An 802.11n And Powerline Mess: High Definition Video Distress

This is an analysis I really wish I didn’t have to write. I’m shaking my head, in fact, as I type these words. But my past testing of wireless and wired networking technologies (see here for an important addendum) begs for an application-specific follow-up, especially given my CES 2008 critique of the premature rush-to-market that IEEE 802.11n experienced.

As I’ve mentioned in the past, I use a Dell XPS M1330 laptop running Windows Vista Ultimate as my Media Center system for time-shifting ATSC broadcasts from CBS, NBC and Fox, as well as a PlayOn server (primarily for Hulu content). I have two Xbox 360s acting as Media Center Extenders, one in the living room and the other in the bedroom. I’ve been connecting the Xbox 360s to my Apple Airport Extreme 802.11n router via HomePlug AV adapters, because Microsoft doesn’t (yet, at least) make an 802.11n adapter for the console (although alternatives are available), and because Microsoft’s 802.11g adapter is both ridiculously expensive and delivers insufficient bandwidth to reliably stream high-definition video content.

I also historically have tethered the laptop to the router via HomePlug AV, even though the XPS M1330 contains a built-in 802.11n transceiver. That’s because high-definition video streaming performance from the laptop to the router over Wi-Fi has historically been horrendous, even if I consciously choose a lower-grade (therefore lower bitrate) recording quality setting (Windows Media Center trans-codes an incoming A/V stream into the proprietary DVR-MS format as part of the recording process, at four different quality options… Fair, Good, Better, and Best). Below are example screenshots of the two Windows Media Center ‘Network Performance Tuner’ application "views’ (accessible from both the PC and Extender), whose attributes I’ll refer to through the remainder of this post:

I generally find the ‘graph’ view to be the more useful of the two, as it provides a versus-time plot of bandwidth variances. Its lower ‘acceptable for TV’ threshold implies standard-definition video, while ‘HDTV’ is self-explanatory. The bar graph display, conversely, reflects varying bandwidth over time, but a single-point-in-time screenshot isn’t terribly useful in a discussion such as this one.

Until a few days ago, whenever I tried to connect the laptop to the router over 802.11n, even with the Xbox 360 router-tethered over either high-speed HomePlug AV or CAT5, I’d end up with bandwidth measured below the ‘acceptable for TV’ threshold. Since the XPS M1330 was fairly new (bought by me last July) and since neither Windows Update nor Dell’s Support Center utility indicated that a newer driver for the Dell Wireless 1505 Draft 802.11n WLAN Mini-Card was available, I assumed I was dealing with a vendor-to-vendor incompatibility regarding the still-unratified IEEE standard.

Yet, HomePlug AV wasn’t a panacea. Brians’ Brain veterans will know that I’ve long struggled with a variety of powerline networking issues; refusals to pass certain protocols, router incompatibilities, insurmountable bit-leaps across 220V phases at the circuit breaker box, and varying performance with ambient temperature, power grid impedance and noise. To that latter point, and in spite of abundant educational ’scars’ from solved past problems (exemplified by, for example, AC noise filters I installed at my refrigerator and whole-house furnace, the two primary motors in this abode), I still struggled with erratic and oft-insufficient network speed. In case you’re wondering, DS2-based UPA adapters were even worse; several times to date, I’ve exhaustively (and ultimately unsuccessfully) tested multiple generations’ worth of units.

Best-case, the test results with both the Xbox 360 and laptop HomePlugAV-tethered would look something like this:

With available UDP speed hovering right at the HDTV line, Windows Media Center was generally able to manage…by buffering a big enough stream to smooth out the bumps and, when necessary, by resorting to more drastic measures such as frame-drop and down-resolution surgery. But past a certain point, I’d notice more egregious errors; visible image artifacts, audio and/or video dropouts, and a temporary loss of audio/video sync. Plus, most every time I lost premises power, the six HomePlug AV adapters here wouldn’t come back up in sync, and the incommunicado unit(s) would subsequently degrade the entire powerline network. I’d be forced to unplug them all, then plug them back in one-by-one to restore full functionality. Ironically, right before I started writing this piece earlier today, a reader posted a relevant comment to a past powerline piece of mine, which I’ll excerpt below:

The major issue with Powerline technology is that it requires an understanding of not only networking, but also an understanding of electrical principals and properties. These are typically above most end users abilities. Couple this with the fact that the majority of the people purchasing these devices do not know how to do a proper site survey prior to implementation and you have a recipe for disaster and disappointment.

While I can’t disagree with anything this person has said, I do think that it’s both an over-simplification and unreasonable expectation of the situation. For one thing, any technology that requires a ‘proper site survey prior to implementation’ is fundamentally incompatible with most consumers’ capabilities and ease-of-installation expectations. Secondly, while this individual may have ‘personally performed numerous installations of Powerline networking with products from several vendors and have not had an issue to date’, my experiences with powerline suggest that a lack of problems at initial setup is no guarantee of continued smooth sailing down the road. What happens if the installer’s customer swaps out routers later, or appends a gadget whose motor injects adverse noise into the mains, or plugs in a widget that adversely affects the power grid’s impedance? If the installation occurs in winter, what happens the following summer when air conditioner use is pervasive throughout the neighborhood? Or what happens if the customer moves a powerline adapter from one outlet to another, thereby incurring a phase jump that didn’t previously exist?

Continue reading with Part Two of this post, ‘An 802.11n And Powerline Mess: A Hybrid Approach May Be Best‘…