Mic and speaker thresholds act directly on the normalized, filtered signals.
Ensures only full events (ones that have a start and a stop to the event) are detected, ones that have a duration for less than 100 ms are excluded.
When running a test, it will run on the full signal (it just won't save them). It runs on the signal as determined by the bottom most start/stop user input values.
Visualization updates:
Color changes to the Event Detection traces, make it easier to see the filtered signals.
In the main window, clean up speech signal auto-scaling.
Event Detection updates:
Visualization updates: