Interesting article on HDR

[Björn Benckert]

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.

tone mapping uses transfer modes etc, but if you just mix / blend a and b then you just get a lower contrast "negative" to start your grade from I think that is what ketch is after... which could have benefits at times but for sure, But I agree that is not really where the revolution of HDR screening comes to play. Nobody on the planet will see if it was shot with a beam splitter or if the DP simply just shot with a bit less contrast in the lights... result for the viewer is identical.

 

[jake blackstone]

tone mapping uses transfer modes etc, but if you just mix / blend a and b then you just get a lower contrast "negative" to start your grade from I think that is what ketch is after... which could have benefits at times but for sure, But I agree that is not really where the revolution of HDR screening comes to play. Nobody on the planet will see if it was shot with a beam splitter or if the DP simply just shot with a bit less contrast in the lights... result for the viewer is identical.

One more time- ANY modern camera and film acquired material already perfectly suitable for HDR finish. There is no need to invent a new camera for the HDR finish.
From IMDB for Tomorrowland, which was an HDR released feature:

Sony CineAlta F65, Zeiss Master Prime and Fujinon Premier Lenses
Sony CineAlta PMW-F55, Zeiss Master Prime and Fujinon Premier Lenses

HDR and tone mapping are not the same thing as are HDR and HDRx.
HDR mainly s a display technology, that can display HDR encoded images.
Tone mapping is a post technique for combining multiple exposures.
HDRx is a camera technique for producing double exposed images used to expand camera dynamic range.
Not the same thing!

 

[Ketch Rossi]

Maik,

It won't work that way as Björn pointed out, but some interesting tests are on the way, we'll see what happens...

Stacey,

The Screen Plane "Ultra Macro" 3D rig, works right now and has been used to capture some amazing extreme close up 3D cinematography. As for use for my HDR testing purposes I a discussing them with Sebastian, Screen Plane's CEO as this rig can't do Absolute ZERO Parallax right now... We'll see what happens... But in it self it's an amazing 3D little Single Lens rig that takes 3D capture beyond standards.

Jake,

I am not, what I referred to as in photography is multiple exposures by in camera automated change of shutter speed timing, which in still photography it's perfectly fine to then bring together and clone them in to one single image...

Tone mapping is used by some while not by others, I have used both, but never liked the Tone Mapped extreme HDR look you see on the net...

HDR as in camera capacity to capture a wider DR is of course not HDRx, we know this very well...

What I meant is that we are in the process of testing multiple exposure captures with and withOUT the assistance of HDRx, for the purpose of acquiring extreme HDR imagery, to then further test on the new HDR display technology.

I, am extremely very well aware of the exact specs and the fact that this new HDR Display technology does NOT need HDRx or multiple camera acquired HDR imagery to work, but we are in the process of deep testing how far this new technology can display the HDR imagery and how much information can be displayed in consideration of pixel per pixel and in the total exposure.

I imagine that for sure you also know that this new standard, does not just start nor finishes at the acquisition of the HDR imagery, or simply by displaying already acquired material, which is perfectly fine, but moves forward to test the capacity in itself of what can be seen and displayed.

Additionally and most importantly this will fall back on the guys like you, in post, during the Color Grading, which needs to be done a bit differently, or a lot differently but in not a such a complex way, while yet intricate in itself, and what I ma talking about here are of course the only HDR playback standard I am currently supporting and interested in, 2084 via Dolby Vision, Dolby Cinema.

I see interesting times ahed for experimentation, for both Cinematographers and Colorist...

 

[Stacey Spears]

HDR and tone mapping are not the same thing as are HDR and HDRx.

While tone mapping does deserve a lot of the negative comments it receives, tone mapping is used by Dolby Vision and other HDR playback solutions to make the HDR content looks its best on a given HDR display. The idea is the mastering display has a wider gamut and more light output than the display you are watching it on at home and therefor tone mapping is used to "make it fit" for lack of a better phrase. Those would be good examples of tone mapping unlike the million bad examples that show up when you Google it.

 

[Ketch Rossi]

While tone mapping does deserve a lot of the negative comments it receives, tone mapping is used by Dolby Vision and other HDR playback solutions to make the HDR content looks its best on a given HDR display. The idea is the mastering display has a wider gamut and more light output than the display you are watching it on at home and therefor tone mapping is used to "make it fit" for lack of a better phrase. Those would be good examples of tone mapping unlike the million bad examples that show up when you Google it.

