Do we know for sure that the widths of the pulses matter at all? Maybe only the frequency of pulses matter. Maybe both.
Does the duty cycle (mark:space ratio) change as engine revs increase? If not, the tachometer can't simply be a low-pass filter feeding an ammeter. Some experimentation may be needed to see how the instrument reacts to various frequencies and duty cycles.
For the 12V signal from the gearbox, a simple voltage divider may be enough, say 4K7 and 6K8?
Paul