Music Information Research Articles

Analytical Approaches to Electroacoustic Music Improvisation

This article introduces a framework for musical analysis applied to electroacoustic improvisation. After an introduction about the benefits of interdisciplinary research and research-creation practice there is a description of the different domains of the framework: acoustic analysis, music analysis itself and the design of graphic representations. The framework is based on the extraction of information from the audio signal (music information retrieval) to create data representations, some of which are very efficient for music analysis, such as the brightness standard deviation or the self-similarity matrix. The last part of the article demonstrates the use of the framework on extracts of electroacoustic improvisations performed by the Paris-based group Les Phonogénistes, of which the author is a member.

Towards A Critical Music Information Literacy Praxis

:Critical music information literacy enables an individual to recognize and navigate the power structures that exist in our systems of music information construction, dissemination, preservation, and access. It relies on an instructor’s critical disposition toward information, which is passed on to the student who can then recognize the relationship between power and scholarship. Together they can seek to resist and tear down inequitable systems and practices. If we understand collections to be a locus of musical study and discourse, it is necessary for library workers, as well as those using music collections like researchers and performers, to engage in critical praxis and critical music information literacy if we hope to create an anti-racist, equitable society that “bend[s] toward justice.” Using this definition of critical music information literacy, we explore theories that inform its praxis within music library workership, briefly summarize challenges due to the materiality of music information, and offer some practical applications for reflection, adoption, and action.

Composition and Analysis as Communication. Outline for a Compositional Theory based on “Musical Information”

The concept of “information” in association with music recently re-emerged in music psychology and cognition in the form of statistical learning and as the basis for models of music perception. The first time music and information were associated in a systematic manner was in the 1950s and 60s when, particularly in North American music theory, researchers applied concepts from information theory to analysis and composition. Among the work of the 1950s was Edgar Coons’ and David Kraehenbuehl’s Information as a Measure of Structure in Music (1958). The authors suggested that information could be developed into a useful compositional tool that would aid decisionmaking in terms of form, but would not hamper the composer’s personal style or freedom of choice with regards to musical materials. In this article, the fundamental concepts behind the development of a musical information theory are outlined, giving some simple examples of its possibilities and proposing directions for future development. The outline embraces the necessity for a contemporary musical theory to combine formal, scientific methods and an awareness of contexts and relationships inspired by the New Musicology and the study of music as a cultural activity. As a theory of communication, information theory is presented as a far more subjective and pliable concept than is implied in Claude E. Shannon’s Mathematical Theory of Communication (1948), the work upon which many of the musical information theories in the 1950s were based. A flexible view of information theory and the concept of information must be maintained to enable them to take part in contemporary musicological discourse.

Research on Music Information Visualization of Chime Bells of Marquis Yi of Zeng

Chime bells of Marquis Yi of Zeng is an important ritual and musical instrument in ancient China, which embodies great cultural value and social influence. However, the particularity of the chime itself is not convenient to go out for display. The audience can only watch the chime through the protective glass, so they can't directly enjoy the unique music of the chime. This paper will intro-duce the music information of chime, think about the problems faced by chime music display at this stage, analyze the implementation method of sound visualization, explore the feasibility of chime mu-sic visualization, and explore the method of chime music visualization.

Development of Creative Skills of Future Teachers of Music Art by Means of Innovative Technologies

An urgent problem of modern Ukrainian education is not only the task of improving the quality of training of competitive specialists, but also the ability of teachers to apply innovative methodological support, effective principles, methods and forms of work in the educational process, to use modern music information technologies and computer programs in musical and pedagogical activities to create, converting and performing musical works. Modern education needs the introduction of innovative approaches, methods, forms and means of learning using information and music computer technologies into the educational process. The analysis of the main studies and publications on this issue confirmed the positive influence of innovative technologies on student motivation and the quality of education. The use of innovative teaching technologies in the educational process contributes not only to the improvement of the quality of training of future music teachers, but also to their introduction to compositional creativity and performing arts. Music information technologies, modern multimedia tools and music production software motivate students’ learning, activate their creative activity, develop musical memory, creative thinking, promote creative composer’s imagination, give students the opportunity to create their own music directly on a mobile phone at any time.

