In various weak signal modes of WSJT-X, the Tone Spacing decreases with increasing T/R Periods. For instance for FST4 / FST4W (as decribed on page 1 of the Quick-Start Guide) we see:
|T/R period (s)||Tone Spacing (Hz)|
Later on, we read that "oscillator drifts and path-induced Doppler shifts must be less than the tone spacing".
Why was this chosen like this? Wouldn't larger tone spacings provide better SNR decodability? At the very least, they should allow for higher Doppler shifts for the longer-period modes, shouldn't they?
Why isn't there an option in WSJT-X to manually increase the spacings for testing? I mean, is it just because there was no demand for it and it simply wasn't implemented, or is it because it inherently doesn't make sense to increase them?
I understand that signals with slower frequency changes generally occupy less bandwidth, but don't such tiny tone spacings also make decoding harder?