Some time ago, I seem to recall hearing or reading of an audio experiment where the very short initial transients of one musical instrument (A) were spliced in front of the much longer steady state or sustain sounds of another instrument (B). Human listeners in this experiment, upon hearing these composite sounds, were reported to recognize or hear the first instrument (A) with some reliability even though the initial transient or attack was only a very short portion of the entire note played. (Anyone have a reference to this research?)

If the above is correct, what type of signal processing analysis of just the initial transient of a note from a musical instrument, before steady-state periodicity or stationarity, might be used to try to do musical instrument recognition or separation?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.