Next-generation home networks must provide adequate reach and performance to support multiple streams of HD video, while also delivering the lowest cost of deployment and an optimal viewing experience. As service providers gravitate to wireless technologies as the easiest, most economical way to deploy video-networking services, there are a number of design issues to consider.
With the arrival larger and higher-resolution flat-panel TVs, a 30Mbps video-encoding rate is needed across many as three or four TVs plus game consoles, and up to 60Mbps is required when including wireless gaming with low latency response. Meanwhile, some HD compressed sources may have much higher peak data rates. Service providers and content providers should plan ahead so that they don't have to re-configure their offerings every few years, which means it would be reasonable to allocate at least a sustained 120 Mbps of compressed HD video rates in the downstream direction (and perhaps higher peak rates) from a variety sources such as STBs, RGWs and NAS boxes.
Available Multiple Input Multiple Output (MIMO) wireless architectures vary widely in their ability to deliver this level of bandwidth in large-scale service roll-outs, while optimizing connection strength and reliability for consumer entertainment. There is a significant performance delta between older 2x2 and 3x3 MIMO solutions and the latest generation 4x4 systems with dynamic beamforming and low density parity check (LDPC), and this has significant implications for video networking performance and viewing satisfaction.
Understanding Viewing Satisfaction
Determining "viewing satisfaction" is a fairly complex and involved process with both objective and subjective components.
Objective components include long-term averages of statistical packet error measurements which, for example, can define a bad connection (low signal-to-noise ratio, or SNR). In this case, a typical consumer may observe a certain number of macro blocks randomly appearing on a digital TV screen. It is possible to quantify and evaluate these connection-quality scenarios using measurements of long-term averages for Packet Error Rate (PER). Enabling consumers to continuously view high-quality, glitch-free images requires a very low PER on the order of .01 percent to .001 percent.
Subjective components involve time-dependent, event-driven statistical system properties. These errors are fairly hard to simulate or measure directly but their combined effects can be seen in the form of residual errors that appear on the screen long enough for the eye to see them -- degraded picture sharpness, color leakage, reduced contrast ratio, etc. This is normally measured by sending a specific set of patterns to the TV so that trained observers can view it, assess image quality, and collectively generate a Mean Opinion Score (MOS) for image quality. The methods used in subjective testing are detailed in various ITU-R recommendations.
There are many parameters that could influence subjective testing results, but on the wireless physical-layer (PHY) side of the system, these are limited to the system's linear dynamic range and its SNR margin. Both parameters are affected when the system is subjected to random interfering events occurring in the air.
The Drive to Higher Data Rates and Bandwidth
Data rate or total capacity achieved in a home environment is a product of time and bandwidth per user. The growth in TV size, contrast ratio and color resolutions is driving demand for higher bandwidth, and higher dynamic range is also required in order to transfer quality HD content. While earlier flat panels displayed video content at a 1080i-30 resolution, a large-scale shift has occurred toward 1080P-60 resolution. At the same time, frame rates of 120/240 frames per second (fps) are increasingly common and the movement toward 3D TV is already signaling a shift toward more bandwidth.
Even more bandwidth is needed for large screens. Viewing very high-quality HD content such as the Super Bowl on a 72-inch TV screen might require 30 Mbps or more of compressed H.264 performance. The result: while 8 Mbps to 12 Mbps performance for transporting compressed HD content may previously have been sufficient for a good viewing experience, speeds of 30 Mbps or higher will soon be required for a satisfactory viewing experience.
Video gaming drives even higher bandwidth requirements. Wireless technology must support low-latency (sub-10 msec) video encoding/decoding technology by delivering as much as 60Mbps for a fairly large screen.
Importance of MIMO Antenna Order
A wireless MIMO channel is a multipath system, which means multiple reflections create many paths between antennas. The cumulative power of those multiple reflections can be harnessed to significantly improve wireless performance.
On the transmit side, any single antenna can transmit signals, which can then be deflected and diffracted into many radio wave branches. On the receive side, each antenna can be the recipient of many of these radio wave branches, in which case each antenna can be viewed as an independent "observer" that performs independent sampling, improving signal-to-noise ratio (SNR). Within this multi-path environment, a wireless channel can be classified in terms of channel rank, or how many independent paths it can support between its transmit and receive antennas. The higher the channel rank (meaning more independent paths), the more the antennas can make use of these independent observation and sampling opportunities that improve SNR.
As the number of antennas increase, we can assign at least one spatial stream per antenna . Each spatial stream can carry a large amount of data-in the 802.11n protocol, the maximum amount is 150Mbps in a 40 MHz bandwidth. As one would expect, not every spatial stream can carry 150 Mbps, and not every spatial stream has equal capacity or is truly independent. As such, sometimes the use of extra antennas may not yield additional throughput. Figure 1 shows how two spatial streams are multiplexed over an array of 4 x 4 antennas in order to utilize the capacity offered with only two available independent paths. The use of extra antennas (i.e., two antennas per spatial stream) benefits the signal reliability by an average factor of two and in some cases even more. At the receiver, reliability can be improved by employing an algorithm known as Maximal Ratio Combining (MRC), which optimizes the received SNR by enabling all four antennas to be used to recover the two spatial streams.
FIGURE 1 • 4x4 MIMO with multiplexing of two spatial streams
In order to achieve the maximum raw data rate of 150 Mbps per spatial stream, sufficient SNR must be available. Adequate SNR enables the receiver to decode the incoming signal, where each spatial stream has a 64 quadrature amplitude modulation (QAM) index. Support of a lower modulation index is required (i.e., 16QAM or QPSK) when the available SNR at the receiver is lower.
The ability to increase the number of antennas per spatial stream is a major advantage of higher-order MIMO systems. In addition to committing two antennas to each of the spatial streams required for home-networking models, systems can employ other reliability-enhancing techniques, such as allowing unequal modulation for each spatial stream (see Fig. 2). Obviously, the ability to leverage the benefits of spatial streams increases as the number of antennas is increased.
FIGURE 2 • This figure illustrates 4 x 4 MIMO with 4 spatial stream multiplexing. In this example, due to this spatial signature, the spatial streams are not equally modulated. It is shown that streams 1 and 3 are QPSK modulated (per subcarrier -OFDM) and streams 2 and 4 are 64QAM modulated (per subcarrier -OFDM).