Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Interesting article on HDR

Yes kinda of,

HDR articles are popping out all over the place, in all actuality they have been since several years ago... Just very quietly, and now with t strong push by some great companies things are changing fast...

The good, we have been shooting HDR for the better part of the past 4 years via HDRx without the need to have a Beam Splitter set at absolute zero Parallax in order to get two exact matches pixel images to edit on different exposures, yet HDRx is limited to 6 stops, with generally only 3 very good, and 4-5 usable...

I have been in touch with few of this new bigger manufacturers and industry leaders in the works for some HDR material, and discussing what the some what similarity of 3D big fuck ups must be avoided here...

Yet I see some of the same repeating itself...

If the standard is not set VERY soon, it will blow out of proportion, with TV/Monitor manufacturers CHEATING again and again to the buyers, saying their sets are Compatible or Ready or what ever BS, when in all actuality HDR TRUE viewing is not as simple as upgrading a TV/Monitor software, but it needs LUMINANCE/LUMINOSITY which comes at the cost of how many "nits" the TV/monitor in question has, and also what type they are, but mostly if they are simply trying to cheat out their customers making them think that their 400 nits topping TV/Monitors will serve them well against the minimum suggested 700/800 & 1000 nits then here we go again...

3D sucked and continues to suck big because of the incapacity to have set a true standard form the start both tat home and in Movie Tethers... which is a far bigger issues then poor Interocular Distance and Parallax setting done by an inexperienced Stenographer or the lack thereof to have hired one by a Director that things He can do without...

I love HDR, and once said to Jim after the announcement of it on the RED EPIC that I would make an entire movie on HDR to have an even better fill then if done in 3D...

My prevision of that is about to show it self...

I truly just hope that it works out well for every one... For us filmmakers is a great opportunity, an other one to have yet an other fantastic tool in our arsenal...

Dolby Vision is my best bet... And my support goes to them... They Dolby Cinema with both Dolby Atmos and Dolby Vision is simply superb...

The Dolby Vision tested and approved Vizio TV's which are a far cry from the 4000 nits one tested, will give a great starting point... The 65 affordably so, the 120 not so much... :)
 
The industry won't settle on a standard, I have no hopes for that since the entire history of film and television is filled with compromises and temporary solutions that got set in stone. Most of the settings we use today with gear capable of much more is the result of old decisions kept by lesser knowing people or lack of visionary ideas. Laziness and profit keeps the industry from being simple and easy.

As for HDR, Ketch, isn't HDR also gonna be able to be applied even on footage that isn't shot in HDRx? I mean, there's a level where an image starts to lack all contrast and dynamic if "everything" is seen, so wouldn't HDR even be applied to creative color grading shots with a lot of contrast, but enhancing the DR of that contrast in order to closer resemble what the human eye is capable of seeing?
 
Unfortunately in HDR as we have seen in 3D, there is already in preliminary tests and in some actually big tests done by some big film companies and manufacturers, we are seen the same thing repeat itself, instead of actually shooting originally acquired HDR imagery, they are processing re-processing older non HDR material, and while in 3D they at last are making some better progress in Post 3D, here they at list have the same image to simply replicate, map out re-arrange within certain parameters and giving it a desired parallax, but this comes form existing material and information...

With HDR what hasn't been captured is not there, if the blacks where not exposed then you simply don't have any info there and you'll be just illuminating a dark area to create noise to then need to eliminate via noise reduction process, opposite ends at the whites, if you din't capture the highlight or burned out that info, then you got absolutely nothing there to protect and show for...

But yes, with proper re-processing and showed on a high luminance (actual nits) and properly contrast adjusted monitor, you can get something more out of standard imagery via a better TV/Monitor by simply allowing for better pixel showing... But is much more complex then this, and original HDR material will have that much more DR to show on this new TV/Monitors and Dolby Vision cinemas...

Then there is also the question of where this will be headed, as I have been doing a series of complex test (very Quietly) shooting some samples with various settings in HDRx with the 2 files x folder with double data acquisition and half the frame rate allowance in slo motion, and then also doing the same shots with a Beam Splitter 3D rig with which I have plenty of knowledge of, and now getting two completely separate yet identically framed images with different settings but same lens aperture, as to create to images that when composed will give a much higher HDR then you ever could via HDRx.

Then also shooting double cams, both in HDRx at different settings, so to now have 4 files in two folders with 4 separate exposures controlled in 2 lines, then companied in a single one for a truly High HDR file... Which this then get showed on to a HDR monitor for grading and then on a HDR TV for viewing...

