Lanning Joshua M, Stilp Christian
Department of Psychological and Brain Sciences, University of Louisville, 317 Life Sciences Building, Louisville, KY, 40292, USA.
Atten Percept Psychophys. 2020 Jul;82(5):2209-2214. doi: 10.3758/s13414-020-01980-w.
Perception of sounds occurs in the context of surrounding sounds. When spectral properties differ between earlier (context) and later (target) sounds, categorization of later sounds becomes biased through spectral contrast effects (SCEs). Past research has shown SCEs to bias categorization of speech and music alike. Recent studies have extended SCEs to naturalistic listening conditions when the inherent spectral composition of (unfiltered) sentences biased speech categorization. Here, we tested whether natural (unfiltered) music would similarly bias categorization of French horn and tenor saxophone targets. Preceding contexts were either solo performances of the French horn or tenor saxophone (unfiltered; 1 second duration in Experiment 1, or 3 seconds duration in Experiment 2) or a string quintet processed to emphasize frequencies in the horn or saxophone (filtered; 1 second duration). Both approaches produced SCEs, producing more "saxophone" responses following horn / horn-like contexts and vice versa. One-second filtered contexts produced SCEs as in previous studies, but 1-second unfiltered contexts did not. Three-second unfiltered contexts biased perception, but to a lesser degree than filtered contexts did. These results extend SCEs in musical instrument categorization to everyday listening conditions.
声音感知发生在周围声音的背景下。当前期（背景）声音和后期（目标）声音的频谱特性不同时，后期声音的分类会因频谱对比效应（SCEs）而产生偏差。过去的研究表明，SCEs会使语音和音乐的分类产生偏差。最近的研究将SCEs扩展到自然聆听条件下，即（未滤波的）句子的固有频谱成分会使语音分类产生偏差。在这里，我们测试了自然（未滤波的）音乐是否同样会使法国号和次中音萨克斯管目标的分类产生偏差。前期背景要么是法国号或次中音萨克斯管的独奏表演（未滤波；实验1中时长为1秒，实验2中时长为3秒），要么是经过处理以强调法国号或萨克斯管频率的弦乐五重奏（滤波；时长为1秒）。两种方法都产生了SCEs，在法国号/类似法国号的背景之后产生了更多的“萨克斯管”反应，反之亦然。1秒的滤波背景产生了与之前研究相同的SCEs，但1秒的未滤波背景则没有。3秒的未滤波背景使感知产生了偏差，但程度小于滤波背景。这些结果将乐器分类中的SCEs扩展到了日常聆听条件。
