The indispensable element in the next-generation set-top box is a decoder chip capable of handling all the highly compressed video streams.
MADISON, Wis. — Is the semiconductor industry prepared to profit from the emerging 4K UHDTV market? If so, how ready is anyone?
As soon as consumers start looking for 4K Ultra HDTV (assuming they do), the indispensable element in the next-generation set-top box is a decoder chip capable of handling all the highly compressed video streams.
Compression will be carried out by using a complex but flexible "toolkit" developed by the High Efficiency Video Coding (HEVC) standard, a successor to H.264/MPEG-4 AVC (Advanced Video Coding).
Leading video chip companies such as Samsung, Broadcom, and ViXS Systems are all said to have developed their own HEVC chips. But these devices are still in early versions.
The nightmare secretly anticipated by many chip designers is the long and arduous task of verifying whether a video chip is truly capable of decoding every HEVC-encoded video stream.
HEVC, also known as H.265, is said to halve the bandwidth requirement for video, thus enabling migration to 4K video.
The issue is that there are so many variable options embedded in the standard's toolkit. Each encoder designer is free to use any tricks available in the toolkit to wring the best compression performance out of the standard.
That flexibility, however, poses a huge headache for decoder chip designers, since they need to make sure that their chip "understands" all the variables present in different HEVC-compressed video streams, explained Alan Scott, CEO of Argon Design.
HEVC in multi-core chips
To complicate the matter further, in HEVC, multi-core chips can encode streams in independent tiles, while hardware implementations may choose to minimize cache sizes by using wavefront encoding. However, decoders need to support all options.
Also, "HEVC has moved from a fixed partitioning scheme to a quad-tree decomposition scheme. This means there are orders of magnitude more variations of block sizes for coding/transform/prediction," Scott told me.
If you've lived long enough to remember when MPEG-1 and MPEG-2 standards were developed (in the 1990s), you also probably remember a test sequence video stream showing tulip fields against Dutch windmills on a screen. The video clip showed a beautiful landscape panned slowly. But pretty soon, you might have wearied of tulip fields. You might have seen tulips in your dreams.
In those days, you could validate your video decoders by looking at the screen, says Scott. But that visual practice won't cut it anymore, he believes, because "HEVC is many times more complex than MPEG-2."
He notes that the new standard "has so many options, variables and settings. It's hard to cope with all that complexity."