Today's high-definition TV and video can look great, but there's always room for improvement. New, higher-fidelity image schemes are beginning to gain traction these days, including 4K, wide color gamut, and high dynamic range.
The idea behind all of these methods is to allow video shown on current and future TVs to look more like reality, whether through improved detail and sharpness (called); a broader range of colors that more closely matches what the human eye can perceive (wide color gamut, or ), or brighter highlights and darker blacks presented on-screen at the same time (high dynamic range, or ).
We already have some 4K TVs to display them. Standards to allow wide color and HDR in home video are being developed now. From demos we've seen, they have the potential to improve image quality much more than 4K can., along with
One issue, however, is that most of those movies and TV shows were mastered for standard dynamic range (SDR) and the narrower color range of HD. That's the issue(ITM) aims to address. It's a method to convert standard dynamic range content (i.e., pretty much everything) into a pseudo-HDR.
When I first heard about ITM, I was skeptical. There are no good examples of converted content that looks remotely as good as the real thing. 2D movies converted to 3D look worse () and rarely looks as sharp as true HD or 4K. For an example of the latter, just turn your HDTV to a standard-definition channel.
So how can something with standard dynamic range match the incredible pop of HDR? I went to Technicolor's labs in Hollywood facilities to find out.
Technicolor's sharply angled and modern steel-and-glass building occupies the Sunset side of the legendary Sunset Gower Studios. Security checks, valet parking and name badges were all typical in this company town, and offered no surprises for "Jeffery Morrison."
Inside looked much like any office building, perhaps with sturdier (soundproof) doors and far more movie posters and blown-up stills than average. After a brief meet-and-greet, I was shepherded into one of the labs by Mark Turner, vice president of partnership relations and business development, and Josh Limor, director of product development for industry video technology. Both enthusiastic about HDR and knowledgeable. I meet plenty of people in technology who are one or the other, but rarely both.
The room (above) was a pretty typical post-production lab. There were 30-inch Sony OLED broadcast monitors that cost about $1,000 an inch; a wide control surface that looked like a mixing board but with more knobs and fewer sliders; and dominating the far end of the small room were two 65-inch SamsungTVs. I'm told they were set up as identically and correctly as possible.
Both Samsung TVs were running the same demo reel. It consisted of a series of clips, commissioned by Technicolor, and edited and processed in house. Each version of this reel was created from the same camera files. Each version was "tuned" to look as good as possible in its specific color space and dynamic range.
The left JS9500 was running a standard image, with theIt was essentially HD-quality, and looked good.
The right JS9500 showed a split screen image. The right half was true HDR content displayed in the. It looks fantastic, as HDR can, with rich cyan skies, deep red bathing suits, bright highlights on metal that really pop and more.
The left half of the split screen was way more interesting to me. It showed the standard dynamic range image, run through Technicolor's ITM process, to create a pseudo-HDR image from the SDR content.
Essentially, that means upconverting SDR to HDR.
It looked...really good, actually. Much better than the SDR version. Did the real HDR look better? Absolutely. Depending on the shot, at least 20 to 30 percent better. But the upconverted version looked so good, that if that was the only image shown, I doubt anyone would have guessed it was "fake" HDR. Which is, of course, the point.
What's goin' on
Of course, I took what I was seeing with a big bag of salt -- I've been shown a lot of demos in my day, this is not my first rodeo drive. However, there are a few things that make this particular demo more promising than most.
The first thing to keep in mind is that recent and current digital video cameras have significantly greater dynamic range than is possible within the confines of the current HD standard. (So do most recent TVs, incidentally, whether they officially support HDR or not.)
During the initial mastering process for HD, the image shot by the director and cinematographer on those cameras is essentially "dumbed down" for delivery to your eyes via today's HD sources, namely Blu-ray, streaming or broadcast TV. During a remastering process, more of this visual information can be used to create an image with greater dynamic range.
The second is that ITM isn't a fully automated process. It's a plug-in for post-production color grading software (color grading is used to make sure the whole show or movie looks consistent and how the director wants it). So don't expect to see an "ITM" button on your next TV.
A colorist, ideally along with the original content creator (be it director, cinematographer or both), makes choices as to what to accentuate, what to hide, and how much color to bring out in the highlights, midtones and shadows. The goal is to get the image you see closer to what they intended you to see.
To show me this process, Technicolor's lab techs put the SDR/709 image on the left TV, then undid the work they did with the ITM for the image on the right. The images were identical, proving their claim from earlier about the TVs.
Then, using sliders and adjustments familiar to anyone who has used Lightroom or Photoshop, they were able to dial in color to various parts of the image. With more dynamic range to play with, for example, more detail can be brought out in both clouds and shadows.
Excitement and concerns
What do with "legacy" content is a concern during any video transition. Upconverting 50 years of NTSC (the video system or standard used in North America that's 30 frames per second) content was often done poorly, and making 2D into 3D was and is rarely done well.
The possibility of converting standard to HDR getting bungled is real and worrying, especially given the great possibilities of the "format."
What I saw gives me hope, though of course it was in a pristine environment as part of a demo. The potential for the conversion to be done correctly -- or at least not badly -- exists.
Will there be poorly converted HDR content? Of course, it's inevitable.
What Technicolor's demo showed me is that for content creators who care, there's a way to convert SDR content cost-effectively, while creating an end-product that looks better than SDR, if not quite as good as native HDR.
And really, that's what matters. The content that you want to look really good ("Game of Thrones" or "Sense8"for example), can look good. Probably even better than what you're seeing now.
Cameras are improving all the time, as are TVs. With HDR and WCG, the content is finally catching up. Technicolor's ITM, and other methods like it, are a promising step to help us bridge the gap between SDR and a wonderful HDR future.
Got a question for Geoff? First, check out all the other articles he's written on topics such as , , and more. Still have a question? Send him an email! He won't tell you what TV to buy, but he might use your letter in a future article. You can also send him a message on Twitter @TechWriterGeoff or Google+.