Most if not all modern radios have an ALC circuit, which at the level of the noise floor will modify the readings of said s-meter, and anything based on said signal that the s-meter reads. At said noise floor, signals that are farther apart power wise will appear closer together on an s-meter (weather it is measured in s units or dBm) due to this circuit. Very powerful signals will have a similar effect. This is the same with ham radio and CB. The 3 dB difference being significant number I gave comes from before these circuits were widespread and commonplace. Said circuits allows the radio to process a much wider range of received signal strengths than before. Because of this, i will acknowledge that modern day equipment that it can take a larger change in signal to be noticed.
When it comes to the plots you made you categorical estimation based on gain and assumed it would be the same for all contacts. The truth of the matter is that for local communications said plot is wholly inadequate as a measure. This is a surface wave chart for one of the antennas above (the 5/8 with mast, both with and without radials to control CMC, mounted one wavelength high). The measurement is the strength of the RF wave front taken at a given distance from the antenna running at 100 watts of power, and from 0 to 11 meters high. This chart does not take into account any losses or gains that a receiving antenna at this location would also present, but is a direct measure of the signal strength available at this distance over a moderate earth.
249 nV/m is 0.249 μV/m, or just above s1 on what an s-meter would show if it received a signal of that strength. 430 nV/m is 0.43 μV/m, and just above s2 on what an s-meter would show if it received a signal of that strength. Because of this, an antenna at this distance and height would see almost 1 s-unit difference between said antennas with and without radials... That comes very close to your minimum number to notice a difference, and given not every possible real world variable is in play here, could easily end up exceeding your number. There have also been real world reports of people who made a similar change that actually exceed these results. That isn't to say everyone who adds radials to said antenna will notice the difference, every installation and location is different after all. All I'm saying is your categorical statement that no one can and will ever notice a difference from this change in any circumstances as going a bit far.
The higher the transmitting antenna is above the earth (in wavelengths) and/or the further the receiving station is away from the transmitting station, the bigger the received difference will be between antennas. Because of this, one far field gain calculation and angle comparrison, wile a useful calculation for one aspect of radio, namely skywave propagation, is irrelevant when discussing long distance local surface wave contacts, i.e. contacts from people at the edge of your local communications range.
When it comes to modulation types, I agree that CW would require the least amount of signal difference to be noticed, in order I would followed that with SSB, then AM, and finally FM, which requires the most bandwidth of all of these modulation types listed, and based on differences with how FM works, there would be no "audio fading" at all at low powers at all as it uses frequency variations instead of the amplitude variations that AM and SSB use to send its information. The 3 dB number I gave above comes from noticing a difference with older SSB radios.
The DB
