HDR Discussion Thread

[bruceames]

Originally Posted by puddy77

HDR, Resolve, and Creative Grading

A very long and thorough overview and commentary of HDR from the perspective of a colorist.

A nice description!

Simply put, HDR (High Dynamic Range) is an escape from the tiny box, as currently defined by BT.709 (governing color gamut), BT.1886 (governing EOTF), and ST.2080-1 (governing reference white luminance levels), in which colorists and the video signals/images they manipulate have been kept imprisoned for decades.

Haven't read the whole article yet, but that part deserves a quote.

[Geoff D]

Good stuff, although the notion of 10-bit being a panacea to cure 8-bit banding woes isn't quite a reality just yet. There's a reason why matey mentions that 12-bit is used for HDR mastering at the source level (and is preferred outright for Dolby Vision): it's above the Barten threshold for banding as outlined in the SMPTE 'Perceptual Signal Coding' presentation from 2012, whereas 10-bit falls below it.

Incidentally this is why the Digital Bits are barking up the wrong tree when they keep mentioning UHD banding being a result of boiling down the original 12 or 16-bit source capture....because the same thing would happen with a 10-bit source in UHD: it's HDR10 itself and how it allocates them bits across the range which is the issue. Hopefully they can hit on a way to dither the source properly when encoding in HDR10.

And a +1 for his bit about any digital camera with a decent dynamic range being able to offer x amount of post-produced "HDR" capability, I mentioned the same thing here just a couple of days ago and is also the reason why I'd be excited for a Tron Legacy HDR redo.

[Penton-Man]

Originally Posted by Richard Paul

SMPTE 2094-10 (Dolby) and SMPTE 2094-40 (Samsung) have proposals for the upcoming HEVC meeting.

Pictorial illustration as to what this means with regards to post houses unwilling and/or unable ($) to pay for a Dolby Vision Color Grading system….

Originally Posted by Penton-Man

and……on the far right, is the CMU working or,what some call, da brains to making it all happen....

to enable workflows such as this –



is that there will be non-proprietary toolset options based upon Applications in the SMPTE 2094 suite to enable dynamic metadata tone mapping using any high quality HDR reference display (ex. Sony BVM -X300) without the Dolby CMU (or Pulsar) required to produce the masters.

[Penton-Man]

Originally Posted by puddy77

HDR, Resolve, and Creative Grading

A very long and thorough overview and commentary of HDR from the perspective of a colorist.

“You just have to know that whatever peak highlights you have in the frame will be preserved, and have the potential to venture into super-bright levels, so you should start planning your highlights within the image accordingly. I’m guessing DPs….”

To that ^, if one were to attend ASC Master Classes like this ….
http://www.theasc.com/site/news/asc-...ion-concludes/ in the future, attendees may learn thru Hollywood colorists that with regards to shooting even in RAW, DPs shouldn’t get a false sense of security with their highlights and as to what they’ll ultimately get when in post production color correction sessions for the HDR grade because DPs are all using LUTs with their dailies which are based on curves which are rolling the highlights over in a different fashion than what color correction does….they roll them over softer and further down than in the color correction post process, before the video signal starts folding over on itself.

[Geoff D]

Yup, most productions these days use a LUT as a quick 'one light' kind of timing for the dailies to get an idea of what it'll look like, sometimes sending a still frame with the relevant corrections to the DIT who can then lay that look over an entire shot before feeding it back to the production.

And to complicate things further it was interesting to hear from the director and DP of Deadpool re: the HDR grade, how it wasn't until the final pass for the home UHD version that they actually got to see all that extra highlight detail in action!

[Penton-Man]

Gamer dudes…https://developer.nvidia.com/preparing-real-hdr

https://developer.nvidia.com/impleme...se-tomb-raider

This is an introductory level course coming up in Anaheim at Siggraph - http://s2016.siggraph.org/courses/ev...cal-challenges . Book early for best savings on hotels (remember, this is Disneyland territory folks…summer tourists, etc.).

Longtime 4K Movies Releases thread readers should already be somewhat familiar with Sci-Tech Academy award winner Josh …..I personally don’t know Tania, other than the quality of her peer reviewed journal publications. Both are TOP drawer.

[Penton-Man]

Originally Posted by Penton-Man

post houses unwilling and/or unable ($) to pay for a Dolby Vision Color Grading system….

another pic will just have to suffice….

[Richard Paul]

SMPTE 2094-30 (Technicolor) has a proposal for the upcoming HEVC meeting and it looks like Philips (who made SMPTE 2094-20) is now supporting it. I remember reading earlier this year that Philips and Technicolor were going to combine their HDR systems.

Originally Posted by Geoff D

Incidentally this is why the Digital Bits are barking up the wrong tree when they keep mentioning UHD banding being a result of boiling down the original 12 or 16-bit source capture....because the same thing would happen with a 10-bit source in UHD: it's HDR10 itself and how it allocates them bits across the range which is the issue. Hopefully they can hit on a way to dither the source properly when encoding in HDR10.

The more dither/noise that is added to compressed video the higher the bit rate goes so dither is a trade-off. The best solution is increasing the bit depth but that would cost more and most people seem to be okay with a bit of banding.

