[BC] AM Fidelity vs FM fidelity

[Robert Meuser]

The link only has some tired old basic stuff comparing AM produced by a 
4 quadrant multiplier and a practical example of one means of high level 
modulation. We wasted too much NAB session time in the early 80s 
re-hashing that one. How a signal is generated is all but meaningless. 
These days all that is required is putting the appropriate equation for 
the desired modulation into a DSP and create it mathematically.

I will stand by my previous comments in that many things are 
theoretically possible in the lab but when made to work as a practical 
system there are real factors to consider. The quoted statement, which 
is really more like the distorted political edited sound bite attack ads 
now in season, actually proves the point because it is presuming a 
bandwidth limitation for a FM *system* while no constraint is applied to 
the part of the statement pertaining to AM.

Put two transmitters in a LAB and a theoretically perfect AM transmitter 
operating at 1 Mhz could be modulated to to just under 500 KHz before 
there would be an alias problem while one on 540 Khz could only be 
modulated to just under 270 Khz. That is in a perfect world. Assuming 
there was actually modulation components that high in frequency an anti 
alias filter would be necessary which would further limit to a degree 
the maximum response. Now, try this with a real transmitter and the 
limitation in a nearly perfect world would the the transmitters' output 
network. I presume one could use AM techniques at VHF to raise the limit 
but that is too far off topic as far as broadcasting is concerned.

Take a theoretically perfect FM transmitter and, yes, in the lab there 
will be many more sidebands for a given modulation frequency; a broader 
more enlightened view of modulation theory will reveal that when 
analyzed as a system other factors come into play. Keeping the two 
transmitters in the LAB, lets add noise and set a transmission systems 
criteria based on signal to noise. If the AM transmitter requires X 
power to meet a target signal to noise ratio with 5 kilohertz of audio, 
it will require 3 db more power to maintain that level for 10 kilohertz 
audio and yet another 3 db to reach 20 kilohertz. The AM system would, 
therefore require 6 db more power to support the added frequency 
response.  The FM signal if left unconstrained, as was the AM signal, 
will deliver a constant signal to noise regardless of the modulating 
frequency if the modulation index is kept constant. FM is essentially 
spread spectrum as it trades bandwidth for immunity to noise and 
interference.

Take either system out of the lab and place them in a real world setting 
and many things change. Either modulation technique when made to work as 
a system must conform to certain channel limitations, external sources 
of noise and interference and constraints of real receivers. In the real 
world an FM system can easily achieve 20 KHz response while it is highly 
improbable to accomplish the same in a real end to end system.
  In that context, Mr Khan's statement is an outrageously edited sound 
byte which is a partial truth not applicable to real broadcasting. It is 
strange to the degree that he often compares himself to Armstrong.

On 6/21/12 11:16 AM, RichardBJohnson at comcast.net wrote:
> No it is not!
> My statement, as was Kahn's, was about Amplitude Modulation. It was not about a particular implementation and was, in fact, the reason why I spent so much time, effort, and money perfecting a specific implementation. It is repeated here:
>
> "In fact, amplitude modulation has the capability of better fidelity than FM because all of its bandwidth is available for audio."
>
> Learn something about Amplitude Modulation before you contradict experts:
> http://www.abominablefirebug.com/Modulation.html
>