Pitch Control in the Human Voice: A Reply to Rostron

Rostron (1976) has been unable to repeat our results (Deutsch and Clarkson, 1959) showing that as a delay is interposed between vocal output and auditory input, so does the frequency of the vibrato decrease and the amplitude of the vibrato increase. This result reported by us is large and obvious (Figures I and 2). In fact it is immediately audible. One wonders then why Rostron was not able to repeat such a result. The most obvious answer is that his subjects could still hear their own undelayed voice in spite of the amplified and delayed signal fed back to them through the earphones. When performing our experiments we found that three main factors would make it possible for the subjects to disregard the delayed signal fed back to them. The first was inadequate loudness of the signal fed back through the earphones. The second was extended practice with the delayed signal. Subjects can learn to focus on that signal which enables them to maintain pitch with the minimum of excursion. The third condition which enables the subject to ignore the delayed voice is a change in its quality. The subject must accept the delayed voice as being his own. If the subject rejects it, then he will not correct his output on the basis of the strange voice that he hears. That one of these factors which militate against the appearance of the delay effect could have been operating is clear from Rostron’s description: “The auditory feedback was delayed by a PDP-8 computer. , . . After receiving instructions subjects experienced delayed auditory feedback (DAF) under the longest delay condition with both speech and singing. The intensity level of the note sung was then decided, followed by the determination of the gain of the auditory feedback. This was done by the experimenter continuously adjusting the gain of the amplifier and the subjects stating whether the feedback signal was too loud for comfort.” From Rostron’s description any of the three factors could have been present.