SDVoE LIVE! on demand
Episode 4 – HDR Explained
HDR creates beautiful pictures, but it can also create headaches in large scale AV distribution systems. Know the potential pitfalls ahead of time, so that your design can account for and avoid them – giving your customers the best picture possible, always. When things go wrong with HDR, you need to understand how to fix them. This episode will introduce you to the basics of HDR technology, but also give you the resources to recognize specific HDR-related problems, and offer troubleshooting techniques and solutions.
Stéphane Tremblay, Semtech
Stéphane Tremblay is CTO of AptoVision products in Semtech’s Signal Integrity Products Group. He has spent his entire professional career leading the conception, development and productization of numerous AV, graphics and image processing technologies. In 2011, he co-founded AptoVision with the vision of transforming AV signal distribution by replacing proprietary switching equipment with off-the-shelf Ethernet equipment and has since led the AptoVision team in delivering on that vision with the BlueRiver series of products.
Justin Kennington (00:10):
Hello. Hello and welcome to SDVoE Live. I’m your host, Justin Kennington and this is TV for pro AV. We have a very important show for you today. We’re going to cover a topic that is important to pro AV, important to the future of consumer video electronics, but also a little bit confusing in the world of pro AV and that is high dynamic range imaging, HDR. We’re going to cover that with our special guest Stéphane Tremblay from Semtech AptoVision, one of the creators of the technology that drives SDVoE and a real expert in HDMI and all kinds of video systems. He’s going to have a lot to share with us about that.
Watch throughout the show, as we have a couple of quiz questions coming at you in some of the breaks. Those of you who are able to successfully answer those quiz questions either in the chat window down below me or on Twitter at #sdvoelive will be entered to win a prize.
Before we move on, I want to pause for a minute and reflect on someone very important that the pro AV industry lost a week ago, someone important to pro AV but also important to me personally and that was Fred Bargetzi, the CTO of Crestron. He is the reason that I moved into the pro AV industry from outside of it. He taught me that with the right team with you, all you have to do is set very big goals and commit yourself and you can achieve almost anything. For that reason I would say he’s the reason that you’re seeing this show today because of what he taught me.
I remember a particular incident. I was in his office complaining about some tasks that he needed me to do and I thought it wasn’t going to be possible and I couldn’t do it and blah, blah, blah.
Fred said, “Just stop. Just go do magic.” And you know what? I don’t remember what it was but I left his office and I figured it out and I made it happen. It was his way of inspiring people to believe that they could do things that they didn’t know they could do that I think stood out the most. Not to mention just what a nice person he was. You meet so few truly successful people that are also just really nice and so I’m really sorry to miss him. I think his family’s asking for donations to Compassionate Care ALS, so if you’re so inclined, check that out. I’d like to dedicate today’s show to Fred’s memory and say one last thanks to him for everything he taught me.
With that, and speaking of teaching things, I’m going to take a break now and hand it off to my co-host Matt Dodd, who you’ll see in a few seconds in our classroom and he’s going to teach you about the basics of high dynamic range, HDR. We’ll see Matt in a minute.
Matt Dodd (03:21):
Hello again, everybody Matt here, Matt Dodd. Great to have you back, hope you had a lovely Christmas and a New Year break. Happy New Year as they say over there in the US and Canada. All right, that’s what they say over here in the UK. It’s a cold evening here in the UK. It’s just cold and damp, but hey, the sun’s always shining in the SDVoE building. I think I’ve just given away the fact that we’re not in the same building, but I think you’ve probably gathered that by now. So let’s press on with a clip from the SDVoE Academy course, “What is HDR?” As with everything in education, it’s really important that we don’t assume anything. We can’t do that, so what we’re going to do, first of all, is just go through the basics of HDR, take a look at this, and then we can move that forward and introduce Stéphane Tremblay. So enjoy.
HDR, or high dynamic range, has been used extensively in the world of photography for years and over the last 10 years, HDR has been introduced to video content to give us a far deeper image quality and definition, especially noticeable on larger panels which have become today’s norm. This course will explain HDR technology and give you clarity on exactly how it improves the quality of an image. Put simply, HDR improves the quality of an image by creating whiter whites and darker blacks while making millions more colors appear on the screen. By increasing the contrast and brightness capabilities of a display, we’re able to make a far wider range of colors and color shades available. In other words, by increasing brightness and contrast levels, we’re able to display deeper and more vibrant shades of blue, green, and red across the screen because the range from dark to light is far higher in a display that can support HDR.
Matt Dodd (05:46):
In the “Color Space and Color Depth” course, we introduced you to the chromaticity chart to explain the differences between Rec 601 or standard definition, Rec 709 or high definition and Rec 2020 or ultra high definition. You can revisit that course by clicking here in the top right.
We’re going to use the same chromaticity chart to help us explain HDR because while it’s not a direct component of HDR, WCG, or wide color gamut, describes the amount of colors which can be shown at any one time on a display and once you understand this, you’ll have more clarity on the benefits of HDR.
