4K video may wow vidiots, but content creators see pitfalls
Skyrockets studio storage costs, strangles editing bandwidth
GTC 2013 Don't expect ultra-high-resolution 4K video to be broadcast onto your living room wall anytime soon. According to the people responsible to building the equipment that can capture, edit, and encode 4K, there are a number of hurdles to overcome – not the least being storage requirements on the production side.
Sure, 4K displays were one of the stars of this January's Consumer Electronics Show, but that gadget orgy – as its name implies – is decidedly on the consumption side. On Monday at Nvidia's GPU Technology Conference in San José, California, a panel of digital-video smarties discussed some of the challenges faced by 4K-video content creators.
One of those panelists was Shailendra Mathur, chief video architect at digital audio and video technology powerhouse Avid. The film industry, he said, which has begun working with digital video cameras shooting in RAW formats – essentially pure, unmediated data pulled directly from the camera's image sensors – has discovered that storage requirements for working in 4K are gargantuan, and therefore expensive.
RAW digital-video content must be stored in order for it to be recalled when needed. After it's used to create slimmed down, compressed versions of itself for use in editing, the RAW content is needed when its chopped up and assembled into a final production version. Problem: that RAW content takes up enormous amounts of space — and, of course, requires additional storage space for backup.
"Everybody's extremely cost conscious," Mathur said, referring to movie studios. "How do you serve up storage with some of the RAW imagery, especially in recent productions with terabytes of information coming in every day, and more?"
Those RAW files are expensive to be kept on hand, but if a content creator can afford all that storage it's worth it. "You're getting a lot out of it in terms of the color information – the color gamuts are larger, and of course the resolution is superior."
And then there's the other advantage of 4K video to a content creator: the fact that if you're going to downsample the image to HD anyway, you have a larger image canvas to work with. You can, for example, pan and zoom within the 4K image and still have plenty of resolution left over for full 1080p video.
However, all those extra pixels combined with a wide, 32-bit color gamut not only fill up storage systems more quickly than does HD video, but there are I/O problems to conquer as well as content creators push 4K frames through suddenly inadequate pipes in their video-editing systems – and lets not even get started on high-framerate video, such as the industry's new 48fps and 60fps darlings.
Of course, when all these high-quality digital video files are transcoded and downsampled to, say, HD video, they won't have all the advantages of 4K RAW – but the better-quality images you start with and the more headroom in those images, the more you have to work with and the better quality your resulting HD images should be.
Not that what you see on your home television – or, for that matter, on any consumer-level display you see on the CES show floor – will be anywhere near as rich as anything you would see if you were working with, say, a full-on 16- or 32-bit color development system such as Avid provides to its pro customers.
The panel's moderator, Nvidia senior product manager in advanced technology Andrew Page, described some of the color step-down pain that digital video content creators face. At the top of the distribution heap, Page said, is a 12-bit DCI-compliant digital cinema projector, which he describe as having good dynamic range – and with ongoing developments in laser illumination, that DCI dynamic range is getting even wider.
Color quality goes quickly downhill from there, however. "Your average home TV?" Page asked, rhetorically. "Not quite so good." Most home HD displays have 6-bit color, he said. "You're lucky if you have 8-bit color – and by the time you get to tablet devices, with the stream compression and usually relatively a limited gamut, then the dynamic range of those devices is even less."
It's all about trade-off, Page said. "It's certainly an interesting topic on the mismatch between the acquisition side and the display side, and keeping an appropriate balance through that." ®