Hasil untuk "Musical instruction and study"

Yinghao Ma, Haiwen Xia, Hewei Gao et al.

While music generation models have evolved to handle complex multimodal inputs mixing text, lyrics, and reference audio, evaluation mechanisms have lagged behind. In this paper, we bridge this critical gap by establishing a comprehensive ecosystem for music reward modeling under Compositional Multimodal Instruction (CMI), where the generated music may be conditioned on text descriptions, lyrics, and audio prompts. We first introduce CMI-Pref-Pseudo, a large-scale preference dataset comprising 110k pseudo-labeled samples, and CMI-Pref, a high-quality, human-annotated corpus tailored for fine-grained alignment tasks. To unify the evaluation landscape, we propose CMI-RewardBench, a unified benchmark that evaluates music reward models on heterogeneous samples across musicality, text-music alignment, and compositional instruction alignment. Leveraging these resources, we develop CMI reward models (CMI-RMs), a parameter-efficient reward model family capable of processing heterogeneous inputs. We evaluate their correlation with human judgments scores on musicality and alignment on CMI-Pref along with previous datasets. Further experiments demonstrate that CMI-RM not only correlates strongly with human judgments, but also enables effective inference-time scaling via top-k filtering. The necessary training data, benchmarks, and reward models are publicly available.

Lancelot Blanchard, Cameron Holt, Joseph A. Paradiso

Vocals harmonizers are powerful tools to help solo vocalists enrich their melodies with harmonically supportive voices. These tools exist in various forms, from commercially available pedals and software to custom-built systems, each employing different methods to generate harmonies. Traditional harmonizers often require users to manually specify a key or tonal center, while others allow pitch selection via an external keyboard-both approaches demanding some degree of musical expertise. The AI Harmonizer introduces a novel approach by autonomously generating musically coherent four-part harmonies without requiring prior harmonic input from the user. By integrating state-of-the-art generative AI techniques for pitch detection and voice modeling with custom-trained symbolic music models, our system arranges any vocal melody into rich choral textures. In this paper, we present our methods, explore potential applications in performance and composition, and discuss future directions for real-time implementations. While our system currently operates offline, we believe it represents a significant step toward AI-assisted vocal performance and expressive musical augmentation. We release our implementation on GitHub.

Chunyu Qiang, Kang Yin, Xiaopeng Wang et al.

Text-to-speech (TTS) and text-to-music (TTM) models face significant limitations in instruction-based control. TTS systems usually depend on reference audio for timbre, offer only limited text-level attribute control, and rarely support dialogue generation. TTM systems are constrained by input conditioning requirements that depend on expert knowledge annotations. The high heterogeneity of these input control conditions makes them difficult to joint modeling with speech synthesis. Despite sharing common acoustic modeling characteristics, these two tasks have long been developed independently, leaving open the challenge of achieving unified modeling through natural language instructions. We introduce InstructAudio, a unified framework that enables instruction-based (natural language descriptions) control of acoustic attributes including timbre (gender, age), paralinguistic (emotion, style, accent), and musical (genre, instrument, rhythm, atmosphere). It supports expressive speech, music, and dialogue generation in English and Chinese. The model employs joint and single diffusion transformer layers with a standardized instruction-phoneme input format, trained on 50K hours of speech and 20K hours of music data, enabling multi-task learning and cross-modal alignment. Fig. 1 visualizes performance comparisons with mainstream TTS and TTM models, demonstrating that InstructAudio achieves optimal results on most metrics. To our best knowledge, InstructAudio represents the first instruction-controlled framework unifying speech and music generation. Audio samples are available at: https://qiangchunyu.github.io/InstructAudio/

Yinghao Ma, Siyou Li, Juntao Yu et al.

