The hardware side of an SDR is a pair of incredibly fast analog to digital converters, capable of tens of millions of samples per second. If you can sample at 40 million pairs of samples per second, you've sampled every radio signal in 20 megahertz of bandwidth, and moved it into a form where you can computer.
There are then techniques that are used to recover individual signals. I'm not really an expert in this, so forgive/correct me if I get this wrong, but my understanding is that multiplying two signals at frequencies A and B is the same thing as adding together a signal with frequency (A+B) and a signal with frequency (A-B).
So for AM radio, if A is the carrier and B is the message, we can multiply (A+B) from the original signal with a copy of A' we generate in the computer. That yields (A+B)-A', which is just B - the signal we want to listen to. (Plus high frequency signals that can be filtered out.)
The thing is, this last step can be done multiple times - as many times as the computer can compute, since it's all done by the process. So that's how it can listen to multiple signals at once.
"The hardware side of an SDR is a pair of incredibly fast analog to digital converters, capable of tens of millions of samples per second. If you can sample at 40 million pairs of samples per second, you've sampled every radio signal in 20 megahertz of bandwidth"
Yes, that makes a lot of sense!
Never thought about that until now; you (and the other posters who have posted virtually the same thought/thing) have expanded my mind as to what I previously thought possible (but then, I'm a software guy who dabbles in electronics -- not the other way around)...
But anyway, thank you! (and the other posters too!)
> There are then techniques that are used to recover individual signals. I'm not really an expert in this, so forgive/correct me if I get this wrong, but my understanding is that multiplying two signals at frequencies A and B is the same thing as adding together a signal with frequency (A+B) and a signal with frequency (A-B).
Yah, this is the gist of how a heterodyne receiver works, along with filters before the mixing to reject the unwanted frequencies.
In practice, in software, you do quadrature. That is, (rf * sin x + rf * cos x) and (rf * sin x - rf * cos x) will each get you just the sum and difference frequencies; this way you avoid any prefiltering.
The hardware side of an SDR is a pair of incredibly fast analog to digital converters, capable of tens of millions of samples per second. If you can sample at 40 million pairs of samples per second, you've sampled every radio signal in 20 megahertz of bandwidth, and moved it into a form where you can computer.
There are then techniques that are used to recover individual signals. I'm not really an expert in this, so forgive/correct me if I get this wrong, but my understanding is that multiplying two signals at frequencies A and B is the same thing as adding together a signal with frequency (A+B) and a signal with frequency (A-B).
So for AM radio, if A is the carrier and B is the message, we can multiply (A+B) from the original signal with a copy of A' we generate in the computer. That yields (A+B)-A', which is just B - the signal we want to listen to. (Plus high frequency signals that can be filtered out.)
The thing is, this last step can be done multiple times - as many times as the computer can compute, since it's all done by the process. So that's how it can listen to multiple signals at once.