The good is that RED now has the REDCAST module that can allow to vie HDR and having 4 outputs then makes things great... But with Weapon, and the not compatibility of the REDCAST we now are hopping to see something soon come out on this end.

Fact is that unless you are shooting HDR and viewing it on a HDR monitor proper exposure will not be possible, and you'll always be guessing the stops of HDRx you'll be using, which is good but not ideal, and for this I love the REDCAST.

Oh, and the double shots graded at different ISO to increase the combined DR is simply not good either unless done within it's limitations, where dies off when the information is no longer there, limiting how far you can push it left or right in your Curve as well.


One thing I can say is that I absolute LOVE where this is finally headed, vs 3D, NO GLASSES, but also even so I was still to use 3D rigs, you won't need a Sterographer any more, as far as you can make sure you know how to set an absolute Zero Parallax... Leaving some other issues to deal with of course as that of insuring proper calibrated lenses as in 3D and perfect focus alinement as with 3D.

But...

Here again, look Cinemas and TV manufacturers fuck things up with double standards and not giving the right equipment and or proper lighting power... :)
 
Good timing, i'm trying to catch up a little on shooting/exposing with HDR in mind. Using HDRx for this makes sense (and sets RED ahead of the competition again).

I think Ketch should write an article, hehe. Nice to read your thoughts on this. Please keep em coming. ;)
 
Ketch, I get your point and it's a shame that the industry isn't doing it right, but I was refering to the things we record right now. If I shoot without HDRx on and expose the scene without crushing blacks or whites, I will surely shoot a lot more information than a usual screen can show off, meaning that a camera like the Dragon can without HDRx shoot for HDR delivery, right?
 
Just to clarify, you don't need to use HDRx for HDR playback. HDRx is to capture more dynamic range of a given scene with the camera. w/o HDRx, the camera captures way more than 8-bit SDR displays can really take advantage of. Even footage shot on the RED ONE M sensor looks great in HDR.

We have at least three competing EOTFs right now. SMPTE ST 2084, Philips and Hybrid Log-Gamma. (HLG) HLG was done for backwards compatibly. The BBC wants to broadcast HDR interlaced content and HLG works in their existing system with no metadata and is also royalty free. This assumes they use the 709 gamut because if they go wide gamut, then they are no longer backwards compatible.

Personally I am a fan of 2084, though HLG has some interesting bits in it. The ITU is hoping to publish something by years end, I believe, using HLG.

Vudu has started streaming UHD / HDR (Dolby Vision) content and it looks great! Amazon has been streaming HDR-10. M Go has download HDR-10 content at 100 Mbps. Takes awhile to download.
 
I'm a fan of the standard that is best, that shows the most promise. Worked great for BluRay. Never really understood why the industry kept using the inferior HD-DVD.
But it has always been like this, inferior standards marketed to hell to cloud the truth of which is actually the best. Goes back even to the days of Tesla and Edison. Edison might have been a great inventor and he was great at showing off his feats, but Tesla was the greater of them two.
 
Maik,

I actually am writing one, but it won't go out till all the HDR material is finished, as the article will have various samples of what i mentioned above.

Chris,

I did understood you, and just went a bit over a wider range of things as to refer to the various differences and my fillings about what is happening in actual facts...

But yes, what we have already captured is good enough for current TV/Monitors within the 8/10bits range and with the topic 400 nits, but this material when showed on 800/1000 nits TV/Monitors already shows that there is more that could be shown, and since we are talking about HDR, the more we have the more a given TV/Monitor can show, is like taking a box of 12 crayons vs the 128 ones, literally, I have seen it in person and was blown away...

Stacey,

You are right, but there is more to it, 2084 is also one of my favorite, but what Dolby Vision is doing is just simply put eye blowing... http://www.dolby.com/us/en/technologies/dolby-vision/dolby-vision-white-paper.pdf I am working closely with some very cool people with what I have done and what I am doing for HDR with the RED Dragon in 6k, and soon in 8k, and the way you show current footage, even none HDRx acquired one, is one thing, but even showing HDRx to 3 stops, shows significant improvement and a more realistic image and the looking out of the window phrase comes out a lot, but when over imposing and post working two separate images to create a wider HDR image the difference i huge, as you now have two really great images which have already a fantastic DR in itself, when extended to over 20+ usable, then you see some insane blacks and whites and colors you din't even know where there when you so the flowers with your naked eye...