Recent advances in audio-text large language models (LLMs) have opened new possibilities for music understanding and generation. However, existing benchmarks are limited in scope, often relying on simplified tasks or multi-choice evaluations that fail to reflect the complexity of real-world music analysis. We reinterpret a broad range of traditional MIR annotations as instruction-following formats and introduce CMI-Bench, a comprehensive music instruction following benchmark designed to evaluate audio-text LLMs on a diverse set of music information retrieval (MIR) tasks. These include genre classification, emotion regression, emotion tagging, instrument classification, pitch estimation, key detection, lyrics transcription, melody extraction, vocal technique recognition, instrument performance technique detection, music tagging, music captioning, and (down)beat tracking: reflecting core challenges in MIR research. Unlike previous benchmarks, CMI-Bench adopts standardized evaluation metrics consistent with previous state-of-the-art MIR models, ensuring direct comparability with supervised approaches. We provide an evaluation toolkit supporting all open-source audio-textual LLMs, including LTU, Qwen-audio, SALMONN, MusiLingo, etc. Experiment results reveal significant performance gaps between LLMs and supervised models, along with their culture, chronological and gender bias, highlighting the potential and limitations of current models in addressing MIR tasks. CMI-Bench establishes a unified foundation for evaluating music instruction following, driving progress in music-aware LLMs.

Jiahao Zhao, Yunjia Li, Wei Li et al.

As large language models continue to develop, the feasibility and significance of text-based symbolic music tasks have become increasingly prominent. While symbolic music has been widely used in generation tasks, LLM capabilities in understanding and reasoning about symbolic music remain largely underexplored. To address this gap, we propose ABC-Eval, the first open-source benchmark dedicated to the understanding and instruction-following capabilities in text-based ABC notation scores. It comprises 1,086 test samples spanning 10 sub-tasks, covering scenarios from basic musical syntax comprehension to complex sequence-level reasoning. Such a diverse scope poses substantial challenges to models' ability to handle symbolic music tasks. We evaluated seven state-of-the-art LLMs on ABC-Eval, and the results reveal notable limitations in existing models' symbolic music processing capabilities. Furthermore, the consistent performance of individual baselines across different sub-tasks supports the reliability of our benchmark.

Paul Haimes

This paper introduces Chord Colourizer, a near real-time system that detects the musical key of an audio signal and visually represents it through a novel graphical user interface (GUI). The system assigns colours to musical notes based on Isaac Newton's original colour wheel, preserving historical links between pitch and hue, and also integrates an Arduino-controlled LED display using 3D-printed star-shaped diffusers to offer a physical ambient media representation. The method employs Constant-Q Transform (CQT) chroma features for chord estimation and visualization, followed by threshold-based filtering and tonal enhancement to isolate the root, third, and fifth. A confidence score is computed for each detection to ensure reliability, and only chords with moderate to very strong certainty are visualized. The graphical interface dynamically updates a colour-coded keyboard layout, while the LED display provides the same colour information via spatial feedback. This multi-modal system enhances user interaction with harmonic content, offering innovative possibilities for education and artistic performance. Limitations include slight latency and the inability to detect extended chords, which future development will aim to address through refined filtering, adaptive thresholds, and support for more complex harmonies such as sevenths and augmented chords. Future work will also explore integration with alternative visualization styles, and the comparison of audio analysis libraries to improve detection speed and precision. Plans also include formal user testing to evaluate perception, usability, and cross-cultural interpretations of colour-pitch mappings.

Yash Vishe, Eric Xue, Xunyi Jiang et al.

Music editing plays a vital role in modern music production, with applications in film, broadcasting, and game development. Recent advances in music generation models have enabled diverse editing tasks such as timbre transfer, instrument substitution, and genre transformation. However, many existing works overlook the evaluation of their ability to preserve musical facets that should remain unchanged during editing a property we define as Music Context Preservation (MCP). While some studies do consider MCP, they adopt inconsistent evaluation protocols and metrics, leading to unreliable and unfair comparisons. To address this gap, we introduce the first MCP evaluation benchmark, MuseCPBench, which covers four categories of musical facets and enables comprehensive comparisons across five representative music editing baselines. Through systematic analysis along musical facets, methods, and models, we identify consistent preservation gaps in current music editing methods and provide insightful explanations. We hope our findings offer practical guidance for developing more effective and reliable music editing strategies with strong MCP capability

