Hasil untuk "Music"

Max W. Y. Lam, Yijin Xing, Weiya You et al.

Autoregressive (AR) models have demonstrated impressive capabilities in generating high-fidelity music. However, the conventional next-token prediction paradigm in AR models does not align with the human creative process in music composition, potentially compromising the musicality of generated samples. To overcome this limitation, we introduce MusiCoT, a novel chain-of-thought (CoT) prompting technique tailored for music generation. MusiCoT empowers the AR model to first outline an overall music structure before generating audio tokens, thereby enhancing the coherence and creativity of the resulting compositions. By leveraging the contrastive language-audio pretraining (CLAP) model, we establish a chain of "musical thoughts", making MusiCoT scalable and independent of human-labeled data, in contrast to conventional CoT methods. Moreover, MusiCoT allows for in-depth analysis of music structure, such as instrumental arrangements, and supports music referencing -- accepting variable-length audio inputs as optional style references. This innovative approach effectively addresses copying issues, positioning MusiCoT as a vital practical method for music prompting. Our experimental results indicate that MusiCoT consistently achieves superior performance across both objective and subjective metrics, producing music quality that rivals state-of-the-art generation models. Our samples are available at https://MusiCoT.github.io/.

Zhuoyuan Mao, Mengjie Zhao, Qiyu Wu et al.

Recent advancements in music large language models (LLMs) have significantly improved music understanding tasks, which involve the model's ability to analyze and interpret various musical elements. These improvements primarily focused on integrating both music and text inputs. However, the potential of incorporating additional modalities such as images, videos and textual music features to enhance music understanding remains unexplored. To bridge this gap, we propose DeepResonance, a multimodal music understanding LLM fine-tuned via multi-way instruction tuning with multi-way aligned music, text, image, and video data. To this end, we construct Music4way-MI2T, Music4way-MV2T, and Music4way-Any2T, three 4-way training and evaluation datasets designed to enable DeepResonance to integrate both visual and textual music feature content. We also introduce multi-sampled ImageBind embeddings and a pre-LLM fusion Transformer to enhance modality fusion prior to input into text LLMs, tailoring for multi-way instruction tuning. Our model achieves state-of-the-art performances across six music understanding tasks, highlighting the benefits of the auxiliary modalities and the structural superiority of DeepResonance. We open-source the codes, models and datasets we constructed: github.com/sony/DeepResonance.

Yumin Kim, Seonghyeon Go

Audio and music generation systems have been remarkably developed in the music information retrieval (MIR) research field. The advancement of these technologies raises copyright concerns, as ownership and authorship of AI-generated music (AIGM) remain unclear. Also, it can be difficult to determine whether a piece was generated by AI or composed by humans clearly. To address these challenges, we aim to improve the accuracy of AIGM detection by analyzing the structural patterns of music segments. Specifically, to extract musical features from short audio clips, we integrated various pre-trained models, including self-supervised learning (SSL) models or an audio effect encoder, each within our suggested transformer-based framework. Furthermore, for long audio, we developed a segment transformer that divides music into segments and learns inter-segment relationships. We used the FakeMusicCaps and SONICS datasets, achieving high accuracy in both the short-audio and full-audio detection experiments. These findings suggest that integrating segment-level musical features into long-range temporal analysis can effectively enhance both the performance and robustness of AIGM detection systems.

