HDR, holy confusion. What did I miss?

O

Oller Visuals

Well-known member
Thought I knew what HDR is, but apparently not..

Heard that Netflix was adding HDR support for Win 10, and I was like "what?" Read up on it and it appears that what people call HDR is simply TVs that can display +10 bit color depth. I don't have a TV so I dunno...

I thought HDR was refering to the process of either combining multiple exposures to create the "HDR look", or shooting with something like a RED to allow scenes with heavy color grading. And when I see all these mumbo jumbo terms like " UltraHD Premium" I kind of cringe.

Can someone explain what this HDR fuzz is about? Posting this in cinematography since it's related to the source material, kind of... I guess? Unless I convert my 8-bit video to 10-bit, which I've been doing for years... oh I guess I've been doing HDR all along!

Huh. So confused.
 
Vimeo and YouTube also support HDR playback provided you are viewing their content with a HDR display.

And you should also get caught up on the HLG video format, because some of the newer Panasonic and Sony cameras can record HLG HDR video.
 
You might find this article I researched and wrote for HD Video Pro two years ago about DolbyVision to be helpful in understanding HDR from a consumer and somewhat from a content creator perspective.
https://www.hdvideopro.com/workflow/capture-workflow/the-art-of-better-pixels/#

You can read what Dolby markets about DolbyVision here https://www.dolby.com/us/en/brands/dolby-vision.html

It has very little to do with 8-bit versus 10-bit and has everything to do with increased dynamic range, contrast ratios and a more lifelike reproduction of images. It also is the main driver behind home cinema at least being perceived from an image standpoint in the same leage as cinema in a proper theater.
 
I went to a Panasonic demo day couple of weeks ago and their line was that their cameras have 14+ stops of DR (that was really stretching it from the footy they showed) so they are HDR. Bwahahahaha- it was so stupid that I felt brain insulted. Here we go again- that's how they destroyed 3D- and haven't learned a thing. I fully expect HDR sticker on EVA1 soon... Ironically- I could not find the "Eva1" designation on that camera anywhere ;-)
 
optitek said:
I went to a Panasonic demo day couple of weeks ago and their line was that their cameras have 14+ stops of DR (that was really stretching it from the footy they showed) so they are HDR. Bwahahahaha- it was so stupid that I felt brain insulted. Here we go again- that's how they destroyed 3D- and haven't learned a thing. I fully expect HDR sticker on EVA1 soon... Ironically- I could not find the "Eva1" designation on that camera anywhere ;-)
Click to expand...
I am not sure what you think is stupid about this. A camera that can record 14 stops can certainly deliver HDR.
 
If it doesn't see what my eyes see it won't . And it doesn't . I saw the footy close and personal. Nothing wrong with the camera but.... I guarantee you that most viewers feel the same. They killed the "3D" and they will kill this travesty as well. In the mean time Panasonic will sell thousands of devices with the "HDR" sticker on them- that's what this is all about- not the actual product quality or features. Sad really....
 
optitek said:
I saw the footy close and personal.
Click to expand...

I'm not sure if you've followed any of the links you were given. I don't think you understand what HDR for video content is.

Typically, most content we watch today is 8-bit Rec. 709. That's good for about 6 stops or so of dynamic range. One way to look at it would be to say that a camera that could deliver more than that is shooting "HDR". Many modern cameras today capture 12 stops or more. Among digital film cameras 14+ stops is expected. Already 12, but certainly 14 stops is firmly in HDR territory.

Without a matching video standard to deliver video in, those nice cameras needed to compress their captured data to fit in a smaller container (or cut the data out). But new HDR standards stipulate a significantly larger color space and 10-bits or more. The new HDR delivery formats together with new TV-panels with much higher average and peak brightness allow for a new type of grading with more lifelike light distribution. Typically, most of the image will have similar contrast to what we are used to from SDR, but then smaller areas of the frame will have much higher brightness—like daylight outside a window, sirens on a cop car, headlights/tail lights, flames of fire and so on...

But… a number of things need to play together before you get HDR. Just downloading a 14 stop demo clip from RED and watching it on your normal computer monitor won't make it HDR. You'll need to grade in Wide Color Gamut and export it as such. Then you need to watch it on a HDR monitor. There are lots of different specs surrounding all of this.

Good TV sets from LG's OLED series that are popular have a brightness of around 700 nits I believe. Dolby Vision's HDR standard goes all the way up to 10.000 nits, even if grading is usually done with a 4000 nit target.

I'm not sure what the brightness capabilities are on Apple's latest devices, but I know the recent iPad Pros and new iPhone X will also trigger HDR mode from Netflix for example.

So: several specifications, several takes on transfer functions and lot's of different display targets. It's a bit of a jungle and quite foolish of me to even start to break it down in a post like this. It's quite technical and any "information" just scraping the top like this is basically an open invitation to nitpicking.

