But sometimes I have a trouble recognizing sound signals (not sure if anybody mentioned it).
Probably not everybody has perfect ear to distinguish sound tones if there is a significant time interval between them. Could you please use more complicated sound samples so that it is easier to distinguish them? Or you could just combine the existing two (instead of "high" and "low" you could play "low-high" and "high-low"). Should be easy to implement.
I could look into this. Down sounds could have a downward pitch ramp, and upward for Up. That way it's not just the same two sounds at different pitch.
That'd be great! Also: Loudness. The default, 100% sound volume can be... painful, when wearing a headset.