Musical Instrument Recognition with a Convolutional Neural Network and Staged Training

Musical instrument recognition is an important task within the broad field of Music Information Retrieval. It helps to build recommendation systems, compute similarity between musical compositions and enable automatic search in music collections with regard to the instrument. The task has two variants differed by the difficulty level. The simpler one is classification based on the sound of a single instrument, while the more difficult challenge is to recognize the predominant instrument in polyphonic recordings. In this paper, we used a convolutional neural network to solve both of these problems. As the analysis of monotimbral recordings is relatively easy, we used the knowledge acquired during solving it to train a neural network to tackle the more complex predominant instrument recognition problem. Within this staged training scenario, we also examined the impact of introducing some intermediate stages during the training sessions. The results showed that such a training approach has a potential to improve the accuracy of classification.

FNIRS-Based Dynamic Functional Connectivity Reveals the Innate Musical Sensing Brain Networks in Preterm Infants.

Humans have the ability to appreciate and create music. However, why and how humans have this distinctive ability to perceive music remains unclear. Additionally, the investigation of the innate perceiving skill in humans is compounded by the fact that we have been actively and passively exposed to auditory stimuli or have systematically learnt music after birth. Therefore, to explore the innate musical perceiving ability, infants with preterm birth may be the most suitable population. In this study, the auditory brain networks were explored using dynamic functional connectivity-based reliable component analysis (RCA) in preterm infants during music listening. The brain activation was captured by portable functional near-infrared spectroscopy (fNIRS) to simulate a natural environment for preterm infants. The components with the maximum inter-subject correlation were extracted. The generated spatial filters identified the shared spatial structural features of functional brain connectivity across subjects during listening to the common music, exhibiting a functional synchronization between the right temporal region and the frontal and motor cortex, and synchronization between the bilateral temporal regions. The specific pattern is responsible for the functions involving music comprehension, emotion generation, language processing, memory, and sensory. The fluctuation of the extracted components and the phase variation demonstrates the interactions between the extracted brain networks to encode musical information. These results are critically important for our understanding of the underlying mechanisms of the innate perceiving skills at early ages of human during naturalistic music listening.

Pedagogical Conditions for the Effective Realization of the Didactic Potential of Musical Information Technologies in the Training of Future Military Conductors

Pedagogical Conditions for the Effective Realization of the Didactic Potential of Musical Information Technologies in the Training of Future Military Conductors

Asymmetric Contrastive Learning for Audio Fingerprinting

Audio fingerprinting methods can compress audio contents into compact signatures so that we can save storage and reduce query time. This technology is widely used in many fields, such as audio retrieval, music information retrieval and audio authentication. However, most of the existing methods cannot balance the recognition accuracy, query speed and storage size well. This letter presents a novel self-supervised learning scheme called asymmetric contrastive learning to generate binary hash fingerprints of audio segments. Meanwhile, we design a new loss function named bidirectional asymmetric pairwise loss to minimize the loss of information. Experimental results show that our scheme can achieve a high top-1 hit rate on both music and speech datasets. Furthermore, the proposed scheme outperforms the previous work of real-value fingerprinting in query speed and storage size.

