Exploring the Impact of Nonlinear Distortion on Perceived Musical Gestures and Intentions

All rights remain with the authors under the Creative Commons Attribution 4.0 International License (CC BY 4.0). Proc. of the 17th Int. Symposium on Computer Music Multidisciplinary Research, London, United Kingdom, 2025 Exploring the Impact of Nonlinear Distortion on Perceived Musical Gestures and Intentions. Agathe Mangialomini1,2*, Richard Berrebi 2*, Etienne Thoret 3+ and Richard Kronland-Martinet 2+ 1 University of Tours, CNRS, Centre d'études supérieures de la Renaissance (CESR) UMR7323 2 Aix-Marseille University, CNRS, Perception Representation Image Sound Music (PRISM) UMR7061 3 Aix-Marseille University, CNRS, Institut de Neurosciences de la Timone (INT), UMR7289 [email protected] [email protected] *The two authors contributed equally to this work +Senior co-authorship Abstract. Distortion is a defining timbral characteristic in many current popular music genres and arises from nonlinear processing that introduces complex spectral changes such as increased brightness and auditory roughness. Despite its cultural and expressive significance, the perceptual, affective and embodied effects of distortion remain underexplored. This study investigates how spectral modifications induced by nonlinear distortion on guitar chords influence listeners’ perception of the emotional valence, physical effort, movement amplitude, and symbolic sound categorization. Using 20 distorted electric guitar stimuli generated with an implementation of the Boss DS-1 pedal, eight participants rated six perceptual and affective dimensions. Acoustic descriptors (spectral centroid and auditory roughness) were computed for each stimulus to model timbral changes. Results indicate that increased roughness correlates with more negative emotional valence and trends toward greater perceived effort, while brightness strongly relates to Bouba/Kiki judgment. Professional experience in sound-related fields and frequency of musical practice were also found to increase valence and amplitude ratings. These findings show that specific acoustic features produced by non-linear distortion differentially shape judgments of emotional valence, physical effort, and sound-symbolic categorization, with these effects Proc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 114 A. Mangialomini and R. Berrebi modulated by listeners’ prior experience. Future work will extend the descriptor set and aims to inform a multidimensional model of distorted timbres. Keywords: Music Cognition, Perception, Nonlinear Distortion, Timbres, Guitar. 1 Introduction Distortion is ubiquitous in current popular music, particularly in genres such as rock, metal, and punk, where it contributes to the timbral identity of guitar sounds which have become a staple of these genres. This characteristic sonic quality arises from non-linear distortion, which introduces new frequency components that significantly modify its spectral structure [15]. This results in complex timbres characterized by high-frequency energy, increased inharmonicity, and pronounced auditory roughness [6, 25]. From a perceptual standpoint, distorted timbres are often linked to negative valence, high arousal, and intense emotions like anger or discomfort [3, 6]. However, their influence extends beyond affective responses. Wallmark et al. [26] show that distorted or more broadly, the category of ‘noisy’ timbres, characterized by spectral features such as inharmonicity, aperiodic components, and auditory roughness [15, 21] also engage motor representations. Their neuroimaging results further support this by revealing activation in motor-related brain regions when listeners hear such timbres, even in the absence of any overt movement. Studies have already highlighted the possibility of recalling movements from natural sounds, such as those produced by the friction sounds of the pencil on the paper when someone is drawing [23, 24] or by impacts [12]. In the auditory perception context, this aligns with the concept of motor resonance [17] and the Shared Affective Motion Experience model [14, 16], which posits that sound perception involves internal simulation of expressive motor actions. In musical contexts such as punk or metal, distortion may therefore function not only as a stylistic timbral signature but also as an acoustic means of eliciting embodied and affectively charged responses [4, 9], shaped by social and cultural frameworks. Despite its cultural and perceptual significance, research on distortion has predominantly focused on its acoustic properties [8, 18] or its impact on speech and music comprehension [15]. More recent work has begun to examine perceptual preferences related to distortion responses to musical chords and the perception of distorted guitar sounds in rock and metal [6, 25, 20, 11]. However, the