HDR Discussion Thread

[Penton-Man]

Ray, this is your conscience speaking. I sincerely hope that your serious family issue has been resolved, e.g. when my mother was suffering from Stage IV colon cancer, my life basically centered around two things….my job and supporting her and my Dad through that ordeal…until she passed. And if I’d acted otherwise, I’m certain I would regret it to this day.

[Penton-Man]

Originally Posted by bruceames

....I wonder how many scenes that are significantly above 1000 nits even though it is mastered at 4000 nits. It's possible that there is not a single scene above 1000 nits in some of those masters.

Depends on the movie. This is the MaxCLL histogram of a popular feature film showing a tad more than 30% of frames having a frame maximum light level of > 1,000 nits.



P.S.
Though, as an aside, in terms of weakness, the MaxCLL specification of ST2086, which is part of the mastering static metadata requirement is really not useful in determining the diffuse white luminance, ergo a shortcoming.

b.t.w. cool shades…but people like it when you smile. Why so serious?

[Penton-Man]

Originally Posted by DanBa

Dan, I still personally prefer the physical social contact with family and friends.....rather than the isolationism of dem goggles…unless of course physically challenged.

[bruceames]

Originally Posted by Penton-Man

b.t.w. cool shades…but people like it when you smile. Why so serious?

Probably because I was anxious about the 30km trail race I was about to do. Plus I'm a fairly serious guy anyway for the most part and most of my smiles look forced (like Arnold in T2, lol).

[K i r a]

Originally Posted by ray0414

I would say yes. The ks8000 has much more wow to its HDR than the 850c. How anal are you about black levels? You may get some bleed with the ks8000, but you may already have some on your 850c anyways.

I mainly watch my movies in a dark room, so I do like to have nice black levels. Both are edge lit, so it might be same.

On a side note, my local Best buy just got in the Sony Z9D on display and... holy ****ing shit. It makes the X94D look like trash in comparison. White levels, detail, black levels... it's probably the best i have seen from an LED that rivals OLED truly. $5,999 for the 65" version, though.... RIP.

[DanBa]

Originally Posted by ray0414

So let's summarize. Will hdr10 now be dynamic metadata instead of static? This was approved by SMPTE or BDA?

HDR10 (i.e. BDMV HDR) is specified by the BDA.
http://www.blu-raydisc.com/assets/Do...per_150724.pdf
It is characterized by the following properties: ITU-T H.265 HEVC, ITU-T BT.2020, SMPTE ST 2084, SMPTE ST 2086.

According to Thierry Fautier, President of Ultra HD Forum, SMPTE ST 2094 Dynamic Metadata is not part of BDA UHD spec.



If one of the 4 different ST 2094 dynamic metadata methods "from Dolby, Philips, Technicolor, Samsung that are considered sufficiently different to make it impossible to rationalize into a single method" is royalty-free, it should be added into the upcoming VP9-PQ HDR technology of the royalty-free approach Internet world (contrary to the HEVC of a HDR10 system, VP9 is royalty-free).
Like HDR10, the VP9-PQ is an incomplete HDR technology: the dynamic metadata adaptation for consistent visual, one of the 3 fundamental HDR building blocks, is missing.
https://forum.blu-ray.com/showpost.p...&postcount=366
https://forum.blu-ray.com/showpost.p...&postcount=362

Originally Posted by Richard Paul

SMPTE 2094-40 (Samsung) was approved a few days ago so now all of the dynamic metadata systems in SMPTE 2094 have been officially approved.

Is the approved SMPTE 2094-40 royalty-free?

According to the document "JCTVC-X0061 Indication of SMPTE 2094-40 metadata in HEVC" submitted to the Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11, like HEVC, SMPTE 2094-40 is protected by patents:
http://phenix.it-sudparis.eu/jct/doc...ch_sub_group=1

"4 Patent rights declaration(s)
Samsung Electronics Co Ltd. may have current or pending patent rights relating to the technology described in this contribution and, conditioned on reciprocity, is prepared to grant licenses under reasonable and non-discriminatory terms as necessary for implementation of the resulting ITU-T Recommendation | ISO/IEC International Standard (per box 2 of the ITU-T/ITU-R/ISO/IEC patent statement and licensing declaration form)."

By the way, it is exactly the same patent rights declaration than SMPTE 2094-10 issued by Dolby.

[FilmFreakosaurus]

No wonder lay consumers are baffled by all the variances of A/V technology specs. to the point of indifference.

Hell, the industry insiders can't even make up their damn minds and get this all finalized and simplified, which should have been done a couple years ago!

