What is so different about using software for signal processing?

Foundations of Amateur Radio

In my ongoing software explorations I've discussed that Software Defined Radio or SDR is a fundamentally different way of dealing with radio. It's been in use across non-amateur circles for decades. Your mobile phone has an SDR on board for example.

The original term of "digital receiver" was coined in 1970, "software radio" was coined in 1984 and in 1991 Joe Mitola reinvented the term "software radio" for a planned mobile phone base station.

So, this idea has been around for half a century and in amateur radio this idea is also catching on. You can buy a few pure SDR devices today, some hybrid ones, or you can begin to experiment in a more indirect manner using your traditional radio and a computer.

One of the things that sets this idea of a software defined radio apart from anything we've done so far is that the bulk of the signal processing is done in software. That sounds obvious, but it's really not.

One of the impacts of this idea is that you can improve your radio communications by either writing better software, or by using a faster computer. Unless you write software for a living, these things aren't immediately obvious, so let me explain.

Imagine that you've written software that detects beeps in a particular slice of audio spectrum that's being fed to your application. As you write better software to detect those beeps, you get a better digital mode, one with a better chance of being decoded, or using radio terms, it has a better signal to noise ratio.

If that's not a familiar term, signal to noise ratio is the a measure that describes the difference between a wanted signal and the background noise. Higher signal to noise means that you can better distinguish between the two.

If you stand in a room full of people talking and you use your hands to cup your ears towards the person you want to hear, you've increased the signal to noise ratio and your chance of understanding them has improved.

As you write this software, it gains complexity. As you deal with more maths, more samples, more tests, you end up running out of time to make your decoder return a relevant answer. There's no point in having a real-time signal being decoded late. If it were to take say 10 seconds to decode 1 second of audio, then the next second would be 20 seconds late and the one after that would be 30 seconds late.

That's where a faster computer comes in.

If you have the ability to do more maths, or do the same maths at a higher resolution, you will essentially improve the reception of your radio without ever needing to change your antenna or anything on the circuit board.

Think of it in another way.

Imagine that your tool has access to 2.3 kHz of audio. It's the equivalent of a Single Side Band audio stream. If you break that down into 23 chunks of 100 Hz each, you can deal with the average of 100 Hz of audio for each calculation. If you have a faster computer, you might be able to break that down into 230 chunks of 10 Hz each, or 2300 chunks of 1 Hz. So instead of doing calculations across 23 chunks of audio, you're doing it across 2300 chunks.

Why is this significant you might ask?

Well, in a traditional radio you get one bite at the cookie. You get to design and build your circuit and then package it and sell it. The end result is something like my FT-857d. It does what it does well, but it will never get any better.

However, if I plug that same radio into my computer, I can extract the audio and do stuff with it. If I get a faster computer, I can do more stuff. I don't have to change my radio, or my antenna, or even my shack. Most of the time I run a different application and I get a different result.

I will point out that I'm deliberately ignoring where and how the RF gets to the computer, or where that computer actually is, or what operating system it's running, since none of those things matter to get an understanding of how changing software can change the performance of your radio.

I've said this before and I'll say it again: "The SDR earthquake will change our hobby forever"

Before I go. I'm not for a minute suggesting that your current radio is obsolete. If it were legal, a spark-gap transmitter could still exchange information today, but if you want to explore what might be just over the horizon, going down the SDR path by connecting your radio to your computer is a really nice place to start.

I'm Onno VK6FLAB