In the early days of Amateur Radio signal reports were always given accurately, and most often without any exact metering system. Reports were given by ear.
In the last days of Amateur Radio signal reports are often not given at all, no matter the signal strength, 599 or 5NN is given.
The majority of those who give accurate signal reports use the RST system and appear to use it correctly as it has always been intended, assigning the correct number to the situation, more on that below.
A minority of those who intend to give accurate signal reports use the RST system but use it incorrectly: they read the value off their S-meter. More on this below, and why this approach is not correct.
First, let us acknowledge that the actual basis of the Amateur Radio License is experimentation and self-education in communications, and this is the basis upon which the ITU world wide defines amateur radio.
For many of us, we still enjoy building our own equipment, even if only our own antennas, and also the study of propagation. To this end, we wish to receive accurate signal reports. Inaccurate signal reports are not helpful at all.
So what is an accurate signal report? And what constitutes an inaccurate report?
It is difficult to give an entirely accurate report. However, there are way to use and not use the S-meter in the giving of "accurate" reports. First, let us look at the "R" part of the RST system:
R1 = UNREADABLE (Comment: that means TOTALLY unreadable, not even occasional words)
R2 = BARELY READABLE, Occasional words distinguishable
R3 = READABLE WITH CONSIDERABLE DIFFICULTY
R4 = READABLE WITH PRACTICALLY NO DIFFICULTY
R5 = PERFECTLY READABLE
The "R" part of the signal report cannot be measured by a meter on a radio, that is clear. The R is an easy part of the report to give accurately. It WILL be influenced in some circumstances by your ability: your hearing, your interpretation skills, your code skills. It will certainly be influenced by factors such as signal to noise ratio, fading, interference. But there are clear differences between each of the 5 steps:
If you have total arm chair copy it is R5. If not, R4 if no difficulty but ANYTHING LESS than PERFECT COPY. If you have any real difficulty copying, but are still able to copy (even if with repeats), then R3 is appropriate. If you only get little bits here or there, it is R2. If you cannot make anything out it is R1. So the R part of the system is very simple.
So what about the S (Signal Strength)? This is the most difficult part of the system and the one that is, these days, used in different ways, which renders it unhelpful. As stated earlier, some always give it as S9. That is WRONG. Others read it directly off the S-meter: that too is wrong, and here is why:
The S-meter is dependent upon many things: receiver model, receiver settings (bandwidth, notch, attenuator, pre-amp, etc). It thus serves ONLY as a guideline (more on that below) but can be useful in direct antenna-switching comparisons. The wrongful use of the S meter leads a minority of stations to directly read e.g. S1 on the meter, for a signal that is actually around S5 to S7, due to having pre-amp off, and perhaps even attenuator on, and/or a sluggish S-meter.
Just as it is more accurate to decode Morse Code even in the age of supercomputers BY EAR (with an experienced operator) so it is also more accurate to give an S report (not S METER report) by EAR. One has to compensate and consider a number of factors, when doing so, for example:
On higher frequencies, such as 28MHz, the noise level may be virtually nothing, allowing for very weak signals to be heard. On a typical receiver, the S meter at 28MHz is very sluggish, it may not even move until a signal is strong. Many QSO the S meter will be showing S0 but the signal is R5 and loud. If we look at the OFFICIAL S-scale, we can see there is NO S0 in existence, since S0 would actually mean NO SIGNAL AUDIBLE AT ALL:
S1 = Faint signal, barely perceptible
S2 = Very weak
S3 = Weak
S4 = Fair
S5 = Fairly good
S6 = Good
S7 = Moderately strong
S8 = Strong
S9 = Very strong signals
From the official S scale above, we can see very clearly that this is NOT AT ALL intended for direct measurement from any S meter! Otherwise, there would be S0, and there would be options for S9+10db etc, and the scale would not be given in broad human interpretation terms, but instead would be given as a measurement of e.g. Micro Volts etc.
So, returning to the example of a very quiet 28MHz band with a loud signal not moving the S-meter on the radio. Clearly it is not RST509, so what is the actual S value to be given? This is where you can use your experience of your radio, and its response, and also the feel of how strong the signal is: if the signal is NOT barely perceptible, it's obviously not S1. Is it VERY WEAK (S2), or just WEAK (S3)? Is it FAIR (S4) or even FAIRLY GOOD (S5)? Or even GOOD (S6)? You decide, and give the report accordingly.
Another example, on 80m, the S meter may be over responsive on some rigs. Noise levels will have a big factor often on lower frequencies. If a signal is buried in the noise, one cannot give an accurate S report, however, the R report will evidently be 3 or less, and that will give a clue as to the S report being a "wild guess", e.g. "339 LOCAL QRM5" will be clear to the receiving station that the "S3" is only a guess. With strong local noise of "S8" the buried signal may be in fact S6.
Fading (QSB) makes giving signal reports on relatively low powered amateur radio signals difficult, as signals will typically fluctuate by even several S scale levels: a signal could be fading between Strong (S8) and Weak (S3) but usually a fade of one or 2 S points is normal on HF even in good conditions. Thus the amateur S scale is not so practical for general use unless a range is given as e.g. RST539/579 or RST53/79 etc.
This is why the commercial QSA code has only a scale of 1 to 5 making it much easier to give a report unaffected by normal levels of fading: E.g. QSA3, would cover from around S4-6, QSA 4 would cover from around S7-8, and QSA5 covers S9 and "above"… Note that on the S scale, there is NO such thing as "S9+10DB" etc. ONLY S9. S9 means VERY STRONG SIGNALS — no matter that is S9 or S9+60dB. The S meter can of course be used, but it is not the official, standard and intended use of the S scale as in the RST system.
For this reasons, some, including this author, have suggested alternative systems to RST, which can be adopted among regular QSO partners or some nets, to make things shorter, quicker, and simpler. E.g. the T in RST is almost invariably T9. If it isn't, QRI is available: QRI1 means good tone, QRI2, variable tone, QRI3 bad tone. Since 99.9% of T reports are T9 these days, we can safely drop it. The "Q" report suggested at http://www.vkcw.net/q is one such system.
Although radio amateurs, especially if taking the modern understanding of the word "amateur" as being "non professional" are under no obligation to follow any standard, and can do as they please. Thus we have those who can hardly copy you, giving a 599 report, and those who copy you very loud and clear giving a 519 report. But it certainly helps if we are on the same page, and if standards are adhered to. Otherwise, one should create alternative systems.
It has been suggested to create an alternative to "599" for those who don't wish to give a signal report. Meanwhile, if using RST, please refer to the above tables for the PROPER STANDARD. As stated, amateurs can do as they please, and it is a hobby and for most of us it is about having fun. From time to time it is necessary for us to remind ourselves of what the actual standard is, otherwise it gets lost and the majority abandon it.