Dariusz Piechota

This article is devoted to musical themes appearing in a selection of recent (as of 2025) streaming series. These works demonstrate a clearly discernible shift in the ways that the medium employs musical scores. One apparent reason for this shift is the shortening of opening sequences, which are now often limited to brief animated skits, at times lasting mere seconds — which severely reduces the viewer’s exposure to a musical score. Another reason must be considered in a broader context: the elevation of the once-overlooked ‘couch entertainment medium’ into a genuine artistic genre. This change encourages the viewer to pay attention to artistic means that exist outside of the plot — be they visual or, more to the point here, aural. In most series produced in the 1980s and 1990s, the opening sequence played a central role, embedding both the musical theme and the appearance of the actors (and characters) in the audience’s memory. Today, musical openings — scored by the show’s main theme — are increasingly relics of the past, joining the growing catalogue of cultural retro-objects. Songs used in contemporary streaming series are not selected at random — they are deliberately chosen by the producers. These musical choices play an important role in shaping the narrative world, as they often reflect the characters’ inner lives or act as commentary on the unfolding story. The article focuses on musical score in the crime series Forst (2024), which features a wide range genres, from mainstream pop, through folk and soft rock, to more indie-aligned dream pop. In some cases, the visual storytelling that accompanies certain songs (such as Wonderful Life) closely mirrors the form of a music video, as it is similarly dominated by fragmentation, multiple shots, and contrast. This interconnection is a pointed example of audiovisuality, a phenomenon based on purposeful crossing of the boundaries between media. The blurring of genre lines becomes especially clear in scenes or sequences that directly incorporate musical pieces. These are typically one-minute long or otherwise short fragments that give the impression of watching a music video. Longer musical sequences, by contrast, require more viewer engagement. They invite interpretative reflection, and even deconstruction, prompting questions such as: Why was this particular piece chosen for this scene? How do the lyrics of the song relate to the world presented onscreen? Recent streaming series encourage viewers to become active participants in their experience rather than passive consumers of entertainment.

Dianna Vidas, Romina Carrasco, Ryan M. Kelly et al.

Music is a valuable non-pharmacological tool that provides benefits for people with dementia, and there is interest in designing technologies to support music use in dementia care. To ensure music technologies are appropriately designed for supporting caregivers and people living with dementia, there remains a need to better understand how music is currently used in everyday care at home. We aimed to understand how people with dementia and their caregivers use music technologies in everyday caring, as well as challenges they experience using music and technology. This study used a mixed methods design. A survey was completed by 77 caregivers and people with dementia to understand their use of music and technology. Of these, 18 survey respondents (12 family caregivers, 6 people living with dementia) participated in focus groups about their experiences of using music and technology in care. Transcripts were analysed with reflexive thematic analysis. Most survey respondents used music often in their daily lives, reporting a range of music technologies such as CDs, radio, and streaming. Focus groups highlighted benefits and challenges of music technologies in everyday care. Participants used music and music technologies to regulate mood, provide joy, facilitate social connection, encourage reminiscence, provide continuity before and after diagnosis, and to make caregiving easier. Challenges of using music technology in care included difficulties staying up to date with evolving technology, and low self-efficacy for technology use expressed by people living with dementia. Evidently, people living with dementia and their caregivers use music technologies to support their everyday care needs. Results suggest opportunities to design technologies enabling easier access to music and supporting people living with dementia with recreational and therapeutic music listening and music-based activities.

ChungHa Lee

In this paper, I have implemented a timbre visualization system called Tailors. Through the experiment with 27 MIR users, Tailors was found to be effective in conveying timbral warmth, brightness, depth, shallowness, hardness, roughness, and sharpness features of music compared to the only music condition and basic visualization. All scores of Tailors in the music imagery and music entertainment surveys were valued highest among the three conditions. Multiple linear regression analysis between timbre-imagery and imagery-entertainment shows significant and positive correlations. Coefficients comparing results from Fisher Transformation show that Tailors made user's music entertainment better through improved music visual imagery. The post-survey result represents that Tailors ranked first for the best timbre expression, music experience, and willingness to use it again. While some users felt a burden in the eye, Tailors left the future work of the data-driven approach of the mapping rule of timbre visualization to gain consent from many users. Furthermore, reducing timbre features to focus on features that Tailors can express well was also discussed, with future work of Tailors in a more artistic way using the sense of space.

Eileen Han, Lauren K Lempert, Francesca Vescia et al.

