Bienhold Grayson J, Meyer David P, Wischhoff Owen P, Chumbley Taylor J, Heimer Elle O, Moreira Elise A, Jiang Jack
Department of Otolaryngology-Head and Neck Surgery, School of Medicine and Public Health, University of Wisconsin-Madison, Madison, WI.
School of Music, University of Iowa, Iowa City, IA; Communication Sciences and Disorders Department, University of Iowa, Iowa City, IA; Janette Ogg Voice Research Center, Shenandoah Conservatory, Winchester, VA.
J Voice. 2025 Jun 24. doi: 10.1016/j.jvoice.2025.05.018.
This study uses acoustic and nonlinear dynamic analysis metrics to investigate the distinct vocal characteristics of five music genres-blues, country, folk, Italian opera, and rock. This research uses cepstral peak prominence (CPP) and correlation dimension (D2) values to quantify the levels of chaos and aperiodicity in the singing voices across these genres, providing insights into the acoustic makeups and vocal behaviors unique to each singing style. The aim is to determine if these measures are sufficiently sensitive to develop a comprehensive Voice Type Component Profile (VTCP), a measure of the portions of voice types in an acoustic signal, for each genre.
Fifty voice samples of the vowel /i/ from the Saarbrucken Voice Database were collected and analyzed as a control group representing healthy and typical phonation. These samples were compared with vocal samples extracted from performances in five musical genres: twenty nine from a popular blues music song, 29 from a popular country music song, 24 from a popular folk music song, 23 from a popular Italian opera song, and 29 from a popular rock music song. In addition to being compared to speech phonation, each musical genre was also compared to the others to explore genre-specific vocal characteristics. The singing samples were randomly selected from YouTube and analyzed for CPP, D2, and VTCP.
Many significant differences in acoustic metrics between the genres were found. Rock singing exhibited higher D2 values and lower CPP values compared with blues, country, and folk singing, indicating greater levels of aperiodicity. Blues singing had lower CPP values than the other genres. VTCP analysis revealed that rock singing had the lowest proportion of Type 1 signals and the highest proportion of Type 3 signals.
The findings highlight the unique vocal demands associated with each genre. Understanding these acoustic differences can inform singing instruction and strategies for optimizing vocal health, particularly for genres that exhibit higher levels of chaos (eg, rock). Future research should involve controlled studies with trained singers and advanced vocal analysis techniques to validate and refine these findings.
本研究使用声学和非线性动力学分析指标，来探究蓝调、乡村、民谣、意大利歌剧和摇滚这五种音乐流派独特的声乐特征。本研究使用谐波峰值突出度（CPP）和关联维数（D2）值，来量化这些流派演唱声音中的混沌和非周期性水平，从而深入了解每种演唱风格独特的声学构成和发声行为。目的是确定这些指标是否足够灵敏，以便为每种流派开发一种全面的嗓音类型成分剖面图（VTCP），这是一种衡量声学信号中嗓音类型比例的指标。
从萨尔布吕肯语音数据库中收集了五十个元音/i/的语音样本，并作为代表健康和典型发声的对照组进行分析。这些样本与从五种音乐流派的表演中提取的声乐样本进行比较：29个来自一首流行的蓝调音乐歌曲，29个来自一首流行的乡村音乐歌曲，24个来自一首流行的民谣音乐歌曲，23个来自一首流行的意大利歌剧歌曲，以及29个来自一首流行的摇滚音乐歌曲。除了与语音发声进行比较外，每种音乐流派还相互进行比较，以探索特定流派的声乐特征。演唱样本从YouTube上随机选取，并分析其CPP、D2和VTCP。
发现各流派之间在声学指标上存在许多显著差异。与蓝调、乡村和民谣演唱相比，摇滚演唱的D2值更高，CPP值更低，表明其非周期性水平更高。蓝调演唱的CPP值低于其他流派。VTCP分析表明，摇滚演唱中1型信号的比例最低，3型信号的比例最高。
研究结果突出了每种流派独特的发声要求。了解这些声学差异可为歌唱指导和优化嗓音健康的策略提供参考，特别是对于那些混沌水平较高的流派（如摇滚）。未来的研究应包括对受过训练的歌手进行对照研究，并采用先进的声乐分析技术，以验证和完善这些研究结果。
