Audio quality degradation over FM transmission

[Tam/WB2TT]

Treble.
There is no real reason why you should lose bass, though it may decay
under
10Hz.

Irrelevant.
It's transmitted in a 75Khz bandwidth, about 32Khz above and below the
nominal frequency, and would have exactly the same information content
if it was centered on 88,108, or even 1Mhz.

Actually, it gets worse than that. In that 75 KHz channel, you have 15 KHz
of LEFT+RIGHT, another 15 KHz of LEFT-RIGHT (or is it RIGHT-LEFT), plus
MUSAC, or whatever. The LEFT-RIGHT channel is transmitted as a double
sideband suppressed carrier AM signal; so, it actually takes up 30 KHz. I
don't know this for a fact, but I suspect that if an FM station is
transmitting digital music, it is in the slot formerly used for MUSAC.

To transmit a 75 KHz bandwidth of signal, the FM transmitter actually uses
about 200 KHz of spectrum. That is just the way FM works. Defined by Bessel
Functions.

The 88 to 108 is immaterial, as has been pointed out. In the FM receiver,
the first thing that happens is that whatever station the radio is tuned to
gets converted to a narrow slot at 10.7 MHz. You can't make the slot wider,
or you would get interference from the adjacent channel.

Tam

 

[Pooh Bear]

Well! In "this recent hot weather", 108°F/42°C, my car radio has been
working just fine ;-)

Isn't it fairly flat ( topologically ) where you are ?

We ( with all our hills and valleys ) get all this atmospherically introduced
multi-path trouble ( at least I think that's the cause ). Never actually got
into really examining the fine detail of RF transmission and reception. I
reckoned I'd leave the finer points to someone else and concentrate on what I
seemed to be best at.

Graham

 

[Pooh Bear]

Actually, these things won't be very relevant. I am using this in my
car! The device will site less than 15 feet from the receiving antenna
in the same vehicle. Can we make some assumptions that way?


No, it's an analog signal propagating from my MP3 player's attached
transmitter to my car's antenna on the back -- which hooks into my FM
stereo Easier now?

Yes but....

My car's stereo has an auxilliary input for such things.

Might it not be simpler just to connect it in some such way ?

I've never come across an mp3 player with an FM transmitter. Is it legal ?
Does the Chinese company making it care ?

Anyways, I think someone answered that since it is encapsulated in a 75
Khz band, I am not going to lose much quality at all -- except for the
15 Khz band instead of 20 Khz. I don't think I can hear below 30 Hz or
above 18.5 KHz anyways...so no problem losing the highs!!!

In a car.... for sure !


Graham

 

[Pooh Bear]

I know that MP3 is lossy. All I am asking is if the audio input can
*physically* (due to transmission) match the audio heard over an FM
receiver exactly. Just assume that you have the highest quality CD
source media, CD player, FM transmitter, Antenna, FM receiver, and
Speakers -- and all that is left to degrade the signal is the
modulation of the input over the FM channel -- will the input match the
output??? This is what I want to know. I mean, the act of taking a
100% clean CD audio input -- piping it over FM (with no constraints but
the theoretical math involved in modulating the signal over say 108
MHz) -- will it end up 100% clean on the other side? I don't care
about any other factors. Mathematically, does the transmission of CD
audio input via FM have the effective output of about 256 Kbps MP3
lossy audio on the receiving end? Is it a degrading function (forget
the very low and very high frequencies)?

And if not, what is the theoretical maximum source input before you
start incurring loss on output? Obviously one could not send audio
sampled at 20 Ghz (insane) and get all that information over to the
receiver -- since 20 Ghz of information can not fit into that 108 Mhz
frequency band. I synnonomize this with mapping a much larger data set
onto a smaller data set (like an MP3 encoding). With two CD recordings
in a studio done at the same time using slightly different audo
equipment, it is mathematically possible for these two differing audio
samples to end up as an identical MP3 file in the end -- even though
there are subtleties in the original source audio, right? Would the
same be true of FM transmission of two similar sources? If not, then
you have a widening transformation of the original vector space.
Otherwise, it is a narrowing transformation resulting in loss of input
during mappings. Wow, I must have reworded this about 50 times
now...lol Forgive my naivety! I know I must be confusing some of you,
because I am confusing myself now...

Errr what ?

If you're sending an analogue signal it doesn't matter how it originated. Or
how good it is / isn't. The transmission medium ( unless digital ) doesn't
care how the signal was encoded. Totally irrelevant.

Graham

 

[Jim Thompson]

Isn't it fairly flat ( topologically ) where you are ?

We ( with all our hills and valleys ) get all this atmospherically introduced
multi-path trouble ( at least I think that's the cause ). Never actually got
into really examining the fine detail of RF transmission and reception. I
reckoned I'd leave the finer points to someone else and concentrate on what I
seemed to be best at.