Abstract BackgroundElectronic cigarettes (e-cigarettes) are the most used form of tobacco products among adolescents and young adults, and Vuse is one of the most popular brands of e-cigarettes among US adolescents. In October 2021, Vuse Solo became the first e-cigarette brand to receive marketing granted orders (MGOs) from the US Food and Drug Administration (FDA), authorizing its marketing and their tobacco-flavored pods. Vuse Ciro and Vuse Vibe, and their tobacco-only (“original”) e-liquids, were authorized for marketing in May 2022 and Vuse Alto tobacco-flavored devices were authorized in July 2024. These marketing authorizations are contingent upon the company adhering to the MGOs’ stated marketing restrictions, including reducing exposure and appeal to youth via digital, radio, television, print, and point-of-sale advertising. ObjectiveIn this study, we analyzed the official social media channels of Vuse (Instagram and Facebook) to examine how Vuse marketed its products on social media and whether these marketing posts contain potentially youth-appealing themes. MethodsWe conducted content analysis of the official RJ Reynolds Vapor Company Instagram and Facebook accounts. We collected all posts from October 10, 2019, when RJ Reynolds Vapor Company submitted its premarket tobacco product application to the FDA, to February 21, 2022, to cover the first winter holiday season after the MGO. Two coders developed the codebook with 17 themes based on the Content Appealing to Youth index to capture the posts’ characteristics and potentially youth-appealing content. We calculated the percentage of posts in which each code was present. ResultsA total of 439 unique posts were identified. During this study’s period, there were no posts on Instagram or Facebook marketing Vuse Solo (the authorized product at that time). Instead, Vuse Alto (unauthorized to date of study) was heavily marketed, with 59.5% (n=261) of the posts specifically mentioning the product name. Further, “Vuse” more generally was marketed on social media without differentiating between the authorized and unauthorized products (n=182, 41.5%). The marketing messages contained several potentially youth-appealing themes including creativity or innovation (n=189, 43.1%), individuality or freedom (n=106, 24.2%), and themes related to art (n=150, 34.2%), music (n=77, 17.5%), sports (n=125, 28.5%), nature (with n=49, 11.2% of the posts containing flora imageries), alcohol imagery (n=10, 2.3%), and technology (n=6, 1.4%). ConclusionsAlthough Vuse Alto e-cigarettes had not yet obtained FDA marketing authorization during the 28 months of data collection, they were the primary Vuse e-cigarette devices marketed on social media. Vuse social media posts use themes that are appealing to and likely promote youth use, including creativity and innovation, individuality or freedom, arts and music, nature, technology, and alcohol imagery. The FDA should (1) prohibit companies from comarketing unauthorized products alongside authorized products, and (2) exercise enforcement against even authorized products that are marketed using youth-appealing features.

Yanling Lv, Kexian Ai, Feng Guo

Aiming at the problem that the coherent signal cannot be estimated when the traditional UCA-RB-MUSIC algorithm is used to detect the partial discharge of the transformer, an improved UCA-RB-MUSIC algorithm is suggested to reconstruct the signals captured by the array so that the number of calls sources is equivalent to the rank of the reconstructed signal matrix. The impact of correlated signals is eradicated. The results of simulation analysis and experimental verification show that when the improved UCA-RB-MUSIC algorithm is used to locate a single and double partial discharge source, the improved positioning error is only half of the mistake before improvement, which enhances the accuracy and resolution of uniform circular array positioning and the accuracy of transformer partial discharge detection.

Zofia Helman

Aleksander Tansman (1897–1986), kompozytor pochodzący z rodziny żydowskiej, urodzony i wykształcony w Polsce, wyjechał za granicę pod koniec 1919 po uzyskaniu nagrody i dwóch wyróżnień na pierwszym konkursie kompozytorskim w powojennej Polsce. Zamieszkał w Paryżu, gdzie szybko wszedł w tamtejsze środowisko muzyczne i stopniowo zyskiwał popularność. W latach 20. i 30. odbył dwukrotnie tournée koncertowe po Stanach Zjednoczonych a w 1932 tournée prowadzące przez 4 kontynenty, w czasie którego występował jako kompozytor, dyrygent i pianista. W przeciwieństwie do tych sukcesów w kraju rodzinnym jego utwory były rzadko wykonywane, a polscy krytycy o nastawieniu konserwatywnym nader nieprzychylnie komentowali jego twórczość. Dwukrotnie odwiedził Polskę: w 1932 i 1936. Zmiany polityczne w Polsce w latach 30. i wzrost nastrojów antysemickich spowodowały jednak, że w 1938 Tansman przyjął obywatelstwo francuskie. W czasie II wojny światowej i w pierwszym dziesięcioleciu powojennym jego kontakty ze środowiskiem polskim zostały przerwane. Władze PRL traktowały emigrantów jako osoby wrogie ustrojowi i nie należące do kultury polskiej. Toteż dopiero w 1958 nawiązali z Tansmanem kontakt dyrygent Stanisław Wisłocki (1921–1998) i muzykolog Krzysztof Biegański (1936–1967), który opublikował o nim kilka artykułów i w dużej mierze przyczynił się do pogłębienia znajomości jego dzieł w Polsce i uznania jego znaczenia w rozwoju nowej muzyki. Korespondencja Tansmana z Biegańskim z lat 1959–1961 stanowi drugą część artykułu.

