AT&T appears to have broken ranks with other US operators by trusting WiFi to deliver premium HD content over its U-verse IPTV service. Its announcement that U-verse will be running over WiFi using Cisco technology is not as dramatic as it sounds, since its customers already had WiFi enabling them to receive Standard Definition (SD) pictures on iPads and other wireless devices.
The move overturns the conventional wisdom that WiFi can only be trusted to deliver HD services within individual rooms and that it needs to be underpinned by some whole home wired backbone, which is usually coax via the MoCA standard in the US. Indeed just about all other US operators delivering multi-room TV services, including Verizon, DirecTV and Comcast, have opted for MOCA as their multi-room delivery medium, along with WiFi in most cases to reach wireless devices such as tablets.
But AT&T has not supported MoCA for a long time anyway, having run with the alternative HomePNA standard, which supports either coax or phone lines for HD services. Now AT&T is offering the alternative of a completely wireless home TV solution, by adding a WiFi bridging solution from Cisco’s Videoscape portfolio, comprising the ISB7005 wireless receiver and VEN401 Wireless Access Point, to deliver HD services. The wireless receiver can be connected to a TV anywhere in the home to deliver live TV channels and interactive services, and can also operate as a whole home HD DVR.
Such a wireless network brings obvious benefits. AT&T cites faster service activations and consumer self-installation with easy-to-use Wi-Fi kits. Having integrated Wi-Fi receivers also allows remote monitoring of the device’s performance via the network, and possibly the ability to do some troubleshooting.
But none of this will count unless the quality is good enough for HD on big screens. This is a different matter from tablets, not least because consumers expect near perfection on their primary screen when they are prepared to tolerate glitches when watching on a tablet or PC. AT&T is betting that a combination of improvements in the IEEE 802.11 WiFi standard and proprietary Cisco extensions do the job. It is true that recent enhancements in the 802.11n standard have crept up on us, and do address video quality as well as just raw bandwidth. The first job 802.11n did was improve average throughput, but initially with little impact on video quality because it failed to cope with the continuous changes in transmission conditions that bedevil the average home WiFi set up. These frequently cause temporary delays and jitter that freeze or pixilate pictures.
However more advanced modulation modes emerged within 802.11n and have now been implemented in a new generation of video grade WiFi devices driven by the tablet boom. These use a variety of smart antenna techniques to improve the robustness against variations in transmission to ensure greater consistency of video delivery. Essentially this is achieved through increased signal processing power applied intelligently to monitor channel condition continuously and adjust the radio output accordingly. The use of MIMO (Multiple Input Multiple Output) was the starting point and is an integral part of 802.11n, as well as for 3G and 4G cellular communications, using multiple antennae to create path diversity.
The more antennae used, the greater the path diversity and therefore the more robust a channel is against interference and fading. This is one reason we doubt whether AT&T’s U-verse will deliver reliable HD services across multiple homes ‘ we believe it uses a dual band 2×2 SoC (System on Chip) from Broadcom, so it creates two streams between each of the two transmitting and receiving antennae. The most advanced SoCs from Quantenna for example are 4×4, allowing four streams and many different beam patterns.
MIMO was inspired by the fact that radio signals tend to take multiple paths between a transmitter and receiver anyway because they bounce off various objects on the way. This used to be a problem, since the signals arrived at the receiver from different paths slightly out of phase with each other, causing loss of information through destructive interference. But this can be turned to constructive interference if the signals are tuned correctly.
MIMO embodies several techniques, one being spatial multiplexing, where a high bit rate channel is split into several lower bit rate ones transmitted from different antennae. This reduces the impact of interference, improving the signal to noise ratio, making video transmission more robust without increasing the aggregate bit rate. However MIMO can also increase capacity through another mode called diversity coding, where the full bit rate channel is transmitted in parallel from each of the transmitters. This exploits the fact that signal fading tends to occur independently in each of the parallel transmissions, so that if a particular bit of information is lost in one it is likely to be retained in another. This again makes the channel more robust and enables the bit rate to be stepped up further because the overall signal to noise ratio is increased at a given speed through the resilience against fading along any single path.
There is still potential to improve MIMO and this is happening, but the IEEE 802.11 committee decided that further innovations were also necessary to bring WiFi up to scratch for HD video. This led to the IEEE 802.1ac and then 802.1ad standards, the a referring to the use of the 5 GHz frequency band, rather than the 2.4 GHz widely used at present. This has pros and cons ‘ the 2.4 GHz band is more congested with a number of household devices such as microwave ovens and handheld phones in that zone sometimes causing interference. But 5 GHz signals, being around half the wavelength, have lower range because they are more readily absorbed on the route by walls and other obstacles.
