This is the seventh in a series of posts from the Hot Chips conference at Stanford University. The previous installments looked at technology and software, process technology, multicore designs, IBM's Power6 efforts, Vernor Vinge's keynote address, and Nvidia. Other CNET coverage may be found here. This is sort of an experiment for me; I usually prefer to have time to review my work before I publish it. If you see anything wrong, please leave a comment!
This session has two presentations--one from SiBeam describing wireless HDTV transmission for home use, the other from Broadcom on new 802.11n Wi-Fi technology.
The SiBeam presentation is easily summarized: It describes a chipset that sends uncompressed HDTV video over a 60GHz wireless link. Yup, that was 60GHz. You 5.8GHz cordless telephones, eat your hearts out.
SiBeam OmniLink60 chipset operates at 60GHz so there's enough bandwidth for uncompressed HDTV, which needs about 4 gigabits per second (for 1080p60 video plus 8-channel 192KHz audio). As long as the signal gets from the transmitter to the receiver, you get every bit of the original program.
The 60GHz band provides 7GHz of RF bandwidth, and effective transmit power can be as high as 8 watts. SiBeam uses a channel width of 2.5GHz, making it relatively easy to squeeze in up to 4Gbps (by way of comparison, ATSC digital television fits 19.4Mbps into 6MHz). There's also a 40Mbps digital back-channel so the video receiver can talk to the transmitter for configuration and security purposes.
The presentation went into all the details of SiBeam's implementation, but I'll just skip to the end--this technology has been demonstrated and SiBeam is preparing to bring it to market. The bottom line is that this technology enables truly wireless home entertainment systems--the TV, stereo, HD DVD or Blu-ray player, cable box, camcorder, etc., can all communicate without any video or audio cables. The technology doesn't require true line-of-sight transmission; the radio signal can bounce around the room on its way to the receiver.
Downsides? Well, the system works over ranges of about 10 meters; I gather from the technology details that longer ranges are possible in time. The SiBeam chip, antennae, and other necessary components are a lot more expensive than copper wire--that is, unless you're one of those people who insist on paying ridiculous amounts of money for copper wire. (I guess SiBeam will have to sell to those people; maybe they can make some economic argument after all.)
The Broadcom presentation described how 802.11n Wi-Fi using MIMO technology (multiple-input multiple-output communications using multiple antennae) will support 600Mbps networking by next year. This new standard uses more spectrum--up to 40MHz channels--and more complex encoding schemes that fit over 15 bits per second per hertz of bandwidth (compare that to the 3+ bits/s/hz for ATSC broadcasting, or the 1.6 bits/s/hz for the SiBeam product).
And that's about all I think I'll explain about that--the presentation is mostly about circuit diagrams and spectrum analysis. I should add, however, that the specific chip Broadcom was describing here delivers only up to about 200Mbps of real throughput.
Up next is the conference's second keynote address, this one from Phil Hester of AMD and titled "Multicore and Beyond: Evolving the x86 Architecture". Stay tuned...