К ВОПРОСУ О ВЛИЯНИИ ТЕМБРА ПЕВЧЕСКОГО ГОЛОСА НА ВОСПРИЯТИЕ СОДЕРЖАНИЯ ВОКАЛЬНОГО ПРОИЗВЕДЕНИЯ (на примере исполнения романса С. В. Рахманинова «Здесь хорошо»)

The article presents the results of the research on the influence of the timbre of the singing voice on the listeners' perception and interpretation of musical content of a vocal work. The author considers timbre as one of the main instruments to transmit semantic and aesthetic information, fixed in detail in the musical text. The peculiarities of this process are revealed by means of interdisciplinary methodology combining the approaches of psychoacoustics, as well as the methods of performing and timbre analyses. The musical samples were the soprano singers' interpretations of S. V. Rachmaninov's romance "It's Good Here", irreproachable from the musical-aesthetic and vocal-technical points of view. Our research revealed that the acoustic characteristics of soprano perfectly complies with the sound images of the musical contents of the romance. Through different versions of the performers' interpretation of this vocal work, it is demonstrated that the semantic accents are mainly in accordance with the acoustic characteristics of a particular singer's voice. Among these are the primary zone of the singer's voice and the concentration of high and low singing formants, which determine the prevailing timbre colours. Besides, certain vocal techniques not only enrich the natural timbre with new overtones, but also result in a number of subtle metamorphoses in the listener's perception of the musical contents of the romance. The view on the musical contents of a vocal composition through the prism of the timbre component determines the novelty of the present article. The study of the special role of timbre in transmitting aesthetic and semantic musical information forms a promising direction course for further researches.

Musical Probability and the Music of Claude Debussy. Expectancy and Information in (…Bruyères) and the String Quartet

This article explores the relationship between musical expectancy and information through the use of both hidden Markov models, Bayesian probability and their application to Claude Debussy’s Prélude (…Bruyères) and his string quartet. (…Bruyères), which is one of the most conservative of the Préludes, is an interesting example of the composer’s use of “vagrant” sonorities which persist despite a perceived tonal hierarchy. An analysis of the varying levels of information, determined by the levels of expectancy which arise throughout the composition, illustrates an aspect of analysis which has not yet been applied to Debussy’s music. A study of musical information will inevitably lead to the study of probabilty, as a greater degree of information is expressed by events which could be considered to be less probable. Much of the research that has been done thus far in musical probability has used as a priori a collection of stylistic precepts, which are referred to as likely events. This article, however, attempts to determine probabilistic events based on “intra-opus” events by using David Temperley’s Bayesian approach to determine likely events within a composition. These are classified as events within a hidden Markov model, which examines the interrelationships of variables, some of which are perceived and some which are simply implied. A statistical “extra-opus” methodology is employed and compared to the Bayesian style. In the seond part of the article, melodic probability in a polyphonic context is discussed by examining the several sections from Debussy’s string quartet. The probabalistic fluctuations between voices in the quartet are then compared to the single melody of (…Bruyères).

The 14th International Conference of Students of Systematic Musicology (SysMus21)

The 14th International Conference of Students of Systematic Musicology (SysMus21) was held in a hybrid format that allowed both in-person and online participants to join. The SysMus conference series aims to foster a dynamic and interdisciplinary environment for students and early career researchers to share and discuss their work in the fields of systematic musicology and its related disciplines. This year at SysMus21, a total number of 26 oral and 35 poster presentations were held, covering a range of topics including well-being, data science, absorption and imagery, social connections, rhythm and groove, music information retrieval, sociology, cognition, and emotion. An introductory talk was given by Peter Vuust (Center for Music in the Brain, Aarhus University, Denmark), and two keynotes were presented by Jonna Vuoskoski (RITMO Center, University of Oslo, Norway) and Nori Jacoby (Max Planck Institute for Empirical Aesthetics, Frankfurt, Germany). Additionally, four workshops were held by Caitlyn Trevor, Joshua Bamford, Niels Christian Hansen, and Svenja Reiner, focusing on research skills relevant for developing a career in academia. In this report, an overview of the conference is provided including a summary of keynotes, presentations, workshops, and social activities, as well as a review of the advantages and challenges of the hybrid set-up.

