HDR video. High-density rendering? Nope. Happy Days Radio? Nope again. It’s high dynamic range video, and while it’s not a buzzword yet, it will be in the next couple of years. What is HDR video, and why will it be more important?
What is HDR?
If you are a photographer, you may be familiar with HDR photography. In still photography, you take (usually) three exposures of an image: one correctly exposed, one slightly underexposed (too dark), and one overexposed (too light), but not by too much. Then the three images are combined with special software to get the best highlights (brightest) and views of the deep shadows (darkest) in an image. Since video moves along at 24 or 30 (or more) frames per second, it gets more complex. At 30 frames per second, you’d need 90 frames combined into 30 frames every second. That’s a lot of processing. What combining the frames does is allow you to shoot a scene, say, from the inside of a fairly dark room to the outside. Normally, either you’d have the interior really dark and the outside exposed properly, or you’d have the inside looking good and the outside all smashed out. HDR is the solution.
What is HDR video?
HDR video all starts with dynamic range. High dynamic range is really an optical illusion. But it’s a good optical illusion. It fools the viewer into thinking the video they’re watching is more lifelike by a huge margin over today’s video displays.
Your eyes see a lot better than sensors in most video cameras. Your eye is also more sensitive to color than the best color displays. And there are two devices you need if your HDR video is going to work at all: an HDR video camera to shoot your video and an HDR video display to show it on.
Displays first: They’re inherently less complex than cameras. There are only a few HDR displays, and they are at the top end in price—not quite mainstream. To make matters even more complex, there are multiple standards for HDR video. It’s sort of like Betamax vs. VHS is happening again right now. There are also what are known as “locally lit” displays. Each pixel not only shows an RBG color, it can also be bright or dim. I’ve seen one of these kinds of displays in demo mode, and it is indeed spectacular. There are several in production and they are enormously expensive, as are all new things tech. A little 36-inch-or-so display still costs about $3,500.
The human eyeball has a dynamic range that can go as high as 20 f-stops (about 1 million to 1) contrast. That simply means we can see things that are really, really dim and we can see things that are really, really bright. This 1-million-to-1 ratio happens in different lighting conditions, so the human eye is also adaptive. Video cameras (even many very expensive cameras) have a dynamic range of 8 – 10 f-stops. That’s a huge difference. There are also video cameras that shoot in the “video raw” format. They can go as high as 14 or 15 f-stops. Explaining f-stops and dynamic range of light is complex, and this is an overview, so I’m not going to go into more depth about them in this article.
The problem
Since HDR video isn’t about resolution, it’s about something. Resolution is pixels, and past a certain point, you can’t see the added detail more pixels give. Really. It’s all about exposure—and more than one exposure at that. Still photography is easy: Just take three different shots at different exposures and combine them. It’s a piece of cake when you’re doing something like a landscape (and it’s not windy), but gets more difficult when you’re shooting a portrait, and impossible when you’re shooting action like car racing or even something slower. Video exacerbates the problem by having all those pesky frames get in the way: 24 or 30 of those frames each and every second. And to make matters more difficult, your subject is moving, even if the movement is nothing more than their lips flapping. So instead of still HDR exposures, in which we can take three shots in rapid succession, video is a wholly different story.
The event horizon is now
A few years ago, a few video camera makers introduced consumer-level HDR video cameras. Are they any good? It doesn’t matter in a sense. These cameras are all under $1,000. And that was in 2015. Here’s a little article about the Panasonic cameras for HDR video production. The cameras do make two exposures and “blend” them on the fly, but it’s not quite HDR. Other makers like Sony have similar cameras and use similar algorithms to create their HDR image. This is about as far as the HDR processing can go today. Real HDR is spectacular. On a large screen like a movie, we only look at one thing at a time on the screen. That one thing covers about 10 percent of the entire viewing area. DSLRs always seem to be behind video cameras in their features, but none of them can do HDR video yet. You can hack some Canon models to make HDR video, but because it’s a hack it voids any warranties. So if you’ve got an older Canon, maybe it’s worth it. The camera still needs to make two or three exposures pretty much simultaneously.
We’ll see a lot about HDR video in the next few years as the display prices begin to come down and cameras get better at it, and—and this is the big and—we start seeing content for HDR. I personally wouldn’t care if it’s HD video or 4K video. I’ll see it the same, and so will you. HDR changes almost everything we know about video and how we’ll perceive what we watch. Even better for eLearning. The more we can show our learners and the clearer it is, the more we can all learn.