Exactly...

And once again, here is to be understood that while this new HDR Display tech is meant to support all currently shot Footage, the fact that it can handle so much more DR, is what has warranted this test, with Multicam and HDRx + Tone Mapping via the new Dolby applied Color Grading.

Tests, are just that, but what might come out of them could surprise...

 

[Ketch Rossi]

Oh,

Jake, totally agree with you, HDR re-mastering will be big game... Let's hope it is done properly, and not in mass, quick and cheap, meaning giving a limited hours/pay budget to do it...

I loved 3D, still do, but so much went wrong with it...

I always loved HDR, and this is an amazing step forward in visual displays in my strong opinion, and truly hope it plays out right...

 

[Ketch Rossi]

THIS. And 2084 for the win.

Yes Mark,

I know that extremely well as I had already replied to Stacey... What I/we are doing is very different...

The fact that this HDR display tech is designed to work with current material by showing more Blacks and more Withes, and that the HDR re-mastering is not only possible but very likely...

Does not mean that there is not the possibility for a much HIGHER DR capacity to be displayed, and this is what we are testing on...

Simply playing with both single RED camera with multiple exposures via HDRx and Multiple RED cams with single exposures and Multiple RED cams with Multiple HDRx exposures, to see how much DR we can pull out and what it looks like on the HDR display...

Playing is fun, discovering even more fun, finding out that the images might look so fucking great then they ever could otherwise, even greater... But for now we are just playing...

And yes, absolutely hope for the 2084 standard as I absolutely love it, via Dolby Vision...

 

[Stacey Spears]

Does not mean that there is not the possibility for a much HIGHER DR capacity to be displayed, and this is what we are testing on...

I have been calling that extreme HDR or HDR-11. Bill Feightner and I discussed this at SMPTE.

How do you plan to blend the frames together? That is the issue I always had with HDRx. Its one thing if you have a static shot with a window, which is fairly straight forward to composite. Its another when you have motion. Too bad Red Rocket Fuel never came to be. The demo / sample of it looked promising on FXPHD.

 

[Ketch Rossi]

Yeah it has been called already several names, I called "Super HDR" when dealing with 4 exposures thru 2 cams+HDRx...

This is the fun part of the testing to be head, the how to properly take advantage of each individual multiple exposure is key, and while GREAT, HDRx, in each blending flavor currently available via RCX does not do the material justice, we need an absolute individual control per each file for a desired blending + Tone Mapping + Color Grade.

Under NDA, so can't discuss anything really, just that I really, really hope things work out, even if some one else's comes out with first, I/we still win...

 

[jake blackstone]

I imagine that for sure you also know that this new standard, does not just start nor finishes at the acquisition of the HDR imagery, or simply by displaying already acquired material, which is perfectly fine, but moves forward to test the capacity in itself of what can be seen and displayed.

HDR imagery has zero to do with camera. As I already said twice, pretty much all modern cameras are already acquiring HDR images. Some cameras can exceed 15-16 stops of dynamic range. Displays on the other hand considered to be excellent, if they can do even half of that. What HDR does is simply introduces a DISPLAY technology, that starts to move toward the approaching the dynamic range of the acquired images. And I'm not even talking about the color gamut of the existing displays, which are mostly limited to just Rec-709. Most high end cameras routinely shoot at 16 bit per color channel. Most displays are limited to just 8-10 bits. That is again about half of the cameras capabilities. HDR need 12 bits per color for displays to display HDR correctly.
So, the bottom line, there is absolutely no immediate need to improve already existing cameras in order to produce HDR images, they already do it very well. The bottleneck are the displays and the competing standards. And when that technology will get sorted out, we'll have to deal with the distribution dilemma, because HDR does require more bandwidth...

 

[Michael Tiemann]

I just cannot resist adding this little bit of confusion to the thread: with a beam-splitting solution you can shim the focus of one of the two cameras so that you get two simultaneous focal distances. Then you can use HDRx to stack exposures one way, and get deeper DOF by focus-stacking. A handy calculator would tell you just how many mm of displacement you'd need to double the DOF either in front of or behind the subject. Or, go crazy and have two different depths of field that you can merge together with high-pass masking (a kind of tone-mapping for focus). Could be great for two-person dialog shots!

