The audio and EMR spectra may be different underlying phenomena, but for the purposes of radio, which transmits over EMR but encodes audio frequencies, the correspondence is significant. It's not possible to encode a higher-frequency signal over a lower frequency carrier, so radio frequencies which are used for the transmission of human-detectable sounds must be at a higher frequency than those sounds.
I'm not nearly enough of a radio / audio / encoding geek to tell you what the minimum overage is, though I suspect it's at least 2x greater than the highest-frequency sounds that are intended to be encoded. For human-range hearing, that would be in the 10-20 kHz range, which would mean that a minimum carrier frequency would be somewhere around 20--40 kHz.
The longest wavelengths used for voice comms so far as I know are the longwave or LW band, which is from 30–300 kHz, with the lower end of that pretty much precisely where I'd suggested it might be.
Most commercial AM broadcasts are in the medium-wave (MW) band, from 531 kHz – 1602 kHz, or roughly 25-50x higher frequencies than the highest tones the human ear can detect. And I'd suspect that fidelity above 10kHz audio is not especially high even at those transmission frequencies.
"radio, which transmits over EMR but encodes audio frequencies"
Radio conveys all sorts of things, not just audio. They're orthogonal issues.
And you actually could modulate 20khz BW of audio onto a 30khz carrier using SSB-AM for example. That carrier could be EM, a voltage in a wire or an ultrasonic tone. Not that it's relevant to the issue but it does reinforce the point that the nature of the medium is independent from the intelligence being conveyed. There is no correspondence. Sonics/audio are pressure waves and do not belong on the EM spectrum.
> It's not possible to encode a higher-frequency signal over a lower frequency carrier..
Actually it is possible. AM modulation is just a matter of multiplying two frequencies together. It's just that the lower sidebands wrap around DC. It's a bit harder to demodulate, but not impossible.
>.. AM broadcasts are in the medium-wave (MW) band...And I'd suspect that fidelity above 10kHz audio is not especially high even at those transmission frequencies.
In the transmitter the fidelity is limited by the bandwidth of the modulator, and in the Receiver by the bandwidth of the IF stage. You can actually get FM like audio from a BC AM radio, but not with cheap and nasty receivers designed for the American market. In Europe many HiFI AM receivers have an optional "wide" IF filter, and the transmitters can transmit 15KHz audio.
"In the transmitter the fidelity is limited by the bandwidth of the modulator, and in the Receiver by the bandwidth of the IF stage."
This applies to heterodyne systems with the bandwidth limitation coming from deliberate filtering.
My understanding is the channel width, set out in the broadcast standard and license, is the bottleneck and not so much the limitations of the electronics. FM must be under 150khz (5 times the maximum source bandwidth of 15khz, then doubled for upper and lower sidebands) and MW-AM must be under 9/10khz (limiting the source to around 5 kHz). So without the transmitter broadcasting outside its allocation, I don't understand how changing the receiver could improve the audio quality.
I could be wrong. Do you have a source on 15khz audio bandwidth over MW-AM?
There is no direct connection between the channel spacing, and the allowed TX bandwidth.
If the bandwidth is wider than the channel spacing (for AM BC), all that happens is that adjacent stations hear a little "monkey chatter". And that is minimised because most stations are arranged on a geographic grid which prevents a local station having another local on an adjacent channel.
In Australia the local Government ABC stations most definitely do transmit 15KHz wide audio.
A quick search turned up this article from September 1980 ETI magazine which discusses stations having 30KHz total bandwidth. https://imgur.com/a/gVpSwoo
Whatever, just tuning around with a high-end SDR shows many stations at 20KHz bandwidth.
The Commercial stations have a problem in that they use aggressive audio processing in order to maximise their coverage, given their licensed power output. So as soon as the commercials decided to concentrate on talk shows, they wound down their bandwidth.
AM modulation is just a matter of multiplying two frequencies together.
Pardon my limited knowledge here, but isn't amplitude modulation actually the practice of varying the amplitude of the carrier frequency alone?
FM, frequency modulation, by contrast does vary the frequency of the carrier itself, which can be modeled mathematically as a function of two (or more) frequencies being combined as with Fourier analysis.
I think there is some confusion here mixing up the time and frequency domains. The poster above is correct that AM is achieved by multiplying the intelligence with a carrier. This changes a carrier's amplitude which must (as indicated as a Fourier series) generate sidebands, which occupy frequencies around the carrier frequency. FM is produced differently but also generates (wider) sidebands.
I'm not nearly enough of a radio / audio / encoding geek to tell you what the minimum overage is, though I suspect it's at least 2x greater than the highest-frequency sounds that are intended to be encoded. For human-range hearing, that would be in the 10-20 kHz range, which would mean that a minimum carrier frequency would be somewhere around 20--40 kHz.
The longest wavelengths used for voice comms so far as I know are the longwave or LW band, which is from 30–300 kHz, with the lower end of that pretty much precisely where I'd suggested it might be.
Most commercial AM broadcasts are in the medium-wave (MW) band, from 531 kHz – 1602 kHz, or roughly 25-50x higher frequencies than the highest tones the human ear can detect. And I'd suspect that fidelity above 10kHz audio is not especially high even at those transmission frequencies.
<https://en.wikipedia.org/wiki/AM_broadcasting#Broadcast_band...>