I have just returned from the 2013 Production Technology Seminar hosted by the European Broadcasting Union (EBU). The Union is an alliance of public service media organizations, with the consequence that its members are willing to lift the lid on their workflows in a way that many commercial organizations can prove more reluctant. The Union is also able to undertake research into broadcast technology, and this is what particularly interested me.
Both broadcasters and viewers are being pressured by CE vendors to embrace 4K in order to sell new displays to everyone who has bought an HDTV, possibly with 3-D support — ongoing obsolescence. During the seminar, a viewing test was conducted that compared video sequences in the UHD-1 format (3840 x 2160 pixels, 50fps, progressive) with HD — 1080i, 1080p and 720p, all uncompressed. The tests used the recommended viewing distance of 1.3 times screen height and the typical viewing distance of 2.7m. I would call the assembled viewers experts, yet the tests showed the advantage of 4K over HD was marginal at a normal viewing distance, and measurable, but not huge, at the 1.3 times screen height viewing distance. Future tests will be conducted with compressed material, as artifacts may well change the results.
It is challenging to conduct proper tests right now. Most engineers outside the film community accept that higher static resolution needs a higher dynamic resolution or frame refresh rate. In other words, motion artifacts look bad at 4K and worse at 8K. This is why NHK has chosen 120fps for 8K Super Hi-Vision.
But can you find a 4K display that operates a frame rates above 30? HDMI 1.4 does not even support 60fps 4K. That awaits ratification of HDMI 2.0, expected later this year. As a consequence, displays have to be kludged with tiled inputs (four quadrants) to display UHD-1 signals. Any consumer who rushes out and spends huge sums on a 4K display is going to be sorely disappointed when high-frame-rate video sources become available. This is definitely not a time to be an early adopter.
There are several issues beyond HFR, including the wider gamut specified in BT.2020 (the ITU standard for UHDTV) which has a wider color space than Rec. 709. UHDTV could also use a greater bit depth, possibly 10-bit rather than 8-bit coding, delivering higher dynamic range image (HDR). BT.2020 also allows constant or nonconstant luminance coding, and that is another big issue.
Broadcasters and studios have yet to figure out the commercial advantage of using HFR, HDR and higher resolutions like 4K, and issues like constant luminance coding, although more efficient to compress, will cause backwards-compatibility issues.
HFR, HDR and 4K all increase bit rates way more than any gain from HEVC over AVC.
The CE guys are leading the pack, but who will follow? 3-DTV seems to have fallen below their expectations, so 4K is the next focus. Ask your nontechnical neighbor: “Would you prefer constant luminance coding, wider gamut, 10-bit 4K pictures or social media apps on a second screen linked to a conventional HDTV?” I must presume no one will turn down improved picture quality at no extra cost for the service, but what will be the premium? I sense a lot more buzz about the release of the latest smartphone than a 4K OLED TV with 22.2 surround sound.
I believe consumer technology is getting ahead of the business models, much as mobile TV did back around 2006. I’m sure we will get there, but not this year.
I await the results of the projected viewing tests that the EBU is planning. They should cut through the hype of the CE marketeers. These same guys use 4K as it is a much bigger number than 1080 — no matter one is the picture width (rounded to one significant figure), and the other is the picture height.
—David Austerberry, editor