Jin Gao

The objective of this research is to ascertain the efficacy of interactive instructional methodologies in the exploration of Chinese folk songs. The observation reveals that the main components of Chinese folk songs are singing in unison, music mode structure, variety of rhythmic patterns, contemplative moods, phonetic peculiarities of language, unity of music and textual elements. These indicators show that duplicate sound effects, the twelve-tone technique (scale), no semitones between sounds, the bright melodic patterns, the conciseness of poetry and the unity of phonetic and musical characteristics are the typical features of Chinese Taoist music. The interactive technologies were used to implement the training modules based on the online applications (Vocaberry, Vanido, Vocalizzo Lite, Forte and Ummo). The investigation into the efficacy of students’ knowledge establishes that before instruction, a heightened degree of practical aptitude was exhibited by 21% within the first group and 19% within the second group. After the training, the indicators were 72% and 68%, respectively. The research’s practical significance is the improvements in the educational process for teaching Chinese folk songs based on interactive technologies.

Jinwon Chung, Ji-hyang Oh, Eunshik Choi et al.

Objectives The purpose of this study is to propose a practical teaching and learning model that organically connects musical experience with Education for Sustainable Development(ESD), allowing learners to express their musical creativity and internalize social responsibility Methods This study aimed to validate the development of a “Music-Centered Sustainable Development Teaching- Learning Model” grounded in the philosophy of practical music education. It explores the importance of fostering values related to life, society, and the world through music education, alongside the cultivation of social responsibility through voluntary practices. To derive effective teaching methodologies, the study analyzed key issues in Education for Sustainable Development (ESD) and examined the applicability of transformative pedagogy as a core teaching method. Initially, a draft of the music-centered sustainable development teaching-learning process model, comprising four stages, was developed. This model was then refined and finalized through expert validation and consultation. Results As a result of the study, a “Music-Based Transformative Learning Model for Sustainable Development” was developed. The model aims to structurally expand understanding and practice by linking music learning with sustainable development goals. It is designed in four stages: “Introduction - Exploration - Application - Participation and Practice.” In each stage of learning, the principles of transformative pedagogy are applied, emphasizing experience, critical thinking and reflection, social interaction and creative problem-solving, and social participation and practice. Through expert validation and consultation, the characteristics of each stage of the model were clarified, and practical applications for music learning were proposed. Conclusions This study confirmed the potential of music education to serve as a practical platform for education for sustainable development. Based on musical experiences, it proposed a more concrete educational approach that can contribute to the holistic growth of learners and the formation of a sustainable future society. The practical model suggested in this study holds academic and educational significance in that it applies cross-curricular values to subject-specific education, and is expected to lay the foundation for the organic linkage between education for sustainable development and music education.

Eun-kyoung Son

Objectives This study explores the essence and causes of burnout experienced by mothers of students majoring in music at arts middle schools through a phenomenological approach. Methods TUsing Giorgi's phenomenological approach, the study explored various challenges faced by mothers in supporting their children's music education, including difficulties in time management, financial burdens, conflicts with their children, concerns about career paths, balancing academics and musical activities, and family disagreements. Results The findings revealed that mothers' experiences of burnout are not merely individual issues but are closely related to structural problems within the arts education environment. Mothers sacrificed their time and resources to support their children's musical achievements, experiencing psychological, emotional, and financial burdens in the process. Furthermore, this burnout was found to cause family conflicts and negatively affect both the children's musical performance and the overall quality of family life. Conclusions This study holds significance in its in-depth understanding of the burnout experiences of mothers with children majoring in music. It provides foundational data for developing parental support programs and improving policies within the music education environment.