But HDR is a thing. And unlike 3D, it's definitely here to stay. All our displays will be HDR soon. For those of us with keen eyes: a joy. For the more casual viewer: totally unobtrusive, even if they might see images with a bit more 'pop' than they're used to.
 
The Panasonic demo

The Panasonic demo

AndreeOnline said:
I'm not sure if you've followed any of the links you were given. I don't think you understand what HDR for video content is.

Typically, most content we watch today is 8-bit Rec. 709. That's good for about 6 stops or so of dynamic range. One way to look at it would be to say that a camera that could deliver more than that is shooting "HDR". Many modern cameras today capture 12 stops or more. Among digital film cameras 14+ stops is expected. Already 12, but certainly 14 stops is firmly in HDR territory.

Without a matching video standard to deliver video in, those nice cameras needed to compress their captured data to fit in a smaller container (or cut the data out). But new HDR standards stipulate a significantly larger color space and 10-bits or more. The new HDR delivery formats together with new TV-panels with much higher average and peak brightness allow for a new type of grading with more lifelike light distribution. Typically, most of the image will have similar contrast to what we are used to from SDR, but then smaller areas of the frame will have much higher brightness—like daylight outside a window, sirens on a cop car, headlights/tail lights, flames of fire and so on...

But… a number of things need to play together before you get HDR. Just downloading a 14 stop demo clip from RED and watching it on your normal computer monitor won't make it HDR. You'll need to grade in Wide Color Gamut and export it as such. Then you need to watch it on a HDR monitor. There are lots of different specs surrounding all of this.

Good TV sets from LG's OLED series that are popular have a brightness of around 700 nits I believe. Dolby Vision's HDR standard goes all the way up to 10.000 nits, even if grading is usually done with a 4000 nit target.

I'm not sure what the brightness capabilities are on Apple's latest devices, but I know the recent iPad Pros and new iPhone X will also trigger HDR mode from Netflix for example.

So: several specifications, several takes on transfer functions and lot's of different display targets. It's a bit of a jungle and quite foolish of me to even start to break it down in a post like this. It's quite technical and any "information" just scraping the top like this is basically an open invitation to nitpicking.

But HDR is a thing. And unlike 3D, it's definitely here to stay. All our displays will be HDR soon. For those of us with keen eyes: a joy. For the more casual viewer: totally unobtrusive, even if they might see images with a bit more 'pop' than they're used to.
Click to expand...
I've seen tons of "HDR" footage at trade shows as well as the demo footy at the mentioned event on a "HDR" display in a darkened room. Nothing "popped" that's for sure and not only my opinion but pretty much everyones. Even that it was a split screen demo. I go for "EDR" (enhanced) however poorly executed. I suspect that a 16 bit aquisition with resulting 18-20 stops would do the trick. As it stands now- nothing more than the usual smoke and mirrors ;-)
 
optitek said:
I've seen tons of "HDR" footage at trade shows as well as the demo footy at the mentioned event on a "HDR" display in a darkened room. Nothing "popped" that's for sure and not only my opinion but pretty much everyones. Even that it was a split screen demo. I go for "EDR" (enhanced) however poorly executed. I suspect that a 16 bit aquisition with resulting 18-20 stops would do the trick. As it stands now- nothing more than the usual smoke and mirrors ;-)
Click to expand...
I could not disagree with you more, the difference between HDR and SDR is very easy to spot.
 
optitek said:
I've seen tons of "HDR" footage at trade shows as well as the demo footy at the mentioned event on a "HDR" display in a darkened room. Nothing "popped" that's for sure and not only my opinion but pretty much everyones.
Click to expand...

I'm all for you making up your own mind—that's healthy.

But depending on background, or previous HDR experience (let's say from photography), it's interesting that HDR for video is almost the opposite from that of photography. Also, if you google 'HDR images' you'll basically get the photography version.

In photography, HDR techniques mean you take a number of bracketed exposures and capture lots of DR (maybe 18 stops, often in 14-bit raw). But then you compress all the data with the help of tonal compression algorithms back into SDR range. So this is essentially like SDR video but where the starting point was way more DR than the format could handle and the assumption that all the data needed to be compressed back down. The result is the utterly unpleasing, totally fake looking phenomenon most people associate with HDR imagery.

OK, I might exaggerate here. I'm not a huge fan of HDR photography with extreme tonal remapping, but I know great results can be achieved if these techniques are applied by someone with taste.

But HDR for video is different. You often do just one take with a high DR camera and capture as much quality DR you can. But instead of compressing the range back down you use new formats and displays to try to display as much of the original scene DR as possible. But this can mean you have something that looks like a normal SDR image with the exception of that one practical light in the right corner being REALLY bright.

Photo HDR = 14 stop scene --> tonal remapping --> 8 stop SDR image with "HDR local contrast look"
Video HDR = 14 stop scene --> new video standards maintain more DR and greater color space --> 10-bit+ displays with high brightness --> 10/12 bit actual HDR image with, ideally, scene contrast maintained.

