Storing radio signals - A radio Tivo?

While sitting in a friend’s car this evening (in motion; we weren’t just sitting there), he was flipping around on the radio trying to find a good song. He finally found one we liked, but it was partway through. He hit a button and the song started anew.:eek:

It turned out that it was actually a cd that he had switched to, and he had just started the song over.

However, that got me thinking. As more and more car stereos are going to be encorporating hard disks or other massive storage for mp3 and other media, how hard would it be to have a recorder built in to the radio that would record what was playing so that, if you changed to a song you liked, you could just rewind and hit the rest?

With, say, a 2 G drive, you should easily be able to store the last 20 minutes (plenty long for even quite a long song or set) of the maybe 100 radio bands you can receive at any given time (FM, every .2 MHz, AM, uh… every 100 KHz?) in 128Kb mp3. And 2 G is pretty small. You could have a bigger drive with other stuff on it too.

The big problem, I think, is the tuner. I have only a dim understanding of the workings of a radio, but I think that you have to actually be tuned in to only one station at a time. To receive and store 100 stations, you’d need 100 tuners. The first thing I thought of was to just have, say, 6, that were set to your favorite stations. But then I wondered?

Is it possible to just capture the entire radio signal at once, digitize it, and store it on the hard disk? Is it reasonable to do so? If you could store the whole signal, then no problem; just go back and tune to what you wanted to hear.

My guess is that it would take significantly more space to store the radio signal. To capture radio at 110 MHz, we need 220 million samples/sec, right?

So, calling electrical engineers: Is this feasible? if not, what’s the limiting factor?

That’s what you need if you want to digitize the raw waveform. This is possible in theory, but extremely difficult in practice. If you digitize at a 4-bit resolution you end up with a 1-gigabit/sec data stream. A gigabit data recorder costs a couple of million dollars, IIRC.

But it’s not quite that bad. A 110MHz signal doesn’t contain valid signals at all frequencies up to 110MHz. All the data is contained in a narrow bandwidth around 110MHz. So if you want to record all the radio stations between 100 and 110 MHz, you can throw out all the components below 100MHz. Essentially you end up with 10 MHz worth of actual data.

But 10MHz at 4-bit resolution is still 40 megabit/sec, which is higher than most commercial hard drives can handle. You need a hard drive array, and a high-speed computer system that can handle the throughput. I think it’ll be a $10,000 system.

The total bandwidth of FM radio is 20 MHz, the same as three and a third TV channels. This suggests you’d need roughly the equivalent of over three TiVos. I think your first idea would work better. Capturing, say, ten stations would give you 2 Mhz, or only a third of a TiVo.

I don’t think you’d necessarily need ten different tuners, though. I think you’d receive the whole band, but extract out the bands you want, and send them off for more processing (probably in parallel), before storing them to the harddrive. All of that would be done digitally. I do NOT design radio receivers, so take this as a WAG.

You’re missing a decimal point there. The FM broadcast band is 20 MHz wide in its entirety. Stations are allocated 200KHz channels. Each channel breaks out into a standard carrier deviation of ±75KHz with a 25KHz guardband on each side.

Hmmm maybe you’re not missing a decimal point. I mis-read that as saying 'The total width of an FM radio station is 20 MHz…

I’m willing to bet that the NSA and other such agencies have devices that do exactly what you what you describe. Perhaps you can call them and ask who their vendor is? :wink:

Huh? I must be missing something big here…

How is it an FM broadcast takes more bandwidth than a video signal which also includes an audio track and that audio track itself is often in stereo?

Also, A friend of mine has an Apple iPod that can store 10GB of data and the whole thing is no bigger than a pack of cigarettes. The iPod merrily plays back MP3’s no sweat. Why can’t an audio Tivo do something similar (encode the audio as a compressed MP3 format…or pick your compression)?

I see no reason why this couldn’t be done. Maybe you lose some of the fidelty in playback due to compression but it’d be good enough. Hell…I wouldn’t even mind a small memory version of Tivo that merely ‘remembers’ that last (say) five minutes of audio. There have been plenty of times I missed a favorite song or news report and wished I could back it up a minute or two (in fact because I have Tivo at home I find I automatically begin to reach for a non-existant button to do that very thing).

Whack a Mole is right. You really want to demodulate the FM signal before you start saving the signal. Tivo certainly demodulates the TV signal before compressing it.

I think the Poster was saying that FM radio, all of it, is work about 3 and hald TV stations. An individual FM radio station is only a portion of that.

The entire FM band is 20 MHz, not a single station. That’s about 100 possible FM stations.

A video signal is amplitude modulated and an FM radio signal is frequency modulated (that’s what FM stands for). In an amplitude modulated signal, you have a carrier on a single frequency and you store information by varying the signal level, kind of like turning the volume control up and down. For a frequency modulated signal, the information is stored by shifting the frequency of the carrier back and forth. The dynamic range of the system (in other words the difference between the loudest sound it can reproduce and the quietest sound it can reproduce) depends on how small of a frequency shift the detector can detect, and how much you swing the frequencies back and forth (aka the bandwidth requried). The greater the dynamic range desired, the greater the bandwidth is required for an FM system.

The advantage of an FM signal is that the receiver locks on the signal and tracks it as the frequency shifts back and forth. Any excess clutter and noise in the system gets filtered out because the receiver isn’t directly looking at the signal, it’s only looking at the change in frequency of the signal.

AM radio is also transmitted by amplitude modulation. You’ll note that AM radio signals require significantly less bandwidth than FM signals, and are also significantly more noisy. Broadcast AM radio is also limted to a single audio channel (no stereo) and is frequency limited as well. The dynamic range of the system is limited by the dynamic range of the receive amplifiers, and therefore is typically a lot less than the dynamic range of an FM system.

In other words, if you transmitted audio signals by AM instead of FM, they would require less bandwidth but they would also sound like crap, even if you did allow it to be in stereo and increased the audio bandwidth to match that allowed by FM.

TV channels are 6MHz wide in the US.


Hey, I’m all for that, but I don’t think you can do it for more than one signal at a time without having some really hefty parallel hardware. Tivo only has to do it with one signal at a time, but, ideally, this system would be able to capture all 100 odd stations in the FM band. Zenbeam suggested that you wouldn’t need multiple tuners, but I don’t know any more about radio design than he. Could you extract and compress the audio from multiple channels at once?

There is a lot of work now being done on software radios. So in theory you could take that 20MHz chunk that is the FM band and with a fast DSP demodulate more than one FM station with only one set of RF hardware.

Actually now that I think about it I don’t think that you could use a software radio to demodulate to FM where one had a strong signal and the other had a much weaker signal. So I think you are out of luck iamthewalrus(:3=.

This is what I was thinking of, but I don’t much about it.