The challenge
I'm deaf. I'm also stupid, ignorant and (according to my wife) practically blind. Well, she asks me "are you blind or something?" enough that I think I need my eye prescription checked again.
But the deaf bit is official, kind of. I've long had tinnitus, with a quiet high-pitched noise in the background. It is only intrusive in a very quiet environment, and you quickly come to ignore it. I suspect I also managed to damage my hearing with a bit too much Walkman listening whilst falling asleep at night or working in grotty student jobs as a youngster. Last night, prompted by a BBC News article, I called up the new service offered by charity RNID to test your hearing. They play various 3-digit spoken sequences with varying amounts of background noise, which you enter on your phone keypad. My score, unsurprisingly, came out as "below normal", although not in the bottom category of "go see your doctor, quick".
(I guess RNID once stood for Royal National Institute for the Deaf, but royalty has lost its lustre, nationalism is passé, and the deaf are merely comprehension challenged.)
Note that the service is landline only (because of sound quality issues on cellular), and may not be callable from outside the UK (let me know if it is). It'll cost you money because it is an 0845 number. The "Local rates apply" stuff is a scam I've written about before. It should really say "Premium call rates apply" if absolutely truthful.
Nonetheless, it highlights the degree of challenge VoIP faces if it is ever truly to supplant the PSTN. Let's examine the hurdles we need to jump:
- Distribution. The service has to be more-or-less pervasively adopted amongst the population, including the non-geek elderly and disadvantaged liable to use this service. This kind of awareness campaign relies on us having a minimum level of social and technical homogeneity.
- Technical quality. The call has to meet a minimum level of quality, with a high degree of certainty that the threshold is crossed. Note that the absolute level of quality doesn't need to be exceptionally high, just the certainty of what you're getting.
- Money. They are using the revenue share from the 0845 number to fund the exercise. Money has to be able to flow in the opposite direction to some of the bits. Furthermore, there was no need to authorise or authenticate yourself to enable this flow, or provide details of any payment method. The regulatory environment has taken care of it so that "low rate premium numbers" such as this are callable as normal calls with no special arrangements.
- User experience. It just worked. The only thing the user needs to know is a string of numbers to enter to call the service. That string of numbers needs to be easily communicated via many different media, since -- guess what -- the callers may be partially deaf. There was no need to provision anything, register, or navigate any mazes.
- Devices. I used my shiny new DECT cordless phone at home. (No, didn't buy a WiFi model, I needed one quick from the local store because the old phone broke. Those wishing to take pity and mail me a fancy phone, address for Xmas gifts available on request.) This meant I was able to move around to an appropriate environment to do the test.
How does VoIP stack up against this challenge? I'll acknowledge in advance that this example (and the whole purpose of this article) is to pick the most extreme possible case that favours the PSTN.
- Distribution. Fuggedaboudit. For all its blinding success, Skype as the poster girl of advanced VoIP service remains at best in the single digits of total population penetration in large countries. Broadband penetration is nothing like high enough to count, let alone adoption of any specific PSTN-over-IP service or "Voice 2.0" successor.
- Technical quality. VoIP end-points generally fail to fully deliver on the "end-to-end" idea, with high reliance on smart nodes scattered around the network to proxy signals and gateway media. As such, they not only offer no QoS guarantees, but fail to support any kind of introspection where the other end point can query "what kind of signal are you getting?". Just being "normally better than the PSTN" isn't enough! The PSTN offers one, narrowly defined, voice quality (at least over TDM). VoIP offers a wider spectrum of outcomes. In Skype 2.0 you can tick the "Display technical call info" box (Tools -> Options -> Advanced) and get a ton of detail on jitter and packet loss pop up on your PC. But this service could not access that data, and tell you "sorry, the call quality you are experiencing is insifficient to support this service".
- Money. Bzzzt. No banana. No next-gen VoIP apps support payments, although eBay with Paypal is only an inch away from true integration with Skype. So far, the only way this can be done is to piggy-back on the PSTN, or use premium SMS. Messy.
- User experience. Oops. It's OK on a PC or Mac with Skype, most of the time. Let's just say that this Xmas day, the Queen's speech at 3pm will be losing audience share to "Son, while you're here could you take a quick look at my PC...". Most other PC services suck like a galactic black hole at basic stuff like download and install. And the more esoteric stuff like pure SIP totally lacks packaging for the everyday user.
- Devices. Carting a laptop, power cord, ethernet cable, and headset about doesn't exactly compare. People need to have an appropriate range of devices available. The WiFi handsets are still klunky, and can be a nightmare to provision. Do you still remember the password to your home WiFi router? QoS within the home network -- where it really matters -- is largely missing.
Now, we can imagine a "Voice 2.0" experience that does exceed that of the PSTN here. For example, the call begins with TV presenter Eammon Holmes slowly reading out instructions. In Voice 2.0, this would be multi-modal (where a display exists) showing the text as spoken. You would be able to review them again. An Eammon avatar might be lip-syncing too, with the text personalised to your name, to keep granny swooning at her daytime TV hero. In my case, I decided the traffic outside was too noisy for a hearing test, and wandered into the kitchen. In doing so, I missed a bit of the prelude. Then 2/3 of the way through the fridge kicked in. No pause button here, folks!
None of this is a criticism of what's been achieved so far. The PSTN isn't perfect either -- call from a mobile and the test will start regardless, despite the insufficent audio quality -- and you won't know how much you're being dinged by your mobile operator for calling an 0845 number. But I hope it shows just how long the journey may be.
I look forward to seeing some real progress on it in 2006!
Posted by Martin Geddes at 1:22 PM
TrackBack URL for this entry:
http://www.telepocalypse.net/cgi-sys/cgiwrap/mgeddes/MT/mt-tb.cgi/640