Meeyeon Lee, Jae-Hyee Kim

Objectives The purposes of this study were to help those who learn and play the horn by examining the analysis and performance method of Mozart Horn Concerto K.495. Methods To this end, the compositional background of Mozart's Horn Concerto No. 4, K. 495, as well as the horn's range and basic playing techniques, were examined. The basic techniques were analyzed in terms of embouchure, stopped horn, mute, and trills, and based on this, the analysis and performance methods of Mozart's Horn Concerto No. 4, K. 495, were explored. Results The Mozart's Horn Concerto K.495 is a representative work that shows the musical characteristics of the classical period well, and although it is based on a traditional sonata form, it has Mozart's unique personality. When playing, you should take this into account, maintain an elegant and balanced style overall, and emphasize re strained beauty and clear structural expression. Conclusions Although this work is based on a traditional sonata format, it has Mozart's unique personality. Therefore, taking this into account when performing, you should maintain an elegant and balanced style as a whole, and focus on clear structural expression. And this concerto requires a high level of technique as well as vari ous performance techniques and musical expression from the horn performer. When this work is played at the same time as the piano accompaniment, the consistency of the overall rhythm should be maintained, and at the same time, it is important to express a bright and pleasant atmosphere, and to express Mozart's unique delicate emotions. Therefore, it was confirmed that this work embodies the beauty of classical music and is an important repertoire that requires the skill and artistic interpretation of the performer.

Meile

This paper explores the aesthetic education value of integrated teaching combining Mongolian long tunes and sanxian (three-stringed instrument). By analyzing the cultural essence and artistic characteristics of these two musical traditions, it examines how their collaborative instruction enhances aesthetic perception and stimulates artistic creativity. The study reveals that this pedagogical approach not only enriches music education content but also promotes the inheritance and development of ethnic musical culture, offering innovative approaches and practical models for music-based aesthetic education.

Yi Huang

This study examines how innovative percussion-teaching approaches—featuring polyrhythm structures and contemporary percussion practices—can be meaningfully integrated into primary and secondary school music education in China. The aim is to support students in developing musical creativity and expanding the expressive scope of their musical language. Guided by concepts of musical creativity, normative musical development, and participatory music practices, this research critically reviews existing scholarship and proposes a new framework for fostering creativity through percussion learning. By drawing together diverse theoretical perspectives and addressing challenges specific to the Chinese educational context, the study articulates an approach to rhythm instruction that balances technical skill with creative exploration. The framework emphasizes the importance of historical context, systematic rhythm training, and collaborative performance experiences in cultivating students’ creative capacities and fluency in musical expression. Ultimately, this study seeks to bridge traditional and innovative pedagogies, offering practical strategies for nurturing a new generation of creative young musicians who can communicate effectively through a rich and expressive musical language, contributing to the ongoing development of music education in China.

Jinyi Yu, Marissa Ejercito – Borines

Based on an interdisciplinary perspective, this study explores the mechanisms by which environmental support and cognitive regulation enhance musical creativity among non-major university students in Orff improvisation. A quasi-experimental design was employed, selecting 400 non-major university students as research participants. The experimental group received Orff improvisation instruction integrated with environmental support and cognitive regulation strategies, while the control group received traditional music instruction, with an intervention period of 16 weeks. The study utilized mixed methods, collecting data through questionnaire surveys, behavioral observations, work analysis, and in-depth interviews, and analyzing the data using descriptive statistics, analysis of variance, regression analysis, and structural equation modeling. The results indicate: (1) Environmental support and cognitive regulation demonstrate significant synergistic effects; when environmental support reaches optimal configuration, students' cognitive regulation efficacy improves by 67.3%, and overall musical creativity enhances by 78.2%; (2) The three stages of Orff improvisation present differentiated environment-cognition interaction patterns: the imitation stage primarily optimizes cognitive load, the exploration stage promotes metacognitive development, and the improvisation stage releases cognitive freedom; (3) Demographic characteristics play important moderating roles in environment-cognition interactions, with gender differences mainly manifested in environmental support preferences and age differences reflected in the maturity of cognitive regulation strategies; (4) Musical creativity enhancement is achieved through four pathways: cognitive load optimization, emotional arousal regulation, social belonging construction, and self-efficacy enhancement, demonstrating significant psychological and social effects; (5) The constructed interdisciplinary theoretical framework successfully explains the complex mechanisms of musical creativity development, with the overall model explanatory power reaching 73.4%. The proposed "ecological-cognitive" teaching model provides scientific guidance for music education practice, and the developed assessment tools for environmental support and cognitive regulation establish a foundation for subsequent research. This study enriches the theoretical system of music education, expands the application domains of environmental psychology and cognitive psychology, and provides important theoretical foundations and practical pathways for cultivating innovative talents and promoting students' comprehensive development.