Shulei Ji, Xinyu Yang

Generating music with emotion is an important task in automatic music generation, in which emotion is evoked through a variety of musical elements (such as pitch and duration) that change over time and collaborate with each other. However, prior research on deep learning-based emotional music generation has rarely explored the contribution of different musical elements to emotions, let alone the deliberate manipulation of these elements to alter the emotion of music, which is not conducive to fine-grained element-level control over emotions. To address this gap, we present a novel approach employing musical element-based regularization in the latent space to disentangle distinct elements, investigate their roles in distinguishing emotions, and further manipulate elements to alter musical emotions. Specifically, we propose a novel VQ-VAE-based model named MusER. MusER incorporates a regularization loss to enforce the correspondence between the musical element sequences and the specific dimensions of latent variable sequences, providing a new solution for disentangling discrete sequences. Taking advantage of the disentangled latent vectors, a two-level decoding strategy that includes multiple decoders attending to latent vectors with different semantics is devised to better predict the elements. By visualizing latent space, we conclude that MusER yields a disentangled and interpretable latent space and gain insights into the contribution of distinct elements to the emotional dimensions (i.e., arousal and valence). Experimental results demonstrate that MusER outperforms the state-of-the-art models for generating emotional music in both objective and subjective evaluation. Besides, we rearrange music through element transfer and attempt to alter the emotion of music by transferring emotion-distinguishable elements.

Zihao Deng, Yinghao Ma, Yudong Liu et al.

Large Language Models (LLMs) have shown immense potential in multimodal applications, yet the convergence of textual and musical domains remains not well-explored. To address this gap, we present MusiLingo, a novel system for music caption generation and music-related query responses. MusiLingo employs a single projection layer to align music representations from the pre-trained frozen music audio model MERT with a frozen LLM, bridging the gap between music audio and textual contexts. We train it on an extensive music caption dataset and fine-tune it with instructional data. Due to the scarcity of high-quality music Q&A datasets, we created the MusicInstruct (MI) dataset from captions in the MusicCaps datasets, tailored for open-ended music inquiries. Empirical evaluations demonstrate its competitive performance in generating music captions and composing music-related Q&A pairs. Our introduced dataset enables notable advancements beyond previous ones.

Jorge Calvo-Zaragoza, Alexander Pacha, Elona Shatri

The International Workshop on Reading Music Systems (WoRMS) is a workshop that tries to connect researchers who develop systems for reading music, such as in the field of Optical Music Recognition, with other researchers and practitioners that could benefit from such systems, like librarians or musicologists. The relevant topics of interest for the workshop include, but are not limited to: Music reading systems; Optical music recognition; Datasets and performance evaluation; Image processing on music scores; Writer identification; Authoring, editing, storing and presentation systems for music scores; Multi-modal systems; Novel input-methods for music to produce written music; Web-based Music Information Retrieval services; Applications and projects; Use-cases related to written music. These are the proceedings of the 5th International Workshop on Reading Music Systems, held in Milan, Italy on Nov. 4th 2023.

Jacopo de Berardinis, Albert Meroño-Peñuela, Andrea Poltronieri et al.

The annotation of music content is a complex process to represent due to its inherent multifaceted, subjectivity, and interdisciplinary nature. Numerous systems and conventions for annotating music have been developed as independent standards over the past decades. Little has been done to make them interoperable, which jeopardises cross-corpora studies as it requires users to familiarise with a multitude of conventions. Most of these systems lack the semantic expressiveness needed to represent the complexity of the musical language and cannot model multi-modal annotations originating from audio and symbolic sources. In this article, we introduce the Music Annotation Pattern, an Ontology Design Pattern (ODP) to homogenise different annotation systems and to represent several types of musical objects (e.g. chords, patterns, structures). This ODP preserves the semantics of the object's content at different levels and temporal granularity. Moreover, our ODP accounts for multi-modality upfront, to describe annotations derived from different sources, and it is the first to enable the integration of music datasets at a large scale.

Andrea Poltronieri, Aldo Gangemi

In this paper we propose the Music Note Ontology, an ontology for modelling music notes and their realisation. The ontology addresses the relation between a note represented in a symbolic representation system, and its realisation, i.e. a musical performance. This work therefore aims to solve the modelling and representation issues that arise when analysing the relationships between abstract symbolic features and the corresponding physical features of an audio signal. The ontology is composed of three different Ontology Design Patterns (ODP), which model the structure of the score (Score Part Pattern), the note in the symbolic notation (Music Note Pattern) and its realisation (Musical Object Pattern).

Lamberto Tronchin, Antonella Bevilacqua, Ruoran Yan

The Ponchielli theater of Cremona was built in 1808 after a fire destroyed the old wooden structure. The interior, the architecture and the shape of the plan layout are reminiscent of the Teatro alla Scala, Milan, a masterpiece by the architect Piermarini, albeit on a smaller scale. The four orders of balconies crowned by the top gallery are typical features of a 19th Century Italian Opera theater. Acoustic measurements have been undertaken across the stalls and in some selected boxes according to ISO 3382. The main acoustic parameters resulting from the measurements have been used for the acoustic calibration of a 3D model representing the Ponchielli theater. The calibration has been used to compare different scenarios involving the acoustic response of the main hall at 50% and 100% occupancy. The outcomes indicate that no significant change can be detected when the seats are provided with robust upholstery, which can be considered a positive result, especially for the actors who are not forced to change their effort between rehearsal and live performance. In order to contextualize the measured values in relation to the optimal ones, a comparison with other Italian Opera theaters provided with similar architectural characteristics has been carried out. Overall, the findings indicate that the acoustics of the Ponchielli theater are suitable for both music and speech in line with the other selected theaters, as these places were mainly created for multifunctional purposes in the 19th Century.

Kiran Kappeler, Noemi Festic, Michael Latzer et al.

Algorithmic selection is omnipresent in various domains of our online everyday lives: it ranks our search results, curates our social media news feeds, or recommends videos to watch and music to listen to. This widespread application of algorithmic selection on the internet can be associated with risks like feeling surveilled (S), feeling exposed to distorted information (D), or feeling like one is using the internet too excessively (O). One way in which internet users can cope with such algorithmic risks is by applying self-help strategies such as adjusting their privacy settings (Sstrat), double-checking information (Dstrat), or deliberately ignoring automated recommendations (Ostrat). This article determines the association of the theoretically derived factors risk awareness (1), personal risk affectedness (2), and algorithm skills (3) with these self-help strategies. The findings from structural equation modelling on survey data representative for the Swiss online population (N2018=1,202) show that personal affectedness by algorithmic risks, awareness of algorithmic risks and algorithm skills are associated with the use of self-help strategies. These results indicate that besides implementing statutory regulation, policy makers have the option to encourage internet users’ self-help by increasing their awareness of algorithmic risks, clarifying how such risks affect them personally, and promoting their algorithm skills.

