"Discrimination of Single- and Triple-cycle Sinusoidal Acoustic Signals" is a great example of a paper on neurology. Mark & Rattay (1990) were among the first to have studied the effect of the number of periods on frequency discrimination with tones as short as one wavelength. They conducted an investigation to determine the reaction of the human ear to a single-, double-, and triple cycle sinusoidal acoustic stimuli of various frequencies. The objective of the investigation was to determine the effectiveness of frequency distribution with short acoustic signals. Tests were conducted to determine the number of cycles for different frequencies to recognize an acoustic signal as a tone.
Four subjects with normal hearing were recruited for the study. Subjects were presented with four frequencies, with one, two, or three sinusoidal periods in the first experiment. In the experiment, a docetic presentation was made through an analog computer driving headphones via a recorder and receiver. At the beginning of each experiment, a training period comprising of multiple experimental runs with feedback and repetitions on the subject’ s demand was conducted. When the subject was satisfied with the training, the experiment was started.
20 pairs of signals, comprising of each frequency and each number of runs were presented to the subject; one pair being the test frequency and the other being another signal with a higher frequency. The signal pairs were identical in structure. The subject had to identify the highest among the two presented. The subjects recognized frequency differences in the range of one semitone for single-period acoustic signals. For 75% criterion, where frequencies were 256 Hz, 512 Hz, 1024 Hz, and 2048 Hz, frequency difference linens were 12.4%, 14.8%, 5.3%, and 8.9% respectively.
Increasing the number of periods for signals from 2 to 3 caused no better perception of frequency difference. The best results were obtained for 1024 Hz. Also, when the frequency difference was 20% or 30%, 100% correct answers were obtained. 75% limit of correct answers was in the order of semitone, on presentation of two or three periods. Five subjects had participated in the study, including a professional musician with absolute pitch. The first experiment had four subjects, without any statement regarding the criteria for inclusion or exclusion.
Results displayed as a percentage of correct answers were a function of frequency difference. Since every point represented a mean value of the subjects, the effect of inclusion of the musician could have had a significant impact on the overall results. The best results were obtained for 1024 Hz, with frequency discrimination of 1 % with 70% correct answers. This has not been represented on any of the graphs, where the minimum frequency difference is 2.5%. As the graphs were showing averages of all subjects, we could assume that there was a large standard deviation of results, which might have been caused by the inclusion of the musician.
Caution must be exercised in interpretation since results of individual subjects or analysis of variance is not available. The study provides a good basis for designing an improved experiment by not including musicians and excluding any bias. Also, the difference in frequency distribution as the number of cycles is increased to more than three cycles could be investigated. The enhanced experiment would provide more information on the effects of the repeats of a waveform.
ReferencesMark, H. & Rattay, F. (1990). Frequency discrimination of single-, double-, and triple-cycle sinusoidal acoustic signals. J. Acoust. Sec. Am. . 88 (1), 560-563.