The brightness of a display is measured in nits and when the technical wizards of yesterday went to work making HDR as effective as possible, the standard definition technology available forced video content to be produced to meet the requirements of the typical CRT TVs available using the Rec 709 color gamut and brightness of around a hundred nits. Contrast this to the human eye, which can perceive up to 10,000 nits of brightness. With SDRs limitations, content creators had to carefully choose how to squeeze a realistic image into the hundred nits of brightness available. It’s no wonder you’ve never confused a TV screen with reality. HDR has dramatically changed this because content is being produced using a color gamut and a peak brightness which is more similar to the capabilities of the human eye. However, not all HDR displays and formats are created equally so content may include wider color gamut, and higher peak brightness than the displays capabilities.
Metadata is the term used for embedded information about the content which helps a lower-brightness display understand how to successfully display a high-peak-brightness image. The world of video has become a lot more complicated, if the metadata wasn’t present, then a display wouldn’t know how to correctly show the HDR content it’s receiving. For example, a media container used for HDR can carry more saturated colors than can be shown on most wide color gamut displays, therefore, the media container is larger than the displays’ native color gamut and that content needs to be precisely mapped to the capabilities of the display.
HDR itself is managed in a number of different ways – HDR 10, HDR 10+ and Dolby Vision. Now these all work in similar ways, by increasing the contrast in colors, however, they do behave differently. HDR 10 is a popular and simple HDR format developed by the consumer technology association. The HDR 10 standard sends static metadata to the display enabling the display to calibrate its picture based on overall characteristics of the video string. HDR 10 aims to produce around a thousand nits of peak brightness. HDR 10+ was developed by Samsung and Amazon, and HDR 10+ works differently to HDR 10, by sending dynamic metadata allowing TVs to calibrate the best possible picture frame by frame. This makes the picture look more realistic and the format supports brightness values up to 4,000 nits. HDR 10 and HDR 10+ both send their metadata from source to display using data passing mechanisms in the HDMI standard, called “AV infoframes”, and it’s these two formats which are commonly used by film production companies, video game manufacturers and TV show production companies.
Finally, Dolby Vision is a proprietary HDR standard introduced by Dolby Laboratories. Like HDR 10+, Dolby Vision also sends dynamic metadata to the TV and further supports 12-bit color depth, which is 4,096 shades of each primary color. Dolby Vision aims at reproducing 10,000 nits of peak brightness, far more than that which is offered by its counterpart HDR standards suggesting that TVs with Dolby Vision can produce 10 times the amount of light than HDR 10, although there are currently very few displays that can actually support a 10,000-nit brightness value.
Matt Dodd (11:06):
Dolby Vision passes its metadata by encoding and hiding the metadata inside the image itself. In HDR 10, HDR 10+ and standard Dolby Vision, the data transmitted from the source to the display is identical regardless of the display and it’s a display using metadata, which adjusts the image to suit itself. The latest iteration of Dolby Vision is called “low latency mode”and in this mode, the source reads out the detailed HDR capabilities of the display via EDID and then passes a customized signal built specifically for that display. The same source would pass different signals when connected to different displays because the TV is receiving a customized signal. There’s no need for metadata in Dolby Vision low latency mode, so none is passed.
So the next time you see the phrase HDR compatible, you’ll be able to appreciate just how much work goes into making your display provide you with images which aim to meet the high-peak-brightness demands of your eyes. The idea of a source customizing its data for a specific display brings up specific challenges for video distribution systems and we’ll learn more about that in our next HDR installment.
Justin Kennington (13:01):
Another great segment, Matt, thank you for breaking down HDR in such a simple way that I hope will help people out there understand the flavors a little bit better.
Matt Dodd (13:10):
Yeah, absolutely. Throughout the course of today’s show, we’re going to help you to add to that explanation. I think Steph’s going to be extremely powerful in helping get that message across.
Justin Kennington (14:45):
That is Stéphane Tremblay, CTO of AptoVision products for Semtech, one of the steering members of the SDVoE Alliance. Stéphane was really the inventor of the technology that became SDVoE so if we want to have a person on the show able to talk about moving high resolution video data over Ethernet networks, there’s really no one better qualified. I’ve known Steph for 10 years now since AptoVision first started trying to sell their products to me and then eventually I joined Stéphane’s team and helped to found this Alliance. So Steph has been there from the very beginning of Video over IP and there’s no better guest to talk about the bits and bytes of HDR.
Matt Dodd (16:31):
Steph, we’re here to talk about HDR and one of the first things I want to throw at you is: “Why does HDR matter in pro AV?”.
Stéphane Tremblay (17:22):
It matters to get quality. Years ago, TV vendors including organizations like the IRT and EBU had to figure out how to improve TVs, how to get better images. Is it by going higher resolution, higher frame rate, wider color range, or higher dynamic range? The answer is, simply put, HDR. This is the way to get deeper color, blacker blacks, whiter whites and this is what they have done. I’ve been going to tradeshows for close to 20 years now and this is the most impressive thing. The first time I saw a real HDR display with real HDR content, it was stunning.
