Spectral pattern and the perceptual fusion of harmonics. II. A special status for added components?

A single even-numbered harmonic can be segregated from an odd-harmonic complex more easily than its odd-numbered neighbors, for low fundamentals [Roberts and Bregman, J.Acoust. Soc. Am. 90, 3050-3060 (1991)]. It has been proposed that this effect occurs because the even harmonic is inconsistent with the regular spectral pattern formed by the odd harmonics. However, in this study we evaluate some alternative accounts of the effect. The stimuli were harmonic complex tones for which one of the components was cued by a preceding pure tone. Subjects were required to listen for the cued component and either to rate its clarity (experiments 1 and 3) or to judge its pitch in relation to the preceding tone (experiment 2). It is argued that the weaker fusion of an added even harmonic depends on its status as an added component rather than as an even harmonic (experiment 1), and on its immediate perceptual salience rather than on auditory learning. An even-odd difference was also found when even and odd harmonics were tested in an identical spectral context (experiment 2), and when one of the neighboring odd harmonics was removed (experiment 3). These results are consistent with the proposal that it is easier to segregate a harmonic from a periodic complex tone when it does not form part of the regular pattern of spectral spacing defined by the other harmonics.