|
Option Limiter |
Enabling the limiter is essentially a choice between signal-to-noise ratio (SNR) and dynamic range handling.
Enabling it is particularly useful in the following situations:
With severe fading (QSB):
When the signal constantly fluctuates between “extremely strong” and “barely audible” (typical on shortwave), the automatic gain control (AGC) can sometimes react too slowly.
The hard limiter radically cuts off all amplitude fluctuations above a minimum and provides the filters with a signal at a constant level. This helps the decoder to more accurately find the signal’s zero crossings.With very strong interference (QRM) in the vicinity:
If a very strong signal appears right next to your RTTY signal, it could “push down” the AGC, making your actual signal too quiet.
A limiter ensures that the frequency information (the zero crossings) takes precedence over the amplitude.When CPU load needs to be drastically reduced (theoretically):
A limiter simplifies the signal so much that subsequent calculations need to be less precise. In our implementation, however, this plays a minor role, since the dual-tone logic is running anyway.
When should it remain OFF (limiterless)?
In most modern scenarios, limiterless (i.e., NO) is better:
With a lot of noise:
A limiter raises the noise level during pauses to the same level as the signal. This leads to more “error start bits.” Without a limiter, the noise remains quiet, and the matched filter integrator can better clean up the signal.With selective fading:
Since we now have the dual-tone logic, which levels the mark and space independently, we are already compensating for fading on a more intelligent level. A hard limiter in front of it would destroy the subtle level differences that our new noise reduction logic uses.
Recommendation
Leave the limiter off by default. Only turn it on if you have a very strong but fluctuating signal (fast flat-fading) where the decoder “loses its composure” despite a good signal level.