I have just discovered the Ctrl+M placing of a label during recording and would like to realise and try out a long-time ambition as a language teacher.

The idea is simple: while recording a conversation with a student (or playing back one previously recorded) I hear a mistake.  I press a key or key-combo and a label is placed between 3 and 6 seconds (random or user-defined) before the key press.

When you play back the recording, you can go directly to the labels and play, and the student can try to spot what he or she got wrong.

It can't be that hard, but I can't see how Nyquist handles keypresses or how to read the event time; placing the label seems ok, but it's taking time and I feel I'm reinventing the wheel!

Any ideas out there?