Jhang Geng-Yan, Ueda Kazuo, Takeichi Hiroshige, Remijn Gerard B, Hasuo Emi
Human Science International Course, Graduate School of Design, Kyushu University, Fukuoka, Japan.
Department of Acoustic Design, Faculty of Design/Research Center for Applied Perceptual Science, Kyushu University, Fukuoka, Japan.
PLoS One. 2025 Jun 5;20(6):e0323964. doi: 10.1371/journal.pone.0323964. eCollection 2025.
Two rapidly alternating tones with different pitches may be perceived as one integrated stream when the pitch differences are small or two separated streams when the pitch differences are large. Likewise, timbre differences between two tones may also cause such sequential stream segregation. Moreover, the effects of pitch and timbre on stream segregation may cancel each other out, which is called a trade-off. However, how timbre differences caused by specific patterns of spectral shapes interact with pitch differences and affect stream segregation has been largely unexplored. Therefore, we used stripe tones, in which stripe-like spectral patterns of harmonic complex tones were realized by grouping harmonic components into several bands based on harmonic numbers and removing harmonic components in every other band. Here, we show that 2- and 4-band stimuli elicited distinctive stream segregation against pitch proximity. By contrast, pitch separations dominated stream segregation for 16-band stimuli. The results for 8-band stimuli most clearly showed the trade-off between pitch and timbre on stream segregation. These results suggest that the stimuli with a small number ([Formula: see text]4) of bands elicit strong stream segregation due to sharp timbral contrasts between stripe-like spectral patterns, and that the auditory system looks to be limited in integrating blocks of frequency components dispersed over frequency and time.
当音高差异较小时,两个音高不同且快速交替的音调可能会被感知为一个整合的流;而当音高差异较大时,则会被感知为两个分离的流。同样,两个音调之间的音色差异也可能导致这种序列流分离。此外,音高和音色对流分离的影响可能会相互抵消,这被称为权衡。然而,由特定频谱形状模式引起的音色差异如何与音高差异相互作用并影响流分离,在很大程度上尚未得到探索。因此,我们使用了条纹音调,其中谐波复合音的条纹状频谱模式是通过根据谐波数将谐波成分分组到几个频段,并去除每隔一个频段中的谐波成分来实现的。在这里,我们表明,2频段和4频段刺激针对音高接近度引发了独特的流分离。相比之下,对于16频段刺激,音高分离主导了流分离。8频段刺激的结果最清楚地显示了音高和音色在流分离上的权衡。这些结果表明,频段数量较少(≤4)的刺激由于条纹状频谱模式之间明显的音色对比而引发强烈的流分离,并且听觉系统在整合分散在频率和时间上 的频率成分块方面似乎存在限制。