Attached is a fair quality ~10min .WAV file with around 300 short auditory probes, ISI 1sec. can run in the background of any ML program if placed as "backsound" in the relevant questionare (Jarvis's idea). mind-you that responses can only be externally recorded (if ML is busy with something else) but in many experiments (e.g. using random number generation as a wmk load task) that's not a problem.