relationship between the specific acoustic features of distortion and the range of perceptual, emotional, and embodied reactions they elicit remains only partially understood. To address this gap, a study in progress [2] conducted an extensive acoustic analysis by applying 2,420 unique parameter combinations of the Boss DS1 distortion pedal, systematically varying distortion, tone, and level settings, to a single guitar chord. From this comprehensive dataset, a subset of 20 stimuli was carefully extracted to span the perceptual timbre space defined by brightness and roughness, serving as representative exemplars of the acoustic variation. These two descriptors were selected because they capture key spectral modifications introduced by distortion and are known to play a Proc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 115 Impact of Nonlinear Distortion on Perceived Musical Gesture and Intention central role in timbre perception, particularly in relation to emotional arousal, perceptual salience, and auditory dissonance. Brightness, reflecting the concentration of spectral energy in higher frequencies [5], was quantified via the spectral centroid [1, 13], while roughness, arising from beating interactions among closely spaced partials in the high-frequency spectrum [10], was modeled following Zwicker and Fastl [27]. These 20 selected stimuli were then used in a perceptual dissimilarity rating task, followed by multidimensional scaling (INDSCAL) to uncover the underlying perceptual structure. Results revealed strong listener agreement (high intraclass correlation coefficients), with two main perceptual dimensions emerging from MDS that were closely aligned with the acoustic descriptors: brightness was associated with the spectral centroid, which in turn varied systematically with tone and distortion settings, while roughness was linked to auditory roughness measures, which increased with higher distortion levels. Building on these findings and previous research [6, 26], the present research investigates how distortion-induced spectral modifications influence listeners’ affective, motoric, and symbolic responses to sound. Using the same dataset of 20 distorted guitar stimuli, participants evaluated each sound along six dimensions: 1. Emotional valence (7-point Likert scale) [19], 2. Perceived Physical Effort required to produce the sound (7-point Likert) [26], 3. Imagined amplitude of movement involved in its production (7-point Likert), 4. Appreciation (7-point Likert) [26], 5. Evoked emotion (categorical: joy, sadness, anger, fear, serenity, disgust, surprise), 6. Sound-symbolic categorization (forced choice: Bouba or Kiki) [22], We hypothesize that increasing levels of nonlinear distortion, reflected acoustically by higher spectral centroid and auditory roughness, will be linked to perceptions of greater physical effort, larger imagined movement amplitude, and more negative emotional valence. We further expect these perceptual dimensions to systematically correlate with the acoustic descriptors, supporting the embodied and affective aspects of timbre perception in distorted guitar sounds. Finally, as an exploratory aim, we investigate whether symbolic sound categorization [22], and discrete emotional labels are influenced by these acoustic features, revealing metaphorical or embodied mappings between timbral structure and semantic meaning. Further investigation will also be conducted to understand the potential influence of other timbre descriptors on the evaluation of the stimuli, while music expertise and listening habits will also be explored in relation to the different scales. 2 Method 2.1 Participants Eight listeners participated in the study, out of the twenty yet to be recruited. Among them, five had prior musical experience. The average age of participants was 29.0 years (SD = 12.06 years), with an age range spanning 21 to 58 years, 37% of them being Proc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 116 A. Mangialomini and R. Berrebi women. All participants were either native French speakers or fluent in French, and none reported any hearing impairments. 2.2 Stimuli and Apparatus We used the twenty distorted guitar sounds generated in the previous study [2] by applying systematically varied presets of a MATLAB implementation of the Boss DS-1 distortion pedal to a A# major guitar chord, fully voiced across all six strings, recorded with Ibanez PF300 via Audient iD14 and no amplification. The resulting sounds were loudness-equalized at -18 LKFS (ITU-R BS.1770). To control for loudness-related bias, six expert listeners adjusted the playback level of each stimulus to match the loudness of a reference stimulus using a custom MATLAB interface. Adjustments were made via a slider until the perceived loudness matched the reference. Final values were derived from the median adjustment per stimulus. Loudness estimates ranged between approximately 46.4 and 47.8 dB SPL across stimuli, as measured using the Neumann KU100 artificial head microphone. Listeners were tested individually in a soundproof booth. Stimuli were presented with a MATLAB interface, on Sennheiser HD 800 S headphones using an Apple MacBook Pro 2024 and an audio interface Focusrite Scarlett 2i2 with a fixed volume. Stimuli were 3882ms long with a volume fade out with a decreasing exponential of 120 milliseconds to avoid clicking. 