[bruceames]

Originally Posted by ray0414

I'd bet the 940c is better than the 940d, and should at least not clip 1000 nit content like the 940d.

Here's the comparison and time stamp of the sun time stamp scene. You can flip back and forth between the dolby vision version (well defined sun) and the hdr10 version (which clips the sun). The sun clipping will be similar on other tvs.


http://www.hdtvtest.co.uk/news/dolby...1606214303.htm

I bought Pan and those scenes are not clipped on my 940c (which I expected). I took these pics in daylight environment, but it looks like they came out OK anyway. The other scene wasn't clipped either but the sun was surrounded by such brightness in the scene that I would need to wait until dark to get a good pic. Anyway, the scenes to me look just like the DV pics posted on HDTVtest (although I have the color set on 65 instead of 50 so there is more color saturation, but the HDR looks the same)

[Richard Paul]

Originally Posted by ray0414

So let's summarize. Will hdr10 now be dynamic metadata instead of static? This was approved by SMPTE or BDA?

I think that the CEA is working on adding dynamic metadata to HDR10 but they probably won't say anything about it until it is officially announced.

Originally Posted by DanBa

HDR10 (i.e. BDMV HDR) is specified by the BDA.

HDR10 was made by the CEA but it has specs that are very close to BDMV HDR. Two differences are that HDR10 doesn't specify a video format and that it can refer to uncompressed video (such as computers and game consoles).

Originally Posted by DanBa

If one of the 4 different ST 2094 dynamic metadata methods "from Dolby, Philips, Technicolor, Samsung that are considered sufficiently different to make it impossible to rationalize into a single method" is royalty-free, it should be added into the upcoming VP9-PQ HDR technology of the royalty-free approach Internet world (contrary to the HEVC of a HDR10 system, VP9 is royalty-free).

SMPTE 2094-40 was finished less than a week ago and it will take time for it to be used.

Originally Posted by DanBa

Is the approved SMPTE 2094-40 royalty-free?

SMPTE 2094-40 is royalty-free and Samsung has released a patent declaration for it on the SMPTE website for SMPTE 2094-40.

Originally Posted by DanBa

By the way, it is exactly the same patent rights declaration than SMPTE 2094-10 issued by Dolby.

Dolby will require a license/royalty for SMPTE 2094-10 though all of that information is under NDA. I don't have a problem with companies wanting to make money but there is something a bit impressive about Dolby pushing for 4,000 nits on HDR movies and than turning around and selling their dynamic metadata system. They are basically creating a problem and selling a solution to that problem.

[bruceames]

Also I have the 930C. That doesn't get as bright, but I'll take pics of those scenes out of curiosity (I'll have to move the player there, so it may be awhile before I get around to it).

[ray0414]

Originally Posted by bruceames

I bought Pan and those scenes are not clipped on my 940c (which I expected). I took these pics in daylight environment, but it looks like they came out OK anyway. The other scene wasn't clipped either but the sun was surrounded by such brightness in the scene that I would need to wait until dark to get a good pic. Anyway, the scenes to me look just like the DV pics posted on HDTVtest (although I have the color set on 65 instead of 50 so there is more color saturation, but the HDR looks the same)

Its hard to tell if there's clipping without a tv known not to clip to be side by side. But from your sun picture it looks to be clipping. The sun isn't small and defined, it looks a little blown out compared to the picture I took of my ks9800. Your colors look excellent though.

[DanBa]

Originally Posted by Richard Paul

I think that the CEA is working on adding dynamic metadata to HDR10 but they probably won't say anything about it until it is officially announced.

HDR10 was made by the CEA but it has specs that are very close to BDMV HDR. Two differences are that HDR10 doesn't specify a video format and that it can refer to uncompressed video (such as computers and game consoles).

Does this "Dynamic HDR10" affect the BDA UHD spec?

Originally Posted by Richard Paul

Dolby will require a license/royalty for SMPTE 2094-10 though all of that information is under NDA. I don't have a problem with companies wanting to make money but there is something a bit impressive about Dolby pushing for 4,000 nits on HDR movies and than turning around and selling their dynamic metadata system. They are basically creating a problem and selling a solution to that problem.

The 10,000 nit PQ was accepted by the experts, it’s too late to take another way.
The 10,000 nit race was launched, apparently the Sony’s Black Master Drive consumer display is able of outputting 4,000 nits of peak brightness.

Originally Posted by Penton-Man

For example, I offer a historical retrospective, which few are aware of to this day….and certainly no AV journalists or tweet hounds vying to amass as many *likes* or *followers* as they can.