[Geoff D]

Okay, forget the specific use of 'dithering' (I was using that as more of a 'catch all' term, I should know better ) but can't they try what Sony does with their SBMV on regular Blu-ray? After all, Sony aren't overly taxing the bitrate by adding noise, they're overdriving the bit depth before laying it back down onto 8-bit with no cost to the compressed bitrate and it works brilliantly.

Yeah, these specific HDR masters are reportedly 12-bit at source but that doesn't seem to be enough of an overshoot when encoding in 10-bit PQ, IMO it's telling that one of the few UHD movies I've seen absolutely zero signs of banding on (ASM2) was originally finished at 16-bit.

[Penton-Man]

Originally Posted by Penton-Man

Longtime 4K Movies Releases thread readers should already be somewhat familiar with Sci-Tech Academy award winner Josh …..

who has also presented at the HPA Tech Retreat in the past - https://forum.blu-ray.com/showthread...h#post10417134

[Penton-Man]

Originally Posted by Penton-Man

Although as expressed last March, I’m more of a proponent of ST 2094 Application #1, as previously noted –
https://forum.blu-ray.com/showthread...4#post12007747

Here is a visual illustration for Samsung’s 2094 Application regarding the value of dynamic metadata as covered by a member of the press corps at the last NAB….

https://vimeo.com/164619204

As readers were (exclusively) alerted to the work of SMPTE Technical Committee 10E SG on HDR Ecosystem last year (https://forum.blu-ray.com/showthread...e#post11444185)
in response to the call for one page contributions from HDR proponents, the original Samsung proposal is on p.27 of the completed SMPTE report.

[Penton-Man]

Originally Posted by Penton-Man

As readers were (exclusively) alerted to the work of SMPTE Technical Committee 10E SG on HDR Ecosystem last year (https://forum.blu-ray.com/showthread...e#post11444185)
in response to the call for one page contributions from HDR proponents, the original Samsung proposal is on p.27 of the completed SMPTE report.

For those that don’t want to deal with taking the survey to get the report, here you go - https://www.smpte.org/sites/default/...-Ecosystem.pdf

[Penton-Man]

Originally Posted by puddy77

HDR, Resolve, and Creative Grading

“Monitoring an HLG image is as simple as getting a Hybrid Log-Gamma-compatible HDR display”….

not so 'simple', prior to this - https://forum.blu-ray.com/showthread...g#post12188250 but some content creators (such as those viewing soccer in post #133 on the last page) and these did - http://www.dbwcommunication.it/?p=441&lang=en

[Penton-Man]

Inquiring minds might ask, in the LA environs in addition to Dolby’s place on Hollywood Blvd. (which if readers recall, can serve both as a grading facility as well as a special invite exhibition facility, e.g. https://forum.blu-ray.com/showthread...e#post11569063 )

where might be located those elusive Dolby Vision capable grading facilities rocking Dolby Vision projectors…..well…..Deluxe, Technicolor, Pixar and Disney have ‘em.

[Penton-Man]

Originally Posted by Penton-Man

When I’m at liberty, I do what I can to release information before it becomes ‘official’ and other websites/forums do, e.g. https://forum.blu-ray.com/showthread...d#post12039620

Similarly, with regards to worldwide standards/solutions for HDR implementation, knowledgably discussing the lead up to the DNR ITU-R BT.[HDR-TV] with the solid membership here….. https://forum.blu-ray.com/showthread...l#post11839356

Now that the broadcast ball is rolling, a little past history for those not having qualified access to the respective ITU Working Party. For approximately the past 3 years there had been disagreement regarding the type of HDR system(s) for ITU recommendation. Three systems were submitted…. two PQ systems (USA and Philips), and the HLG system (BBC/NHK). In Feb., ITU-R BT.[HDR-TV] agreed upon a total of 2 systems….. the USA(Dolby) PQ system (defined by SMPTE ST 2084, ushered in by Dolby) and the HLG system (BBC/NHK).

As for movies (feature films), Sony coming out with a 55” 4K HDR OLED client monitor is a significant milestone (not to be underestimated), again to aid in HDR adoption.

In regards to the last paragraph above ^, a follow-up pictorial illustration showing Quad view comparison of S-log3, PQ (ST 2084) HDR, HLG HDR and Rec. 709 (SDR)…

[Penton-Man]

Regarding this launch > http://www.ses.com/4233325/news/2016/22275782

Originally Posted by Penton-Man

I have a special interest in determining whether this Identical suggestion in the header -



will hold true for all or some (i.e. content dependent) live productions…

and also how they compare in a non-reference viewing environment since some would say that HLG holds a theoretical upper hand over ST 2084 based approaches (HDR10 and Dolby Vision) in being a scene referred system.

[Penton-Man]

Originally Posted by Penton-Man

HLG

[Penton-Man]

Both HDR approaches (PQ and HLG) from draft ( http://www.itu.int/md/R15-SG06-C-0039) to achieving international ratification, planned in July….if all goes well.

[Richard Paul]

The HEVC meeting just ended and not much happened. The SMPTE 2094 proposals were rejected and I guess they won't be added to HEVC until they are published by SMPTE.

Originally Posted by Penton-Man

Both HDR approaches (PQ and HLG) from draft ( http://www.itu.int/md/R15-SG06-C-0039) to achieving international ratification, planned in July….if all goes well.

That is a nice surprise and things seem to be going well in the ITU.

[bailey1987]

Was there anything in SMPTE 2094 that will be missed?