2.3 Design and Procedure Participants initially familiarized themselves with the stimulus set by freely exploring all 20 sounds presented in a pseudorandomized order. Subsequently, they completed a training phase consisting of four example stimuli accompanied by the evaluation scales. During the main perceptual task, participants listened to each stimulus without time constraints before providing ratings on six distinct dimensions: emotional valence, perceived physical effort, aesthetic appreciation, imagined movement amplitude, evoked emotion, and phonological symbolism (Bouba vs. Kiki). Ratings were obtained using Likert scales for continuous dimensions and categorical selections for discrete judgments. 3 Analysis and Results 3.1 Inter-listener agreement Inter-rater reliability was assessed using Intraclass Correlation Coefficients (ICC) to quantify agreement across raters. ICC(2,1) measures absolute agreement, while ICC(3,1) captures consistency in relative ranking regardless of systematic bias. The Valence, Physical Effort, Amplitude, and Appreciation scales showed low to moderate absolute reliability (ICC(2,1) = 0.14–0.23 [95% CI]), but good relative conProc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 117 Impact of Nonlinear Distortion on Perceived Musical Gesture and Intention sistency (ICC(3,1) = 0.76–0.86). In contrast, Evoked Emotion exhibited very low reliability (ICC(2,1) = 0.02, ICC(3,1) = 0.28), with high variability (SD ≈ 2.0, 6-point range), indicating weak agreement and subjectivity. The Bouba/Kiki scale showed negative ICCs, reflecting almost no inter-rater reliability and a lack of consensus. 3.2 Correlational Analysis Between Acoustic Features and Perceptual Ratings Correlation analyses reveal significant links between acoustic descriptors and perceptual evaluations. Roughness correlates negatively with emotional valence (ρ = -0.693, p = 0.026, df = 8), indicating that greater roughness is associated with more negative judgments, consistent with prior findings on aversive sound qualities [7]. Its correlation with perceived physical effort is marginal (ρ = 0.593, p = 0.071), with wide variability across responses. Other dimensions, such as amplitude and appreciation, show no significant associations, suggesting roughness relates more to emotion than motor evaluations. Brightness, in contrast, strongly correlates with Bouba/Kiki categorization (ρ = 0.738, p = 0.015, df = 8), with brighter timbres linked to the “Kiki”. It also shows moderate, non-significant correlations with appreciation (ρ = 0.607) and amplitude (ρ = - 0.611), indicating possible trends for further study. Notably, brightness does not correlate with valence, counter to expectations of positive or negative bias. 3.3 Effects of Musical Practice and Sound-Related Professional Experience Regression analyses indicated that valence ratings increased with both professional sound experience (β = 0.40, SE = 0.18, t = 2.20, p = 0.029) and frequency of musical practice (β = 0.32, SE = 0.09, t = 3.64, p < 0.001). Imagined movement amplitude was higher in participants with sound-related occupations (β = 0.46, SE = 0.17, t = 2.65, p = 0.0085), whereas perceived effort and appreciation were not significantly affected. Overall, professional experience exerted a stronger influence on valence and imagined movement, while practice frequency primarily impacted valence. In contrast, neither listening to a specific musical genre nor playing a specific instrument significantly altered participants’ overall mean ratings. 3.4 Mapping Perceptual Ratings onto Acoustic Dimensions Berrebi et al. [2] constructed timbre spaces by applying nonmetric multidimensional scaling to listener dissimilarity ratings of the stimuli. Participants rated the perceived differences between pairs of sounds, and these ratings were aggregated into dissimilarity matrices for each listener. Using the INDSCAL model, a group-level two-dimensional perceptual space was derived that captures shared patterns in these similarity judgments while accounting for individual differences. The dimensions of this space correspond to underlying acoustic features, with one space (fig.1A) primarily reflecting brightness as measured by spectral centroid and the second reflecting roughness (fig.1B) based on Zwicker and Fastl’s model. Proc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 