 

[Ketch Rossi]

HDR imagery has zero to do with camera. As I already said twice, pretty much all modern cameras are already acquiring HDR images. Some cameras can exceed 15-16 stops of dynamic range. Displays on the other hand considered to be excellent, if they can do even half of that. What HDR does is simply introduces a DISPLAY technology, that starts to move toward the approaching the dynamic range of the acquired images. And I'm not even talking about the color gamut of the existing displays, which are mostly limited to just Rec-709. Most high end cameras routinely shoot at 16 bit per color channel. Most displays are limited to just 8-10 bits. That is again about half of the cameras capabilities. HDR need 12 bits per color for displays to display HDR correctly.
So, the bottom line, there is absolutely no immediate need to improve already existing cameras in order to produce HDR images, they already do it very well. The bottleneck are the displays and the competing standards. And when that technology will get sorted out, we'll have to deal with the distribution dilemma, because HDR does require more bandwidth...

Again,

Jake, I understood and not disagreeing with you, on the contrary, but adding to that the the current, does not mean anything to what we are testing, I understand most perfectly the current capacity of cameras and Displays, and what they can do, and where this HDR new standard to come is coming from or better said for, yet it does not mean that you can't see what else you can get out of them, and this is exactly what we are talking about...

15-16 stops might be fine for some, same as 1080p might be fine for some, while others go to 4k, and are very happy with it, then again, here we have 8k...

HDR Displays, and the current prototype Displays, and not talking about just the 8000-1000 nits one that will hit the market after the 400 nits ones, with completely new lighting technology superior to current ones, which are otherwise getting too hot to handle continuos high nits output...

My input, has been EXCLUSIVELY join regards to been doing some interesting tests to see how far HDR cam be pushed on this new Technology within the highest Dolby Vision standards I 2084, that is all...

Not sure if people here are just NOT reading or simply completely misunderstanding what I am saying...

I just cannot resist adding this little bit of confusion to the thread: with a beam-splitting solution you can shim the focus of one of the two cameras so that you get two simultaneous focal distances. Then you can use HDRx to stack exposures one way, and get deeper DOF by focus-stacking. A handy calculator would tell you just how many mm of displacement you'd need to double the DOF either in front of or behind the subject. Or, go crazy and have two different depths of field that you can merge together with high-pass masking (a kind of tone-mapping for focus). Could be great for two-person dialog shots!

And then there can be that too...

 

[jake blackstone]

Again,

Jake, I understood and not disagreeing with you, on the contrary, but adding to that the the current, does not mean anything to what we are testing, I understand most perfectly the current capacity of cameras and Displays, and what they can do, and where this HDR new standard to come is coming from or better said for, yet it does not mean that you can't see what else you can get out of them, and this is exactly what we are talking about...

15-16 stops might be fine for some, same as 1080p might be fine for some, while others go to 4k, and are very happy with it, then again, here we have 8k...

HDR Displays, and the current prototype Displays, and not talking about just the 8000-1000 nits one that will hit the market after the 400 nits ones, with completely new lighting technology superior to current ones, which are otherwise getting too hot to handle continuos high nits output...

My input, has been EXCLUSIVELY join regards to been doing some interesting tests to see how far HDR cam be pushed on this new Technology within the highest Dolby Vision standards I 2084, that is all...

Not sure if people here are just NOT reading or simply completely misunderstanding what I am saying...

And one more time, you're talking about increasing the dynamic range of a camera, which is fine But you keep talking about increasing camera's dynamic range in context of display HDR technology, like Dolby Vision, which is strictly a display technology. Existing cameras already produce images, that far exceed existing display technologies in resolution, dynamic range and color gamut. Increasing dynamic range of the camera will have no impact on display HDR technology.

 

[Ketch Rossi]

And one more time, you're talking about increasing the dynamic range of a camera, which is fine But you keep talking about increasing camera's dynamic range in context of display HDR technology, like Dolby Vision, which is strictly a display technology. Existing cameras already produce images, that far exceed existing display technologies in resolution, dynamic range and color gamut. Increasing dynamic range of the camera will have no impact on display HDR technology.

Ha ha,

And you are talking about EXISTING, and I said CURRENT... So we are talking about one and the same thing...

EXCEPT...

I am saying that I/WE are testing NEW stuff... Which has to do with upcoming new technology in HDR Displays... NOT existing NOR current one...


And the word, TESTING should have been enough... I guess I should have posted nothing to begin with...


