If you use up too much of the error correcting capability to correct errors produced by non-linear operation of your equipment, there won't be enough capability left to correct errors produced by effects illustrated below.
The next three plots are used to show some propagation effects that may cause errors which need to be corrected. In all three plots, the horizontal scale is time and the vertical scale is in phase change units. The tic marks along the horizontal scale are one symbol period (about 8.16 milliseconds) apart.
This first, of three, plots shows the "ideal case" for the first four subcarriers, which are spaced at 230 Hz intervals. This case is "ideal" in the sense that it represents what went into the transmitter. If no corruption was introduced by the channel, then the same signals should come out of the receiver.
Note that there are four waveforms that, over the center portion of the plot, are essentially time shifted versions of each other. Also, note that the order is: red, green, blue, and magenta.
This plot shows the corresponding signals from a transmission by VK3LM, as received by W8ZCF.
Note that even though the energy represented by the green trace (centered around 805 Hz) left VK3LM AFTER the energy represented by the red trace (centered around 575 Hz), it was received EARLIER by W8ZCF.
Comparison of the red, blue, and magenta traces of this plot and the previous plot shows that they have not maintained the correct time relationship to each other either.
This plot shows the corresponding signals from the same transmission by VK3LM, as received by W8ZCF about 5 seconds later that those shown in the previous plot.
Note that even though the energy represented by the red trace (centered around 575 Hz) left VK3LM BEFORE the energy represented by the other three traces, it was received LAST by W8ZCF.
In addition to it's relative time shift, the energy centered around 1265 Hz (shown by the magenta trace) suffered additional corruption, which has caused a significant error in the received waveform
Note the absence of any indication of non-linear operation
Note the extra lines (other than the ones at 1180 Hz and 1520 Hz), which indicate non-linear operation at this highest level of the 2 Tone Test Signal. This highest level is 3db above that used for the HDSSTV transmissions. Thus, some evidence of non-linear operation at this highest level of the 2 Tone Test Signal indicates that you are not significantly below the maximum level you could use, and still remain in the linear region for the HDSSTV signal.
Even though this spectrum shows evidence of non-linear operation, it was not severe enough to prevent successfully correcting all of the errors in this reception.
Although some people prefer to operate their equipment with the RF gain control at its maximum setting, this is generally counterproductive here. Unless the signal is so weak that you have to use all of your RF gain to get any signal at all, you can most likely improve the signal to noise ratio by reducing the RF gain, from its maximum value.
A louder signal, if it is distorted, should be expected to use up more of the error correction capability that one that is not as loud and not distorted.
Louder and distorted is not as good as quieter and not distorted.
Links to: