here with an early heads-up -> https://forum.blu-ray.com/showthread...m#post11099711 to improve upon that fundamental discrepancy between various display devices with their devised tone mapping and color saturating mapping solutions using a scene based (rather than data driven (Dolby Vision) approach.