Retracted] The New Approach Research on Singing Voice Detection Algorithm Based on Enhanced Reconstruction Residual Network

With the development of Internet technology, multimedia information resources are increasing rapidly. Faced with the massive resources in the multimedia music library, it is extremely difficult for people to find the target music that meets their needs. How to realize computer analysis and perceive users’ needs for music resources has become the goal of the future development of human‐computer interaction capabilities. Content‐based music information retrieval applications are mainly embodied in the automatic classification and recognition of music. Traditional feedforward neural networks are prone to lose local information when extracting singing voice features. For this reason, on the basis of fully considering the impact of information persistence in the network propagation process, this paper proposes an enhanced two‐stage super‐resolution reconstruction residual network which can effectively integrate the learned features of each layer while increasing the depth of the network. The first stage of reconstruction is to complete the hierarchical learning of singing voice features through dense residual units to improve the integration of information. The second stage of reconstruction is mainly to perform residual relearning on the high‐frequency information of the singing voice learned in the first stage to reduce the reconstruction error. In the middle of these two stages, the model introduces feature scaling and expansion convolution to achieve the dual purpose of reducing information redundancy and increasing the receptive field of the convolution kernel. A monophonic singing voice separation based on the high‐resolution neural network is proposed. Because the high‐resolution network has parallel subnetworks with different resolutions, it also has original resolution representations and multiple low‐resolution representations, avoiding information loss caused by serial network downsampling effects and repeating multiple feature fusions to generate new semantic representations, allowing for the learning of comprehensive, high‐precision, and highly abstract features. In this article, a high‐resolution neural network is utilized to model the time spectrogram in order to correctly estimate the real value of the anticipated time‐amplitude spectrograms. Experiments on the dataset MIR‐1K show that compared with the current leading SH‐4Stack model, the method in this paper has improved SDR, SIR, and SAR indicators for measuring the separation performance, confirming the effectiveness of the algorithm in this paper.

Automatic Music Classification Model Based on Instantaneous Frequency and CNNs in High Noise Environment.

Automatic music classification has significant research implications because it is the foundation for quick and efficient music resource retrieval and has a wide range of possible applications. In this study, DL is used to extract and categorize musical features, and a DL-based model for music feature extraction and classification is created. In this study, the instantaneous frequency and short-time Fourier transform are used to estimate the sine of a mixed music signal. Based on peak-frequency pairs, the DL algorithm is then used to calculate multiple candidate pitch estimates for each frame, and the melody pitch sequence is then obtained in accordance with the pitch profile duration and continuity characteristics. With this approach, the pitch can be calculated without reference to the fundamental frequency component. A music feature classification approach using spectrogram as input data and CNN as classifier is proposed at the same time in light of CNN's benefits in image processing. Studies reveal that this model's categorization and music feature extraction accuracy is as high as 94.18 percent and 95.66 percent, respectively. The outcomes demonstrate the efficiency of this technique for the extraction and classification of musical features. The field of music information retrieval is a good fit for it.

PipaSet and TEAS: A Multimodal Dataset and Annotation Platform for Automatic Music Transcription and Expressive Analysis Dedicated to Chinese Traditional Plucked String Instrument Pipa

Music information retrieval (MIR) is developing these years rapidly. As the fundamental MIR tasks, automatic music transcription (AMT) and expressive analysis (EA) are gaining momentum in both Western and non-European music. However, the annotated datasets for non-Eurogenic instruments remain scarce in terms of quantity and feature diversity so that general evaluations and data-driven models on various tasks cannot be well explored. As one of the most popular traditional plucked string instruments in Asia, which is barely studied in the MIR community, pipa has lots of distinctive national and local characteristics including the fake nails, intrinsic pitch shift, rubato, as well as a high diversity of sophisticated playing techniques that greatly enhance the music expressiveness. Our work aims to systematically clarify a creation procedure of a pipa dataset with audio, musical notation and multiview video modalities for traditional Chinese solos. The use of 4-track string vibration signals captured by optical sensors paves a path for high quality annotations. Furthermore, a transcription and Expressiveness Annotation System (TEAS) was transparently implemented to ensure the scalability of dataset. Three expressive analysis approaches in this system were newly proposed and evaluated in this paper. Finally, two AMT models were investigated and a series of the existing and emerging MIR tasks enabled by this dataset were enumerated for the future exploration.

