The last log was about a fixed-width short-pulse generator. I don't have the PCBs back to weigh in with the final results, but indications are that it will generate a brief, well-defined pulse width for each input edge. Using two of these, you can build a coincidence detector:

There is one of the pulse generator circuits on each input. When an input transitions high, the corresponding pulse generator emits a short pulse. If the pulses overlap for long enough, the middle AND gate outputs a pulse, triggering the output pulse generator to create a pulse. This final stage could be replaced with something else, if you wanted.

There is some dead-time, of course, but that just limits the repetition frequency.

A larger problem may be metastability in the output stage. It's possible the input signals to be just inside the coincidence window, which could cause a short pulse out of the AND gate, leading to a metastable state in the output flop. I'm not sure how to deal with this.

Ignoring the metastability problem for the moment, the idea simulates well:

The bottom two traces are the inputs. The next two are the pulse generator outputs. Second from the top is the AND gate output, and the top trace is the output pulse.

This is mildly interesting, but I would like to detect coincidence within a much smaller window.