Justin Kennington (18:22):
It was the same story for me. I remember seeing it in some private technicolor backroom suite and they showed it to me on this hot-rodded flat panel display that they built custom to do HDR for the first time and I was like, “Wow, this is special. It’s so different than just increasing resolution.” So why is that Steph? Why is changing the quality of the pixel so much more important today than just increasing resolution on its own, which is the path we’ve been on for decades, I guess?
Stéphane Tremblay (18:55):
Well, first, there’s a diminishing return. I was looking at the numbers for 8K, for example. If you want to appreciate an 8K display you have to get a horizontal viewing of 100 degrees. That means from the angle you describe by looking at the screen from left to right has to be 100 degrees. In my living room, I don’t even have 30 degrees, so yeah, having more pixels, if you’re really close to the screen, you might see a difference but other than that, you need better pixels and not more.
Justin Kennington (19:39):
Matt mentioned in the education piece, the human eye being able to perceive 10,000 nits instead of the 100 that TV was putting out.
Stéphane Tremblay (20:00):
Increasing the brightness is the main thing and the real HDR standard for that is Dolby Vision that aims at 10,000 nits. So everything under the hood is capable of doing 10,000 nits, but most of the content today is mastered at 4,000 nits. They don’t pass that 4,000 nits for many reasons, one of them is no display today can do 10,000 nits.
Justin Kennington (20:34):
You touched on something that I wanted to jump to next, which is Dolby Vision aiming for 10,000 nits. Is Dolby winning that standards battle and what’s the latest from them? Are they doing new things or is that standard staying where it is? What’s going on with Dolby Vision?
Stéphane Tremblay (20:54):
Dolby Vision is by far the best standard for HDR. Just to put things in perspective, they are, from the get-go, 10,000 nits ready, 12 bit, whereas HDR 10 and HDR 10+ are only 10 bit. They are dynamic metadata, so it’s really the more advanced standard, in fact all the demos I’ve seen make it the better standard. So that’s number one.
Matt Dodd (21:27):
Just to press pause there a second, guys, 12 bit, 10 bit, there’s tons of content here in SDVoE Academy after this show, you can head in and start to really remind yourselves or learn about these different bit rates. It’s important and they’re all explained. Sorry, Stéphane, carry on.
Stéphane Tremblay (21:47):
I was about to say Dolby Vision has lots of traction with most of the TV vendors, despite it being a payed standard. HDR 10 and HDR 10+ are free to use, but the TV vendors trying to do Dolby Vision will have to pay a royalty to Dolby. In return Dolby makes sure with, let’s say Sony, that their TV is capable of rendering properly. Speaking of which, Dolby Vision is aiming at PQ as a transfer characteristic function for the electrons to the light and it’s really referral based. In other words, let’s say I put a 709 code then it means 100 bits, so that’s the only standard with that type of fidelity.
Matt Dodd (22:48):
So what we’re going to do here is take a sneak preview at the next SDVoE Academy course “Implications of HDR for Signal Distribution” because this really goes into detail of everything that Steph’s been saying.
SDVoE technology fully supports LLDV and over the past two years, all appliances claiming support for Dolby Vision are also supporting it. Furthermore, SDVoE is able to accommodate all the displays which are not even 4K by using advanced processing to scale the original image and display it in the correct format. In this example, we see just that – a 1080P TV, which needs to subscribe to the LLDV stream. By using the SDVoE fast switching mode, the image can be shown and while it’s never going to match the quality of the original source, once again, it’s perfectly viewable. Compared to standard Dolby Vision, there’s no more finicky data packing of metadata within the active picture pixel data making LLDV much friendlier to distribute. Even a non- Dolby Vision capable TV receiving an LLDV signal will correctly display that LLDV signal. In the case of standard Dolby Vision, image corruption was the result we were previously plagued by. Latency is another huge improvement of LLDV and AV synchronization between different TVs is now possible.
Justin Kennington (25:11):
I was just going to say, it’s worth touching on for people that term LLDV, we can hit it a little more in the aftershow. That’s low latency Dolby Vision. The course you saw here today was walking you through the basics of HDR but now what does it mean when it’s time to distribute this beyond one source to one display. Compare a simple home system to the large scale of pro AV and AV over IP. Check out the SDVoE Academy course “Implications of HDR for Signal Distribution” to get the background.
Matt Dodd (27:30):
We’re coming to the end of the show piece itself. JK, what’s happening on the next show?
Justin Kennington (28:09):
Two weeks from right now, you’re going to see Chris Chinnock, president of the 8K Association, come and tell us why we do need more pixels. Frankly, I think they’re both right, we need more pixels and we need better pixels and we’re going to see how those fit together in a show about 8K and how that fits in the world of pro AV and HDMI.
Matt Dodd (28:27):
Is he as clever as Steph?
Justin Kennington (28:30):
He’s clever in a different way, you’ll see, you’re going to like Chris, you really will. Stick around for the aftershow, which is very easy to do. All you have to do is stay right where you’re sitting and don’t touch anything. We’ll be back with you with Steph for your questions live after the credits.
Matt Dodd (28:45):
Brilliant. Let’s go get changed JK, we’ll see you in a minute. Bye-bye thank you again.