Rifdah Fadhillah, Dewi Meyrasyawati, Johny Alfian Khusyairi

This study analyzes music's role in shaping visitors' hyperreal experiences in Kampung Jelita, Surabaya. Using a qualitative case study approach, the research focuses on the Japanese and Balinese thematic zones and applies Jean Baudrillard's theory of hyperreality. Data were collected through in-depth interviews, direct observation, literature review, and documentation conducted in the Kampung Jelita area, Manukan Lor IV E Street, RT 05/RW 01, Banjar Sugihan Subdistrict, Tandes District, Surabaya. The findings reveal that instrumental music is a cultural simulator that evokes illusion and fantasy, supporting Baudrillard's view that simulation can substitute reality. The soundscapes in each thematic zone, Japanese and Balinese, enhance the immersive experience by harmonizing with visual ornaments, prompting visitors to engage emotionally, experience nostalgia, and participate in performative acts such as renting traditional costumes. However, some visitors noted inconsistencies between the music and the intended cultural themes. Theoretically, this study contributes to hyperreality discourse by emphasizing the role of music in reinforcing sensory simulation within thematic tourism village contexts. Practically, the findings suggest that curating culturally coherent soundscapes can strengthen visitors' emotional attachment and enhance destination branding. Thus, music should be regarded as a strategic medium in constructing cultural identity and tourist experience. Hiperrealitas Musik di Kampung Jelita: Studi Kasus Tematik Wisata di Surabaya Abstrak Penelitian ini menganalisis peran musik dalam membentuk pengalaman hiperreal pengunjung di Kampung Jelita, Surabaya. Dengan menggunakan pendekatan studi kasus kualitatif, penelitian ini berfokus pada zona tematik Jepang dan Bali serta menerapkan teori hiperrealitas Jean Baudrillard. Data dikumpulkan melalui wawancara mendalam, observasi langsung, studi pustaka, dan dokumentasi yang dilakukan di kawasan Kampung Jelita, Jalan Manukan Lor IV E, RT 05/RW 01, Kelurahan Banjar Sugihan, Kecamatan Tandes, Surabaya. Hasil penelitian menunjukkan bahwa musik instrumental berfungsi sebagai simulator budaya yang membangkitkan ilusi dan fantasi, mendukung pandangan Baudrillard bahwa simulasi dapat menggantikan realitas. Lanskap bunyi (soundscape) di setiap zona tematik Jepang dan Bali memperkuat pengalaman imersif dengan menciptakan keselarasan antara elemen audio dan visual, sehingga mendorong pengunjung untuk terlibat secara emosional, merasakan nostalgia, serta berpartisipasi dalam tindakan performatif seperti menyewa kostum tradisional. Namun, beberapa pengunjung mencatat adanya ketidaksesuaian antara musik yang diputar dengan tema budaya yang dimaksudkan. Secara teoretis, penelitian ini memberikan kontribusi terhadap wacana hiperrealitas dengan menekankan peran musik dalam memperkuat simulasi sensorik dalam konteks kampung tematik wisata. Secara praktis, temuan ini menyarankan bahwa pengelolaan lanskap bunyi yang selaras secara budaya dapat memperkuat keterikatan emosional pengunjung dan meningkatkan citra destinasi wisata. Dengan demikian, musik perlu dipandang sebagai media strategis dalam membangun identitas budaya dan pengalaman wisata. Kata kunci: hiperrealitas musik; Kampung Jelita; musik instrumental tradisional; simulakra

Yixiao Zhang, Yukara Ikemiya, Woosung Choi et al.

Recent advances in text-to-music editing, which employ text queries to modify music (e.g.\ by changing its style or adjusting instrumental components), present unique challenges and opportunities for AI-assisted music creation. Previous approaches in this domain have been constrained by the necessity to train specific editing models from scratch, which is both resource-intensive and inefficient; other research uses large language models to predict edited music, resulting in imprecise audio reconstruction. To Combine the strengths and address these limitations, we introduce Instruct-MusicGen, a novel approach that finetunes a pretrained MusicGen model to efficiently follow editing instructions such as adding, removing, or separating stems. Our approach involves a modification of the original MusicGen architecture by incorporating a text fusion module and an audio fusion module, which allow the model to process instruction texts and audio inputs concurrently and yield the desired edited music. Remarkably, Instruct-MusicGen only introduces 8% new parameters to the original MusicGen model and only trains for 5K steps, yet it achieves superior performance across all tasks compared to existing baselines, and demonstrates performance comparable to the models trained for specific tasks. This advancement not only enhances the efficiency of text-to-music editing but also broadens the applicability of music language models in dynamic music production environments.