118 A. Mangialomini and R. Berrebi To examine the relationship between the perceptual ratings and the underlying acoustic dimensions, we focused on four of the six scales (emotional valence, perceived physical effort, appreciation, and imagined movement amplitude) that could be meaningfully represented as continuous dimensions. For each of these scales, a multiple linear regression was conducted, with the two MDS dimensions as predictors. The resulting regression coefficients were then used to construct directional vectors from the origin, allowing interpretation of the perceptual space. Fig. 1. (A) Regression vector projections of perceptual ratings in the MDS space derived from brightness. The distortion dimension (Dimension 2) is consistently associated with more negative valence, greater perceived effort, increased appreciation, and reduced imagined movement amplitude. This latter effect may relate to the prominence of low-frequency content preserved through filtering. Additionally, increased distortion combined with stronger high-frequency energy (Dimension 1) is linked to a greater tendency to categorize sounds as ‘Kiki’, suggesting a connection between spectral timbre features and phonological symbolism. (B) Regression vector projections of perceptual ratings in the MDS space derived from roughness. Although the perceptual structure was less clearly defined, similar patterns emerge: greater distortion was associated with more negative valence, higher perceived effort, greater appreciation, and lower imagined movement amplitude. The categorical emotion scale was excluded from the regression analysis due to challenges in interpretability. 4 Discussion This work in progress offers an initial investigation into how nonlinear distortion affects listeners’ perceptual, affective, and motor-related responses through timbral changes. Results show a significant negative correlation between roughness and emotional valence, indicating that increased distortion relates to more negative emotional responses. This trend corroborates the theoretical prediction that increased auditory roughness evokes affective states associated with tension or discomfort and aligns with previous findings in the literature [7, 26]. While roughness also trends toward a positive association with perceived physical effort, this relationship does not reach statistical significance, making it not yet possible to conclude on any link with other sensory-motor responses. Furthermore, moderate, albeit non-significant, associations between brightProc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 119 Impact of Nonlinear Distortion on Perceived Musical Gesture and Intention ness and other perceptual dimensions indicate that multiple acoustic cues could collectively inform an integrated affective-sensorimotor response to distorted timbres. Individual differences were notable, highlighting the importance of considering listener characteristics. Regression analyses revealed that professional experience with sound and frequency of musical practice influenced valence ratings and imagined movement amplitude, with professional experience exerting a stronger effect. In contrast, neither listening to a specific musical genre nor playing a specific instrument significantly altered overall ratings. These results suggest that prior auditory expertise shapes how listeners perceive and emotionally respond to timbral distortion. Overall, these findings highlight the necessity of a multidimensional analytical approach to fully characterize the perceptual consequences of nonlinear distortion. In this sense, this analysis represents only a first step in understanding the complex interplay between acoustic features and embodied or affective judgments. The perceptual scales used here will be further examined to explore their internal structure and interrelationships. In parallel, data collection is ongoing, and results will be consolidated upon completion of the full sample of 20 participants, which will help improve statistical power and the robustness of observed effects. Concurrently, the study will be expanded by incorporating additional timbre descriptors, including inharmonicity, spectral flux, or temporal envelope characteristics. These parameters may provide a more nuanced understanding of how distortion transforms timbral identity and how listeners map those changes onto expressive, emotional, and motor experiences. This work thus seeks to contributes to bridging low-level acoustic descriptors and higher-order embodied judgments, and to advance our understanding of how distortion shapes the experiential quality of musical timbre. Acknowledgments. Research supported by grants ANR-16-CONV-0002 (ILCB), ANR24CE38-4175 (PSIND), and the Excellence Initiative of Aix-University (A*MIDEX). AM doctoral fellowship is supported by a CNRS 80-prime fellowship. The authors are thankful to Solveig Serre and Luc Robène for their thoughtful comments and support along this study. Disclosure of Interests. