Robin Hanson tells me this will be huge

by on July 14, 2004 at 5:22 am in Science | Permalink

Software Radio. I only wish I understood the idea:

As communications technology continues its rapid transition from analog to digital, more functions of contemporary radio systems are implemented in software - leading toward the software radio. A software radio is a radio whose channel modulation waveforms are defined in software. That is, waveforms are generated as sampled digital signals, converted from digital to analog via a wideband DAC and then possibly upconverted from IF to RF. The receiver, similarly, employs a wideband Analog to Digital Converter (ADC) that captures all of the channels of the software radio node. The receiver then extracts, downconverts and demodulates the channel waveform using software on a general purpose processor. Software radios employ a combination of techniques that include multi-band antennas and RF conversion; wideband ADC and Digital to Analog conversion (DAC); and the implementation of IF, baseband and bitstream processing functions in general purpose programmable processors. The resulting software-defined radio (or "software radio") in part extends the evolution of programmable hardware, increasing flexibility via increased programmability.

Got that?

It appears to hold the promise of eliminating radio interference (now viewed as simply a product of bad receivers) and making the electro-magnetic spectrum far less scarce. Your software radio or cell phone would search for a spot in the spectrum that is not being used and exploit it. The corporate owners of the spectrum will hate this one, but for how long can they stop it? Users will claim they are enjoying a free lunch, and will you win by suing them all? The more that problems of hardware become problems of software, the harder it is to regulate a sector.

Here is the best introduction I can find to the idea, from Here is an incomprehensible tekkie discussion on Slashdot.

Stay tuned, as they say...

Addendum: Perry Metzger writes me the following important and optimistic explanation...

There’s a big advantage that you missed, btw. Check out, for example, the “gnu radio” open source software radio project. Then consider this: the FCC can pass any rule that it likes saying you can’t copy digital television off the air or what have you, but anyone who decodes HDTV with a software defined radio (SDR) can just change the software to ignore the desires of the regulators, and open source SDRs are already here. This is yet another instance where general purpose computers make regulation very difficult.

Really, though, SDR isn’t that interesting until it is coupled with another major advance that is creeping up — phased array active antenna technologies. These have the capacity to literally make the amount of bandwidth available in radio unlimited. I do not use the word “literally” figuratively here — I mean “literally”. With phased arrays, you and I can broadcast as much as we want on exactly the same frequency, and so long as we’re slightly spatially separated no one will care.

Consider the way radio works now. Two stations can’t be broadcasting at the same frequency unless they are so far apart that receivers can only hear one of them at a time. Bring them too close together and the
primitive sensors we use for electromagnetic radiation will get both signals and become confused.

Now consider the situation with another sort of electromagnetic sensor — the human eye. There is no rule that says you and your friend can’t wear the same color shirt in the same room for fear that people will
be unable to distinguish the two of you. The eye has no trouble seeing multiple sources of the same electromagnetic frequency and distinguishing them. You can easily focus on any one of several sources of the same electromagnetic frequency even though you are receiving others.

What’s the difference between your eye and your FM radio? Entirely that your eye is directionally sensitive and the radio is not. The radio has much the same view of the world that a non-directional photocell with a color filter would have — it can tell that a particular frequency has arrived but not from where and it can’t distinguish multiple signals directionally. Imagine if you had to see the world that way.

Well, as it turns out, you don’t have to see the world that way. There’s a technology out there (the details aren’t important) called “phased arrays” that allows you to broadcast a signal very directionally (like shining a light at one and only one person) or to focus very directionally on a single signal (so that you can listen to
one FM broadcast on 99.5 and not hear another originating one mile away at the exact same frequency.)

This is not a theoretical technology — it has been in use in military radar systems (such as the one used on the very expensive Aegis ships the navy has) for decades. However, it depends on doing very complicated signal correlations, and until now that has been very expensive. However, if you do the work in software, it turns out to be pretty straightforward and the expense falls by 50% every year.

There are already companies playing this game — look at Vivato ( for example, which is making phased array WiFi equipment.

With phased arrays, there’s no reason why everyone can’t be using the same spectrum at once. Couple phased arrays, spread spectrum (a modern technology for reducing interference) and SDR and suddenly the way we use radio today looks totally silly.

Comments on this entry are closed.

Previous post:

Next post: