Pitch and attention

Abstract

Listen

Translate article

We have been investigating the ability of listeners to detect Morse-Code-like messages solely on the basis of pitch from a mixture of three such messages heard binaurally through headphones, Each message consists of tone bursts of a particular frequency having two “longs” (tone bursts of duration 400 msec) and two “shorts” (tone bursts of duration 200 msec) in various combinations. Silent intervals between bursts for a particular message invariably have durations of 100 msec. Tones are adjusted beforehand to have equal loudness if each were “on” solely and continuously. All messages have the same duration, Two basic variants of the experiment have been investigated so far. In the first, listerners receive no pitch cue and are asked to find the “middle” (or “low” or “high”) pitch message. In the second, listeners hear a pure tone of frequency equal to the desired message (pitch cue) prior to hearing the message mixture and are asked to determine the message of that pitch. Cues and/or messages can be heard as often as the subject desires. If all three messages have frequencies below roughly 1500 Hz but are well separated with respect to the critical band, then “getting the message” is easy, pitch cue or not, unless two of the frequencies are an octave apart. In that case, the pitch cue is essential for high recognition rates. If at least two of the constituent frequencies are above 1500 Hz, then “getting the message” is highly dependent on the pitch cue. [Work supported by NIH.]