But there is more space for wider channels around 5 GHz, and this is exploited in 802.ac and 802.ad by doubling the maximum channel bandwidth to 80 MHz, and possibly even 160 MHz as an option. For video the objective is not so much to achieve ultra-fast gigabit plus speeds, but to ensure that sufficient bit rates of say around 100 Mbps are delivered sustainably and robustly throughout the network area right to the edge. Research has shown that wider channels combined with intelligent beam forming does this most effectively, with increased average bit rates being almost an agreeable side effect.
Beam forming allows a multi-antenna transmitter to focus its energy towards the target device antennae, and achieves optimum efficiency for a MIMO channel. Instead of emitting energy indiscriminately in all directions, the transmitter estimates the direction that maximizes the power at the receiving antennae. The transmitter also manipulates the phase of the signal coming out of each antenna such that all the signals interfere constructively rather than destructively at the target device antennae ‘ they add rather than subtract from each other.
Beam forming is already supported in 802.11n, but with several alternative modes of operation, which has led to confusion among chipset vendors, and problems with interoperability. In 802.11ac and 802.11ad therefore, beam forming has been refined down to a single method involving use of a special frame sent by the transmitter to elicit feedback from the receiver. This communication allows the transmitting antennae to adjust their signal direction, as well as phase and amplitude of the radio waves, to optimize the received signal. This has immense potential to improve both robustness and bandwidth, and is now accepted as an essential component of a reliable WiFi HD video link. But it has yet to be refined fully, which again suggests that we are not yet ready for prime time multi-room HD delivery over WiFi.
One important new extension with IEEE 802.ac and 802.1ad is the Multi User (MU)-MIMO. This is not implemented in 802.11n, which applies single user MIMO across multiple antennae but only between two devices. In single user MIMO, the receiving device requires multiple antennae, as well as the sending device. But MU-MIMO allows the receiving device to have just one antenna, while a number of such devices can receive signals simultaneously. This is a huge advantage since it reduces the complexity and power requirements of receiving devices, while boosting the spectral efficiency of the overall network.
Indeed MU-MIMO is widely considered to represent almost as big a breakthrough as MIMO itself. The fundamental mode of operation is similar, in that while a single user MIMO transmitter obtains feedback for beam formation from multiple transmitters at a single receiving station, MU-MIMO gathers the feedback from multiple users. In effect, a directed beam is sent to each user simultaneously. This arrangement adds another layer of redundancy, since there is the potential to relay signals between users to provide further protection against problems on a particular path.
But MU-MIMO also adds another level of complexity to the transmitter, which now has to cope with the differing requirements of multiple users as well as make the decisions over beam formation to individual points. There will be greatly increased scope for optimisation by adjusting over the air bit rates, numbers of streams, and other parameters.
In this sense the emergence of MU-MIMO conforms with a continuous trend towards greater complexity for WiFi. The push for ever higher data rates to keep up with wired options, and now for ability to deliver HD video around the home, has led to technical contortions that mean multi-vendor interoperability will be as elusive as ever. It also means that scope for innovation and differentiation between SoC and adapter vendors will increase, and that WiFi equipment becomes less of a commodity. We will hear more claims like those made by Cisco and AT&T over the U-verse offering, that it is because of proprietary extensions satisfactory QoS has been delivered ‘ a claim we tend not to hear from vendors of MoCA kit, although we do get in the Powerline arena.
Meanwhile debates are still raging over some features of IEEE 802.1 ac and 802.1ad, like whether the wider beams are such a good idea after all because they limit the number of available channels, which could mean there are not enough of them to avoid interference in densely populated areas. But it is likely different deployment options will be allowed to cater for some of these differences. In any case devices supporting these new variants will not be available until 2013 at least. We will then be able to assess whether WiFi really can cut it on its own for multi-room HD video, even in Europe, where buildings are supposed to be more resistant to delivery of radio signals, being more likely to have thicker walls with metal in them. In such buildings use of the 5 GHz band can increase the risk of poor coverage, although this will be mitigated by MIMO and MU-MIMO. The overall message is that it is still far too early to declare WiFi fit for HD video.
All of which means that we suspect that the AT&T system will work in homes with minimal obstructions., but not where the signal has to content with ‘brick’ walls.
Cisco however took the opportunity to call it the industry’s first integrated wireless TV solution and made it clear that AT&T has only ordered this from Cisco, not from its other major supplier Motorola, and that the device is entirely of Cisco’s making. Of course it isn’t, because Cisco doesn’t make video capable wireless chips and these usually comes from Broadcom and we understand that the BRCM4717 SoC router is used in this device, a relatively inexpensive 2×2 dual-band 802.11n Wi-Fi radio. The device also has Ethernet and HomePNA wired interfaces which are found in all U-Verse set tops and DVRs.
We seriously doubt that a 2 x 2 MIMO chip is man enough to reach into people’s basements and perhaps Cisco should have gone to Quantenna for the chips which offers full 802.11n with 4 x 4 MIMO and which Motorola has signed up to use.