EDIT: and by the way, with the new displays emerging, I'm quite sure it won't be long until HDR photographers will get a new format for HDR images will be equally amazing. I know Dolby is working on JPEG-HDR that will be backwards compatible and allow for the luminance data to be maintained over 8-bit jpeg and where the images look consistent over both SDR and HDR displays.
 
Last edited:
Hey Optitek, what monitor and footage did Panasonic use to show the HDR? They came to my company a few months ago to show their new cams but didn't have a HDR monitor in production at the time. It didn't look very good when they needed a very expensive Sony Trimaster ($40K) monitor to show their own 4K HDR cameras. The reps were very knowledgeable, the demo went very well and everyone could easily notice the difference between the HDR and non HDR, but we still went with Sony. Just wondering what they brought and displayed at your demo. Thanks
 
Volt said:
It didn't look very good when they needed a very expensive Sony Trimaster ($40K) monitor to show their own 4K HDR cameras.
Click to expand...
What is the problem with using the best available monitor to display the capabilities of their cameras?
 
My wife couldn't tell the difference. I know what to look for, so I can, and she saw it when I pointed it out, but it really is not a revelation's worth of difference on final display.

Not to mention that out of the box, many TVs' HDR display is rather dark.
 
AndreeOnline said:
Typically, most content we watch today is 8-bit Rec. 709. That's good for about 6 stops or so of dynamic range.
Click to expand...

Where are you getting this "6 stops or so of dynamic range" for BT.709? That would be a measly 64:1 contrast ratio. By any measure, BT.709 exceeds that by a lot. I've often seen the 6-stop figure repeated, but never backed up.

There are a lot of different ways to think about dynamic range, but none of them give 6 stops for BT.709. If you are considering the actual BT.709 OETF, the black point is 0% reflectance, which gives infinite contrast. The next highest grey in 8 bits is 10 stops below the white point. If you are considering SDR cameras, for years now cameras have had 8-10 stops or more of dynamic range. If you are considering SDR LCDs, even without local dimming, they have contrast ratios of hundreds or a thousand or more to one. OLED displays have virtually infinite contrast even in SDR mode. A gamma-2.2 display, which is the most common EOTF for HDTVs, has its darkest grey (in 8 bits) 17 stops below the white point.
 
I agree with you entirely when you say cameras have long been able to cope with 8, 10, 12 stops or more. The fact is on Rec 709 displays we can't display any more than 6 stops so we somehow have to compress all our lovely 10-14 stop or whatever range into the capability of the display technology. It is very easy to confirm this by using a good quality light meter on a properly adjusted 709 display. Put up a true black then white signal to the display and measure the difference between the black and the white. In fact just to confirm this I did a quick and dirty test. I stuck a 1 volt 100 IRE white test signal in to a middle of the road pretty representative $2,000 Panasonic 50" LED TV which was set to Standard mode. The exposure reading was f/9.5 at 1/50th second. I then sent a Black 0 volt 0 IRE signal to the screen so the only visible output from the screen was the reflected ambient room light. This measured up at f/1.4 at 1/50th. In this instance the display output range was just on 5.5 stops. This is what defines the 6 stop range of any 709 compliant displays. This is why SMPTE define ITU R-709 as a display measurement.

Alistair goes into 709 in this FS5 video in a fair bit more depth but I think most people would follow what he is saying relative to the 709 discussion. The pertinent 709 part is from 1:09:00 onwards.

https://www.youtube.com/watch?v=U-HrXmRqpME

ChrisYoung
 
balazer said:
Where are you getting this "6 stops or so of dynamic range" for BT.709? I've often seen the 6-stop figure repeated, but never backed up.
Click to expand...

It probably shows up a lot because it's what Wikipedia quotes for the system.

The actual value is derived from having 220 possible quantisation levels in the 8 bit video signal and the assumption that if adjacent levels in luminance differ by 2% or less, then it's imperceptible. But this threshold changes over luminance values—it would for example increase and be more forgiving in lower light levels. If a threshold of 5% is assumed, to allow for a dim display, using an 8 bit signal and a gamma of 2.4 the Weber fraction (the mathematical ratio between quantisation levels) gives a DR of 5.27 stops.

You could use a different threshold and gamma to account for a different system, to land at another DR value, but to answer your question—I think this is where that comes from.

balazer said:
There are a lot of different ways to think about dynamic range…
Click to expand...

I don't think so. It's simply the ratio between the darkest and brightest value, or the ratio between peak signal and noise floor. But if we look at the whole system, you're right that we have many variables: OETF (the way a camera captures light non-linearly), EOTF (the way it's displayed, non-linearly) and the resulting OOTF for the system as a whole, since OETF and EOTF might not cancel each other out.

Anyway, this is not my area of expertise and the quote you use was simply framing for further discussion to compare where we come from and where we're heading.
 
You must log in or register to reply here.
Back
Top