Now extend that to 2 sets of HDRx acquired images by two separate cameras, and you'll be blown away by what you see... Of course you won't see it the same on your current monitor, not even close, and you won't see it the same even on current prototype monitors with 800 nits, but on a 4000 nits one you'll start crying... When the 10000 nits ones will come to be then I see kids, dogs and cats trying to take food out of the monitor... :)
 
Christoffer Glans said:
I'm a fan of the standard that is best, that shows the most promise.
Click to expand...

Currently there is no standard for HDR, which is a problem. Dolby has a complete end-to-end solution that is well thought out. Sadly many don't want to pay Dolby a royalty and simply took pieces of it, which include SMPTE 2084 and SMPTE 2086 and not really understood how all of it works as you can tell by what Samsung is doing on HDR playback. (or rather not doing)

It will all come together, but we may have to wade through a format war first and first and second gen HDR displays need to be considered disposable.

I have looked at doing some HDR tests using two Weapon's in a 3D mirror rig, which would not have the motion blur issue that HDRx has, but does have the image alignment and color issue that HDRx does not have. This has been done using a pair of Arri cameras.

PDF Creating Cinematic Wide Gamut HDR-Video
Slides: Creating Cinematic Wide gamut HDR-video
Video: https://vimeo.com/85754951 <-- Non-HDR version.
 
Stacey,

You have made two good points, let me expend a bit on them...

Dolby, has done (again) and insane job with Dolby Vision, after Dolby Atmos, and now continued to Dolby Cinema for an end to end deliverable and partnership for a playback of HDR at it's best, while maintaining current technology allowance at it's best yet relatively affordable, even so not the case with the Vizio 120", but as you point out other manufacturers prefer as usual to just still and borrow or reinvent here and there bits and pieces of tech, instead of paying royalties which they are btw pretty fair, this is sad, but I hope that more stand behind Dolby, even so with such a huge corporations such as Samsung, Sony Sharp etc. it will be very hard to get a standard down and get them to come on board with Dolby for a real HDR end to end solution...

And by this the issue I have already seen in the prototype TV/Monitors set for HDR, with supposed high nits luminance, is already in the bad end of the deal for consumer, again...

Personally for both at home and in cinema I will be pushing Dolby, even so I am not directly involved with them, at list not yet... :)

As your second point...

The issue with HDRx and motion blur, as you know (others might not) comes directly from the change in shutter speed/angle which many of us have taken great advantage of for extracting stills form our Motion Capture, aside from using the HDRx for simply expanding or DR and protecting the highlights...

Of course the issue will be less evident when dealing with only 2 stops of HDRx protection, but when going to 4 is already evident, sweet spot for me has been 3 stops, 6 for still extraction, but then at 6 stops I also am exposing for the X frame and not the A, and when I say exposing I really mean guess exposing, unless thanks to the REDCAST and a suitable HDR monitor you are then truly monitoring the HDR exposure as well...

When shooting on a 3D Beam Splitter rig with an absolute Zero Parallax to create two perfectly matched pixel per pixel images, you deal with the slight color change as you do in 3D, one do to the fact of the Beam Splitter Mirror itself, and the other do to the change of exposure settings used on A to B cams, but the is actually a very small issue and well adaptable in Post, with simple color correction.

In short, when using a proper set up 3D Beam Splitter Rig, alinement should never be an issue if the operator knows what He/She is doing, this is not as complex as shooting 3D where a Stereographer must absolutely be present, learning how to set absolute Zero Parallax is a simple process, and you don't have to deal with Interocular, so HDR shot this way is way simpler then shooting 3D.

However, there is still the issue of Calibrated Lenses, and perfectly sinked Follow Focus, and if shooting with Zooms, then when zooming as well, with the obvious attention to the Iris... :)

I think that at the end the use of a single Lens 3D like rig might be the best call, yet even here you have the issue that as of now is impossible to set an absolute Zero Parallax, at just 2mm is nothing, but still something that need correction in Post, a simple yet tedious work that takes time and money...

I am discussing this option in particular with Sebastian over at ScreenPlane, to see if we can come up with a solution for my upcoming HDR projects... If not then the best solution will remain to simply find a middle ground with HDRx.

Btw, this is the Single lens 3D rig I am discing about...

pic0.jpg
 
Ketch beem spliters and hdrx is to capture a wider dr then the camera sensor can handle. That we all know but any picture from any deasent curent camera that is not burnt out and does not have any black pixels is way enough and plenty to use for these new standards. Also the normal range of dragon goes a long way. So in other words shooting for hdr is not so different from how things have been done in the past same rules apply expose well and you have no issues. A beam splitter will not save the shot when there is not enough light in the darks to give the sensor a signal read, on the conterary a beam splitter steal light... So all depending on the shot at hand but again controling the light levels and make sure more or all of the frame is whithin the sensors capability is where its at.
 