The sound of music: from increased personalization to therapeutic values

Music providers like Spotify leverage music recommendation systems to connect users with relevant music. Based on content-based and collaborative-filtering statistical methods, these machine learning algorithms quantify user-song probabilities and present the highest-ranked songs. However, most music providers do not fully address their users’ music seeking and retrieval needs. Likewise, the fields of Recommender Systems (RecSys), Music Recommendation Systems (MRS) and Music Information Retrieval (MIR) remain disconnected from real-world use cases of music seeking. In this conceptual paper, we review the literature of the RecSys, MRS, MIR and Music Therapy (MT) academic fields. We discuss trends towards greater user control and personalization in the MRS and MIR fields and the connections between MT and positive health outcomes such as reductions in stress, anxiety and heart rate. We argue that greater control and visibility into the characteristics of songs and recommended items can generate positive downstream benefits. We recommend features that empower users to better seek, find, store, retrieve and learn from their musical catalogs.Results. We suggest design enhancements that recognize music’s wider psychological and physiological benefits and create opportunities to build domain knowledge. Unlocking music’s myriad benefits through the enhancements proposed would catalyze positive outcomes for business stakeholders, users and society.

Musical Information Visualization System

The visualisation systems are spread widely as personal computer’s software. The system, that are processing audio data are presented in this article. The system visualizes the ratio of spectrum amplitudes and has fixed frequency binding to colours. The technology of audio signals processing by the device and components of the device were considered. For the increasing information processing speed was used 32bit controller and graphic equalizer with seven passbands. Music visualization it is function, that are spread widely in mediaplayer’s software, on a different operation systems. This function shows animated images that are depends on music signal. Images are usually reproduced in the real time mode and synchronized with a played audio-track. Music and visualization are merges in the different kind of art: opera, ballett, music drama or movies. Dependencies of auditory and visual sensations are used for increasing the emotional perseption for ordinary listeners . In the systems, that are currently being actively promoted, are used several tools for personal computers, such as: After Effects – The Audio Spectrum Effect, VSDC Video Editor Free – Audio Spectrum Visualizer, Magic Music Visuals. The software, that are mentioned above, has a one disadvantage: the using of streaming video is not possible with the simultaneous receipt of audio and requires processing and rendering of the resulting video series. The purpose of the work is to determine the features of spectral analysis of music information and taking into account real-time data processing. Propose a variant of the music information visualization system, which displays the spectral composition of music and the amplitude of individual harmonics, and filling the LED-matrix with the appropriate color depending on the amplitude of the audio signal, with the possibility of wireless signal transmission from the music source to the visual effects device. The technology of frequency analysis of the spectrum with estimation of amplitude of spectrum’s components of the musical data, that is arriving on the device is chosen for this project. The method is based on the analysis of the spectrum in the selected frequency bands, which in turn simplifies the function of finding maxima at different frequencies. The proposed variant of the musical information visualization system provides display on the LED-matrix of colors that correspond to the frequencies spectrum’s components in the musical composition. Moreover, the number of involved LEDs is proportional to the ratio of the amplitudes of the signal’s frequency components. The desired result is achieved by using a Fast Fourier Transform and selecting Khan or Heming windows for providing a better analysis results of the signal spectrum. The amplitudes of the individual components of the spectrum are estimated additionally and each frequency band has its own color. The work of the system is to analyze the components of the spectrum and frequency of musical information. This information affects the display of colors on the LED matrix. The using of a 32-bit microcontroller provides sufficient speed of audio signal processing with minimal delays. For the increasing the accuracy and speed up the frequency analysis, the sound range is divided into seven bands. For this purpose was used seven-band graphic equalizer MSGEQ7. Music information is transmitted to the system via Bluetooth, which greatly simplifies the selection and connection of the music data source.