Ermelinda Paz

Este artigo tem como objetivo atualizar a biografia de Edino Krieger – crítico, compositor e produtor musical – publicada em dois volumes em 2012, pelo selo SESC Nacional. Em 2014, o compositor Ronaldo Miranda, em sua resenha crítica sobre a citada obra para a Revista da Academia Brasileira de Música a considerou como: “Uma biografia singular para um compositor plural”. A partir de então, nos propusemos a alterar o seu status de biografia desatualizada para biografia atualizada, mas reconhecendo que o potencial do conjunto de realizações do compositor, aliado à grandeza e à projeção natural de seus feitos motivariam o surgimento de novos fatos, revisões e inserções. O artigo pretende atualizar a biografia, em especial, não só no que tange às novas obras, mas também levantar as respostas da comunidade acadêmica através de novas gravações, artigos, dissertações e, ainda, teses. No Vol. II da biografia, nas páginas 58 - 59 há uma menção ao Projeto Bem-me-quer Paquetá. Por fim, é intenção do artigo trazer à baila através da análise minuciosa desse projeto, o Edino Krieger, Educador Musical, cuja importância nos pareceu relegada pelo próprio compositor a um plano menor.

Abby Haynes, Abby Haynes, Anne Tiedemann et al.

IntroductionDance can positively impact older people’s health and wellbeing across cultures and socioeconomic groups, countering age-related physical, sensorimotor and cognitive decline.Background/objectivesThe RIPE (Really Is Possible for Everyone) Dance program aims to improve older people’s physical, mental, cognitive and social wellbeing by integrating engaging dance sequences with evidence-based fall prevention exercises. We sought to identify what mechanisms support observed long-term participation in this program, including by people living with challenging health conditions.MethodsFollowing a realist evaluation approach, we co-developed and tested program theories iteratively with participant interviewees (n = 20), dance teachers (n = 2) and via observation of a dance class. Initial data were dual-coded and emergent findings were interrogated by the research team. Findings were organised to express Program activities + Context + Mechanism = Process outcomes configurations.ResultsWe identified four program theories comprising 14 mechanisms which explained long-term attendance: 1. RIPE Dance benefits my body and mind (trust in the program, belief in health benefits), 2. RIPE Dance helps me feel good about myself (self-efficacy, pride in achievement, psychological safety, defying expectations, feeling valued), 3. RIPE Dance creates camaraderie (social connection, mutual support, rapport with the teacher), and 4. RIPE Dance is uplifting (raised spirits, fun, synchrony, musical reactivity).ConclusionThe RIPE Dance program provides effective and enjoyable ‘exercise in disguise’ for older people with diverse mobility profiles.Significance/implicationsThis research confirms that participation in dance can contribute significantly to healthy, happy ageing. Findings detail program activities that were most strongly associated with process outcomes, offering guidance for further program development, implementation and scaling up.

Hajime Fukui, Kumiko Toyoshima

Since the time of Darwin, theories have been proposed on the origin and functions of music; however, the subject remains enigmatic. The literature shows that music is closely related to important human behaviours and abilities, namely, cognition, emotion, reward and sociality (co-operation, entrainment, empathy and altruism). Notably, studies have deduced that these behaviours are closely related to testosterone (T) and oxytocin (OXT). The association of music with important human behaviours and neurochemicals is closely related to the understanding of reproductive and social behaviours being unclear. In this paper, we describe the endocrinological functions of human social and musical behaviour and demonstrate its relationship to T and OXT. We then hypothesised that the emergence of music is associated with behavioural adaptations and emerged as humans socialised to ensure survival. Moreover, the proximal factor in the emergence of music is behavioural control (social tolerance) through the regulation of T and OXT, and the ultimate factor is group survival through co-operation. The “survival value” of music has rarely been approached from the perspective of musical behavioural endocrinology. This paper provides a new perspective on the origin and functions of music.

Jorge Calvo-Zaragoza, Alexander Pacha, Elona Shatri

The International Workshop on Reading Music Systems (WoRMS) is a workshop that tries to connect researchers who develop systems for reading music, such as in the field of Optical Music Recognition, with other researchers and practitioners that could benefit from such systems, like librarians or musicologists. The relevant topics of interest for the workshop include, but are not limited to: Music reading systems; Optical music recognition; Datasets and performance evaluation; Image processing on music scores; Writer identification; Authoring, editing, storing and presentation systems for music scores; Multi-modal systems; Novel input-methods for music to produce written music; Web-based Music Information Retrieval services; Applications and projects; Use-cases related to written music. These are the proceedings of the 4th International Workshop on Reading Music Systems, held online on Nov. 18th 2022.