Back in early spring of 2012, Dolby submitted the PQ curve for an EOTF based upon human perception for use with UHDTV to the respective ITU working party (the same group responsible for producing the recently released milestone ITU-R BT.2100). Anyway, during the discussions in that particular spring meeting, committee members expressed much skepticism about the value or need for having a TV capable of higher brightness and deeper black levels. The working party did not accept this objective based proposal (containing measurements, numbers, graphs) partly because there was not an appreciation that increased dynamic range would have value in a new TV system and partly because it was too new (visionary/revolutionary).

Soooo, 6 months later, Dolby provided another input document to the working party, this time which contained and described the subjective test results for viewer preference for white and black light levels. There were 34 study participants (among other things, who were classified into 3 age groups in order to approximate age-related effects of the human visual system <- a detail which I don't think has been mentioned in any of the online SMPTE presentations available to the public).

This later proposal was accepted and voilà….years later, we are where we are today.

By the way, do you know if the SMPTE 2094-10 dynamic metadata proposed by Dolby are exactly the same as Dolby Vision dynamic metadata?
Thanks!

[bruceames]

Originally Posted by ray0414

Its hard to tell if there's clipping without a tv known not to clip to be side by side. But from your sun picture it looks to be clipping. The sun isn't small and defined, it looks a little blown out compared to the picture I took of my ks9800. Your colors look excellent though.

Could be that's it's clipping the sun a little. I'll lower the color setting to 50 and look at it again at night. It might be that the ambient light is washing it out a little. It certainly did on that other sun scene (at least, far as taking the pic was concerned).

[bruceames]

Ray if you get a chance could you take a pic from you TV of that sun shot? The DV pic from the HDTVtest site is pretty small so it's a little hard to see how much more defined it is. I watched the scene again later and the pic I posted is pretty accurate. Lowering the color had no effect on it (except for less color of course).

[ray0414]

Originally Posted by bruceames

Ray if you get a chance could you take a pic from you TV of that sun shot? The DV pic from the HDTVtest site is pretty small so it's a little hard to see how much more defined it is. I watched the scene again later and the pic I posted is pretty accurate. Lowering the color had no effect on it (except for less color of course).

[spwill]

Originally Posted by bruceames

Could be that's it's clipping the sun a little. I'll lower the color setting to 50 and look at it again at night. It might be that the ambient light is washing it out a little. It certainly did on that other sun scene (at least, far as taking the pic was concerned).

Try turning live color on as it increases the clipping point on hdr patterns. I don't have any discs mastered to 4000 nits, but I have seen the difference in highlight details from the sony camp demo(the sun) and the lg oled demos.

[Penton-Man]

Originally Posted by puddy77

I just saw this article from the Washington Post raving about the Dolby Cinema presentation of Kubo and the Two Strings:
Dolby Cinema is the way to see ‘Kubo and the Two Strings’

There’s an in depth article discussing the stop-motion, native stereo aspects, etc. in the just published September issue of American Cinematographer magazine on Kubo and the Two Strings along with some coverage of the Summit which included on one day,
this gathering
^ which is a close up pic of the more wide angle shot at the top of p.80 of the magazine….once viewers got seated.

[Penton-Man]

Originally Posted by DanBa

By the way, do you know if the SMPTE 2094-10 dynamic metadata proposed by Dolby are exactly the same as Dolby Vision dynamic metadata?
Thanks!

I'm pretty sure I know.

[PaulGo]

Hurdles Ahead for 4K, HDR

...So what's holding 4K TV back? Some of the problem is the same chicken-and-egg dilemma that high-definition video faced when it debuted. Many mainstream consumers don't want to buy 4K TVs until there's more UHD content to watch. And content producers don't want to invest large sums of money in UHD video until there's a wider base of viewers ready to tune in. But that problem is solvable.

Full article at:

http://www.lightreading.com/video/4k...a/d-id/725703?

[DanBa]

Originally Posted by Richard Paul

I think the HDR timeline might look like this:

2014: SMPTE 2084 added to HEVC
2015: SMPTE 2084 used in consumer products
2016: HLG added to HEVC
2017: HLG used in consumer products
2018: SMPTE 2094 and ICtCp added to HEVC
2019: SMPTE 2094 and ICtCp used in consumer products

Originally Posted by Richard Paul

I think that SMPTE 2094 will get added to the HEVC draft this year but that the next version of HEVC could get released in 2018. Granted they might release the next version of HEVC in 2017 if there are enough companies pushing for it.