Graham

I'm in somewhat mountainous terrain here. But about the only
multi-path stuff I ever see is in the downtown "canyons"... probably
all that steel.

...Jim Thompson

 

[John Perry]

Charlie Edmondson wrote:
...
And if not, what is the theoretical maximum source input before you
start incurring loss on output? Obviously one could not send audio
sampled at 20 Ghz (insane) and get all that information over to the
receiver -- since 20 Ghz of information can not fit into that 108 Mhz
frequency band. I synnonomize this with mapping a much larger data set
...

Kristian, you seem to have missed the repeated statement that there is
no 108Mhz bandwidth -- it's _always_ a 100KHz bandwidth centered around
88.1 MHz, 88.3 Mhz, 88.5 MHz, ..., 107.9 MHz.

The FM modulation technique restricts the audio bandwidth to 15KHz for
reasons others have stated. FM trades off extravagant bandwidth for
excellent resistance to interference and good audio fidelity.

MP3 will be the limitation, even if its parameters are adjusted for
maximum fidelity (if I understand the MP3 paradigm correctly, which I
may well not). But a standard CD will have better fidelity than the FM
channel because the CD is limited to 20KHz, and the FM channel is
limited to 15KHz.

John Perry

 

[Roger Lascelles]

Actually, these things won't be very relevant. I am using this in my
car! The device will site less than 15 feet from the receiving antenna
in the same vehicle. Can we make some assumptions that way?


No, it's an analog signal propagating from my MP3 player's attached
transmitter to my car's antenna on the back -- which hooks into my FM
stereo Easier now?

Anyways, I think someone answered that since it is encapsulated in a 75
Khz band, I am not going to lose much quality at all -- except for the
15 Khz band instead of 20 Khz. I don't think I can hear below 30 Hz or
above 18.5 KHz anyways...so no problem losing the highs!!!

Kristian Hermansen

Firstly, FM involves some distortion even at the theoretical level, because
you receiver has to discard some of the ("infinite") FM sidebands. I don't
know how much distortion. The distortion will be worse at high deviation.
Your car radio will have narrow band filters, which increase distortion.

Secondly, most FM radios are not kind to audio. Phase shifts in the IF, and
non-linearity of tuned circuit discriminators add distortion. Especially at
full deviation, distortion is not negligable. With a clean music source, I
would expect to easily hear the degradation caused by a car radio. FM
receivers do have decent treble, but it I think it is often audibly
distorted. High end MP3 should easily beat FM radio. Just what bit rate
you need for "FM" levels of degradation, I don't know.

Thirdly, your transmitter will have distortion. 2% THD or more would be not
surprise me.


Your whole transmitter - receiver chain at full deviation could do anything
from 7% THD down to about 1% THD.

Roger

 

[Ian Stirling]

Nope, the MP3 player doesn't have that quality. The reason there are
MP3s is because of compression, they eliminate some of the 'unneeded'
data that is stored on the CD so that it will take up less space. This
compression means that there is more noise and less dynamic range in an
MP3 than the original. You also loose some of the highs.

Which is usually unnoticable, except for the lowest bitrates, and the
best listening conditions.

Next, you have your MP3 player. When he said designed down to a price,
he was talking about the MP3 player, not the transmitter. They often do
not have the 'best' audio stages, being just good enough to not sound

I've measured the outputs of my cheap MP3 player with single tone outputs.
It vastly outperforms my ears, and shows good results on test equipment,
coming up with a dynamic range of about the same as my CD player.

Many/most low-power FM transmitters are of very bad quality, and come
nowhere near what FM can achieve.

 

[gwhite]

Firstly, FM involves some distortion even at the theoretical level, because
you receiver has to discard some of the ("infinite") FM sidebands. I don't
know how much distortion. The distortion will be worse at high deviation.
Your car radio will have narrow band filters, which increase distortion.

I doubt many, if any, car radios have bandwidth switches anymore.

Secondly, most FM radios are not kind to audio. Phase shifts in the IF, and
non-linearity of tuned circuit discriminators add distortion. Especially at
full deviation, distortion is not negligable.

But that is not such a big concern, as they are only brief transients.
Besides, the car environment is loaded with other noise.

With a clean music source, I
would expect to easily hear the degradation caused by a car radio. FM
receivers do have decent treble, but it I think it is often audibly
distorted. High end MP3 should easily beat FM radio. Just what bit rate
you need for "FM" levels of degradation, I don't know.

For S/N relative to some distortion level ("big signal ref"), you can
probably expect > 65 dB even in car radios.

Thirdly, your transmitter will have distortion. 2% THD or more would be not
surprise me.

The cheezy low powered TX'ers are the weaker part of the link, from my
experience.

Your whole transmitter - receiver chain at full deviation could do anything
from 7% THD down to about 1% THD.

Example of what a decent FM RX'er can do:

http://www.fanfare.com/ft1a-sht.html

I don't know about car RX'ers -- I just thought the linked one might be
an easy reference.