Shared and Unshared Feature Extraction in Major Depression During Music Listening Using Constrained Tensor Factorization.

Ongoing electroencephalography (EEG) signals are recorded as a mixture of stimulus-elicited EEG, spontaneous EEG and noises, which poses a huge challenge to current data analyzing techniques, especially when different groups of participants are expected to have common or highly correlated brain activities and some individual dynamics. In this study, we proposed a data-driven shared and unshared feature extraction framework based on nonnegative and coupled tensor factorization, which aims to conduct group-level analysis for the EEG signals from major depression disorder (MDD) patients and healthy controls (HC) when freely listening to music. Constrained tensor factorization not only preserves the multilinear structure of the data, but also considers the common and individual components between the data. The proposed framework, combined with music information retrieval, correlation analysis, and hierarchical clustering, facilitated the simultaneous extraction of shared and unshared spatio-temporal-spectral feature patterns between/in MDD and HC groups. Finally, we obtained two shared feature patterns between MDD and HC groups, and obtained totally three individual feature patterns from HC and MDD groups. The results showed that the MDD and HC groups triggered similar brain dynamics when listening to music, but at the same time, MDD patients also brought some changes in brain oscillatory network characteristics along with music perception. These changes may provide some basis for the clinical diagnosis and the treatment of MDD patients.

Organization and Representation of Musical Information (ORMI) in Portugal

This article presents the results of a documentary research regarding the current state of Organization and Representation of Musical Information (ORMI) in Portugal. Many authors describe the national scene of ORMI as very shortcoming, due to the usual difficulties: time vs. detail in the description and lack of knowledge of musical language by the technicians Information professionals.
 A survey of the monographs and articles as of 2011 that make a critical revision to the works of ORMI of several institutions was made, aiming to: 1) perceive their current state and 2) understand its strengths and weaknesses. We identified in these studies the analytic categories to create afterwards a classification structure by disciplinary areas, which shows semantic, structural and quality heterogeneity. The majority of Finding Aids are being produced by musicology projects with structures that respond to their information representation needs.

FAIR Interconnection and Enrichment of Public-Domain Music Resources on the Web

Vast amounts of publicly licensed classical music resources are housed within many different repositories on the Web encompassing richly diverse facets of information—including bibliographical and biographical data, digitized images of music notation, music score encodings, audiovisual performance recordings, derived feature data, scholarly commentaries, and listener reactions. While these varied perspectives ought to contribute to greater holistic understanding of the music objects under consideration, in practice, such repositories are typically minimally connected. The TROMPA project aims to improve this situation by interconnecting and enriching public-domain music repositories. This is achieved, on the one hand, by the application of automated, cutting-edge Music Information Retrieval techniques, and on the other, by the development of contribution mechanisms enabling users to integrate their expertise. Information within established repositories is interrelated with data generated by the project within a data infrastructure whose design is guided by the FAIR principles of data management and stewardship: making music information Findable, Accessible, Interoperable, and Reusable. We provide an overview of challenges of description, identification, representation, contribution, and reliability toward applying the FAIR principles to music information, and outline TROMPA's implementational approach to overcoming these challenges. This approach applies a graph-based data infrastructure to interrelate information hosted in different repositories on the Web within a unifying data model (a 'knowledge graph'). Connections are generated across different representations of music content beyond the catalogue level, for instance connecting note elements within score encodings to corresponding moments in performance time-lines. Contributions of user data are supported via privacy-first mechanisms that retain control of such data with the contributing user. Provenance information is captured throughout, supporting reproducibility and re-use of the data both within and outside the context of the project.