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. References 1. Beauchamp, J.: Synthesis by Spectral Amplitude and “Brightness” Matching of Analyzed Musical Instrument Tones. J. Audio Eng. Soc. (1981). 2. Berrebi, R. et al.: The impact of nonlinear distortion on the perceived brightness and roughness of electric guitar chords, (Submitted). 3. Daniel T. Blumstein et al.: The sound of arousal in music is context-dependent. Biol. Lett. 8, 5, 744–747 (2012). https://doi.org/10.1098/rsbl.2012.0374. Proc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 120 A. Mangialomini and R. Berrebi 4. Godøy, R.I., Leman, M. eds: Musical Gestures: Sound, Movement, and Meaning. Routledge, New York (2010). https://doi.org/10.4324/9780203863411. 5. Grey, J.M., Gordon, J.W.: Perceptual effects of spectral modifications on musical timbres. Acoust. Soc. Am. J. 63, 1493–1500 (1978). https://doi.org/10.1121/1.381843. 6. Herbst, J.-P.: Distortion and Rock Guitar Harmony. Music Percept. 36, 4, 335– 352 (2019). https://doi.org/10.1525/mp.2019.36.4.335. 7. Juslin, P.N., Laukka, P.: Communication of emotions in vocal expression and music performance: Different channels, same code? Psychol. Bull. 129, 5, 770–814 (2003). https://doi.org/10.1037/0033-2909.129.5.770. 8. Le Brun, M.: Digital Waveshaping Synthesis. Audio Eng. Soc. (1979). 9. Leman, M. et al.: Sharing Musical Expression Through Embodied Listening: A Case Study Based on Chinese Guqin Music. Music Percept. 26, 3, 263–278 (2009). https://doi.org/10.1525/mp.2009.26.3.263. 10. Lichte, W.H.: Attributes of complex tones. J. Exp. Psychol. 28, 6, 455–480 (1941). https://doi.org/10.1037/h0053526. 11. Lilja, E.: Dealing with the 3rd: Anatomy of distorted chords and subsequent compositional features in classic heavy metal. In: Karjalainen, T.-M., Kärki, K. (eds.) Modern Heavy Metal: Markets, Practices and Cultures. International Academic Research Conference, Helsinki, Finland: Conference Proceedings, pp. 393–403. Aalto University, Helsinki. (2015). 12. Lutfi, R.A. et al.: Auditory discrimination of force of impact. J. Acoust. Soc. Am. 129, 4, 2104–2111 (2011). https://doi.org/10.1121/1.3543969. 13. McAdams, S. et al.: Discrimination of musical instrument sounds resynthesized with simplified spectrotemporal parameters. J. Acoust. Soc. Am. 105, 2, 882–897 (1999). https://doi.org/10.1121/1.426277. 14. Molnar-Szakacs, I., Overy, K.: Music and mirror neurons: from motion to ’e’motion. Soc. Cogn. Affect. Neurosci. 1, 3, 235–241 (2006). https://doi.org/10.1093/scan/nsl029. 15. Moore, B.C.J.: Measuring and Predicting the Perceived Quality of Music and Speech Subjected to Combined Linear and Nonlinear Distortion. J Audio Eng Soc. 52, 12, (2004). 16. Overy, K., Molnar-Szakacs, I.: Being Together in Time: Musical Experience and the Mirror Neuron System. Music Percept. 26, 5, 489–504 (2009). https://doi.org/10.1525/mp.2009.26.5.489. 17. Rizzolatti, G., Craighero, L.: THE MIRROR-NEURON SYSTEM. Annu. Rev. Neurosci. 27, Volume 27, 169–192 (2004). https://doi.org/10.1146/annurev.neuro.27.070203.144230. 18. Roads, C.: A Tutorial on Non-Linear Distortion or Waveshaping Synthesis. Comput. Music J. 3, 2, 29 (1979). https://doi.org/10.2307/3680281. 19. Russell, J.A.: A circumplex model of affect. J. Pers. Soc. Psychol. 39, 6, 1161– 1178 (1980). https://doi.org/10.1037/h0077714. Proc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 121 Impact of Nonlinear Distortion on Perceived Musical Gesture and Intention 20. Scotto, C.: The Structural Role of Distortion in Hard Rock and Heavy Metal. Music Theory Spectr. 38, 2, 178–199 (2016). https://doi.org/10.1093/mts/mtw013. 21. Sethares, W.A.: Tuning, timbre, spectrum, scale. Springer, London (2005). 22. Spence, C.: Crossmodal correspondences: A tutorial review. Atten. Percept. Psychophys. 73, 4, 971–995 (2011). https://doi.org/10.3758/s13414-010-0073-7. 23. Thoret, E. et al.: From sound to shape: Auditory perception of drawing movements. J. Exp. Psychol. Hum. Percept. Perform. 40, 3, 983–994 (2014). https://doi.org/10.1037/a0035441. 24. Thoret, E. et al.: Hearing elliptic movements reveals the imprint of action on prototypical geometries. Cognition. 238, 105478 (2023). https://doi.org/10.1016/j.cognition.2023.105478. 25. Virtala, P. et al.: Distortion and Western Music Chord Processing. Music Percept. 35, 3, 315–331 (2018). https://doi.org/10.1525/mp.2018.35.3.315. 26. Wallmark, Z. et al.: Embodied Listening and Timbre. Music Percept. 35, 3, 332–363 (2018). https://doi.org/10.1525/mp.2018.35.3.332. 27. Zwicker, E., Fastl, H.: Psychoacoustics. Springer Berlin Heidelberg, Berlin, Heidelberg (1999). https://doi.org/10.1007/978-3-662-09562-1. Proc. of the 17th International Symposium on CMMR, London, UK, Nov. 3-7, 2025 122