Also my last comment on the matter, s I either just simply can't make myself understood, or the fact that I can't fully disclose what we are working on is clouding the way I am describing or trying to, so I'll stop here...

If we succeed you'll all see the results next year, if we don't, then never mind... No one will hear about it... And now it makes me re-tink of why I even shared my thoughts of it here...

 

[sam karr]

Under NDA, so can't discuss anything really, just that I really, really hope things work out, even if some one else's comes out with first, I/we still win...

Here we go again ....

 

[Ketch Rossi]

Here we go again ....

Why am I not surprise that you had to Chaim in...

 

[Maik Müller]

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....

Yes, it's overlapping with HDRx and would not with a flickering ND. The difference would be that there is the same amount of blur in both exposures. So no "magic-moton" side-effects. With a simple setup in mind (24fps 1/48th target) one would shoot 48fps with 1/48th. The result is a seesaw-like stream for time and exposure.

The idea is to fill the gaps in time with interpolated frames. Interpolation would have access to all 48fps (both exposures) and that way even more motion data than you often have with stuff like Twixtor today. The algorithm would need some overlapping in exposure to keep track of objects, pixels and of course needs to know that every second image has a different exposure to it.

...and a bit like with "Hypercinema" such a thing would become more fun with more frames per second.

Even if this idea isn't working out (i'm just thinking out loud here) i'm pretty sure the Motion-Mount has some potential for HDR...

 

[jake blackstone]

Ha ha,

And you are talking about EXISTING, and I said CURRENT... So we are talking about one and the same thing...

EXCEPT...

I am saying that I/WE are testing NEW stuff... Which has to do with upcoming new technology in HDR Displays... NOT existing NOR current one...


And the word, TESTING should have been enough... I guess I should have posted nothing to begin with...


Also my last comment on the matter, s I either just simply can't make myself understood, or the fact that I can't fully disclose what we are working on is clouding the way I am describing or trying to, so I'll stop here...

If we succeed you'll all see the results next year, if we don't, then never mind... No one will hear about it... And now it makes me re-tink of why I even shared my thoughts of it here...

I'm sorry Ketch, but I'm totally confused. You said this:
This is the fun part of the testing to be head, the how to properly take advantage of each individual multiple exposure is key, and while GREAT, HDRx, in each blending flavor currently available via RCX does not do the material justice, we need an absolute individual control per each file for a desired blending + Tone Mapping + Color Grade.
Nowhere have i seen in all your posts in this thread you discussing a DISPLAY technology, you're only discussing camera and post technology utilizing multiple exposures in order to increase it's dynamic range. And now all of a sudden after all this you're talking about the "upcoming new technology in HDR Displays"?
What am I missing?

 

[Ketch Rossi]

I'm sorry Ketch, but I'm totally confused. You said this:
This is the fun part of the testing to be head, the how to properly take advantage of each individual multiple exposure is key, and while GREAT, HDRx, in each blending flavor currently available via RCX does not do the material justice, we need an absolute individual control per each file for a desired blending + Tone Mapping + Color Grade.
Nowhere have i seen in all your posts in this thread you discussing a DISPLAY technology, you're only discussing camera and post technology utilizing multiple exposures in order to increase it's dynamic range. And now all of a sudden after all this you're talking about the "upcoming new technology in HDR Displays"?
What am I missing?

Sorry Jake,

Most of the time I confuse myself.... Because that original post was abut the article of HDR Displays, I took form it to talk about what we where working on testing even higher DR for the HDR Displays, of course there are not just ONE type of them...

But just to be clear, what we are testing with Higher DR captures is for all displays...

This is mainly a test for a new software to allow the working of multiple images to create true HDR, for motion, and be Displayed on the new HDR monitors, or the standard ones, of course with very different results.

Again, this are just tests, and yes, they are for HDR Software, HDR Display, and a Camera system...

 

[Marc Wielage]

Existing cameras already produce images, that far exceed existing display technologies in resolution, dynamic range and color gamut. Increasing dynamic range of the camera will have no impact on display HDR technology.

I believe Jake is 100% right here. We've been compressing the dynamic range of images for many, many years, just from film and regular HD cameras, let alone contemporary wide-range CMOS cameras.

Anybody who doubts this, read Dolby's paper on Dolby Vision:

http://www.dolby.com/us/en/technologies/dolby-vision/dolby-vision-white-paper.pdf