Björn,

No doubt about it...

Good exposure is key to everything... No matter how you shoot or what camera you shoot with despite of actual capability of DR in such used cameras, if you don't expose properly one camera exposing incorrectly two cameras won't save you...

However true HDR as in photography is an image acquired by multiple exposures and what I am talking about is actually marrying the new HDR technologies exposed with the upcoming new brighter TV/Monitors and specifically with Dolby Vision and their Dolby Cinemas, where you have true HDR via multiple exposures, which in the case of Motion Picture, can only be done via HDRx, or via Beam Splitter.

Current tests are to be put to the bench to see what actually can be seen and what benefits are to be head with such new technologies becoming available in the playback arena as in HDR TV/Monitors and Dolby Vision.

Is a science of image vs DR vs visible difference on screen against standard single image.
 
Great to hear about those tests. This is like HDR on dope. Getting all the visible "stops" during the acquisition and exposing and lighting for it instead of stretching a standard single exposure to fill the HDR-latitude or add some bright highlights. I totally see where you are going Ketch, just amazing. This is the HDR i want to see.

This, plus 8k, plus HFR and maybe even 3D. :) Would love to see some samples on a proper screen.

P.S. cool single-lens rig too.
 
Ketch, just an idea. Using HDRx for this is great, but as you pointed out, it has the issue with motion-blur.

Here is another approach to shoot more DR with a single camera, which would involve the Motion-Mount and allow up to 8 stops while staying at the same shutter-speed. Basically it would be shooting two exposures, one after each other, with the MM switching ND on and off. I'm assuming the MM is fast enough.

The negative side-effect would be a time-difference between the two exposures - but smart software-interpolation (which can use both exposures) could do the trick. There would be enough information to fill the gaps.

Higher fps and more than two exposures are thinkable too, as long as the MM can do it.

Well, i admit, this is probably a big feature-request for something experimental. ;)
 
I hope they standardize around a perceptual EOTF like SMPTE.2084, not some inefficient layered or log (with "knee" etc) approach.
 
Ketch,

That is an interesting single lens system, thank you for sharing. Is that something that works today? Much smaller than a typical 3D rig.

The link I shared above was the OETF portion of HLG. There is another doc (WHP309) for the EOTF. The thing about HLG is that the gamma changes based on viewing environment. On a tablet or phone they use the camera to get the ambient light. On a fixed system either the display will need a sensor so it can adjust based on viewing conditions or you measure the surround and give it a fixed value.
 
KETCH ROSSi said:
true HDR as in photography is an image acquired by multiple exposures
Click to expand...
Ketch.
You're mistaken. What you're describing is a good old fashion tone mapping, which, I hope to god, will never see the light of day. HDRx has absolutely nothing to do with HDR. HDRx can only be useful for tone mapping. Here are some images done using tone mapping with multiple exposures:
https://www.flickr.com/search/?q=tone mapping
Tone mapping is fine for some still pictures, but it is an awful idea for moving images.
Pretty much ANY modern camera and film is capable of producing material suitable for HDR mastering. HDR is all about the display technology and it has nothing to do with camera technology. That is why, I think, HDR re-mastering of old movies will be big business, like HD re-mastering used to be.
 
Maik Müller said:
Ketch, just an idea. Using HDRx for this is great, but as you pointed out, it has the issue with motion-blur.

Here is another approach to shoot more DR with a single camera, which would involve the Motion-Mount and allow up to 8 stops while staying at the same shutter-speed. Basically it would be shooting two exposures, one after each other, with the MM switching ND on and off. I'm assuming the MM is fast enough.

The negative side-effect would be a time-difference between the two exposures - but smart software-interpolation (which can use both exposures) could do the trick. There would be enough information to fill the gaps.

Higher fps and more than two exposures are thinkable too, as long as the MM can do it.

Well, i admit, this is probably a big feature-request for something experimental. ;)
Click to expand...

No this would not do the trick in any better way than hdrx does. the motion blur issue from hdrx is the fact that the sensor first offload a portion of the photosite filling then all of it. so the motion blur of hdrx is actually better than what you sugest as what you sugest the motion blur is actually not overlapping....
 
You must log in or register to reply here.
Back
Top