Hao Zhao, Maksym Andriushchenko, Francesco Croce et al.

In-context learning (ICL) allows LLMs to learn from examples without changing their weights: this is a particularly promising capability for long-context LLMs that can potentially learn from many examples. Recently, Lin et al. (2024) proposed URIAL, a method using only three in-context examples to align base LLMs, achieving non-trivial instruction following performance. In this work, we show that, while effective, ICL alignment with URIAL still underperforms compared to instruction fine-tuning on the established benchmark MT-Bench, especially with more capable base LLMs. We then uncover the most relevant elements for successful in-context alignment, finding the crucial role of the decoding parameters. Based on these insights, we show that the approach of URIAL can indeed be improved by adding high-quality, potentially carefully selected via greedy search, demonstrations in context, getting closer to the performance of instruct models. Finally, we provide the first, to our knowledge, systematic comparison of ICL and instruction fine-tuning (IFT) for instruction following in the low data regime, where ICL can be a viable alternative to IFT. Overall, our work advances the understanding of ICL as an alignment technique and its relationship to IFT. We provide our code at https://github.com/tml-epfl/icl-alignment.

Miguel Rios

Large Language Models (LLMs) have shown promising results on machine translation for high resource language pairs and domains. However, in specialised domains (e.g. medical) LLMs have shown lower performance compared to standard neural machine translation models. The consistency in the machine translation of terminology is crucial for users, researchers, and translators in specialised domains. In this study, we compare the performance between baseline LLMs and instruction-tuned LLMs in the medical domain. In addition, we introduce terminology from specialised medical dictionaries into the instruction formatted datasets for fine-tuning LLMs. The instruction-tuned LLMs significantly outperform the baseline models with automatic metrics.

Maria Rosa De Luca

The article addresses some of the main challenges related to the transmission of music-historical knowledge in today’s educational context, both at the school and academic levels. The focus is on the construction of historical knowledge, a crucial issue for the modes of historiographical narration. In this context, the analysis of teaching-learning procedures in music history extends to the music museum, understood as a space dedicated to promoting the learning of music history knowledge through innovative teaching approaches. The epistemological assumptions of didactic work are also discussed, taking into account the urgent problems that teachers face today in the process of didactic transposition, starting with the resistance to the growing phenomenon of presentism.

Zekun Li, Baolin Peng, Pengcheng He et al.

Large Language Models (LLMs) have demonstrated exceptional proficiency in instruction-following, becoming increasingly crucial across various applications. However, this capability brings with it the risk of prompt injection attacks, where attackers inject instructions into LLMs' input to elicit undesirable actions or content. Understanding the robustness of LLMs against such attacks is vital for their safe implementation. In this work, we establish a benchmark to evaluate the robustness of instruction-following LLMs against prompt injection attacks. Our objective is to determine the extent to which LLMs can be influenced by injected instructions and their ability to differentiate between these injected and original target instructions. Through extensive experiments with leading instruction-following LLMs, we uncover significant vulnerabilities in their robustness to such attacks. Our results indicate that some models are overly tuned to follow any embedded instructions in the prompt, overly focusing on the latter parts of the prompt without fully grasping the entire context. By contrast, models with a better grasp of the context and instruction-following capabilities will potentially be more susceptible to compromise by injected instructions. This underscores the need to shift the focus from merely enhancing LLMs' instruction-following capabilities to improving their overall comprehension of prompts and discernment of instructions that are appropriate to follow. We hope our in-depth analysis offers insights into the underlying causes of these vulnerabilities, aiding in the development of future solutions. Code and data are available at https://github.com/Leezekun/instruction-following-robustness-eval