Musical Sources Research Articles

Music Source Separation PyTorch Checkpoints

Music Source Separation PyTorch Checkpoints

Music Source Separation PyTorch Checkpoints

Music Source Separation PyTorch Checkpoints

Music Source Separation PyTorch Checkpoints

Music Source Separation PyTorch Checkpoints

Sonic obstacles and conceptual nostalgia: Preliminary considerations on musical conceptualism and contemporary art

This paper is concerned with the aesthetic and discursive gap between music and contemporary art, and the recent attempts to remedy this in the field of New Music through a notion of “New Conceptualism.” It examines why, despite musical sources being central to the emergence of conceptual artistic strategies in the 1950s and “60s, the worlds of an increasingly transmedial “generic art” and music have remained largely distinct. While it takes New Music’s New Conceptualism as its focus, it argues that the perspective on New Music it takes has wider implications in music and art. It begins by defining what exactly “New Music” refers to, and outlines some of the conditions for the recent rise of conceptualism in New Music. It then takes the work of the composer Johannes Kreidler as a key example of some artistic tendencies and theoretical presuppositions in New Conceptualism. Following this it draws on work in the field of sound studies in order to critically examine the theoretical attempt to connect New Music with contemporary art that is found in the notion of “Music in the Expanded Field.” To conclude it offers some reflections on how a more robust conversation between contemporary art and New Music can begin to be conceived.

The LongSTEP approach: Theoretical framework and intervention protocol for using parent-driven infant-directed singing as resource-oriented music therapy

ABSTRACT Introduction: Despite medical advances, preterm birth and neonatal intensive care (NICU) hospitalization are demanding and pose risks for infants and parents. Various music therapy (MT) models have suggested parental singing to promote healthy bonding and development in premature infants, but evidence on long-term effects is lacking. Method: We present the theoretical framework and intervention protocol of a resource-oriented MT approach for premature infants and their caregivers used in the international LongSTEP trial (ClinicalTrials.gov NCT03564184). We illustrate how guiding principles manifest in MT sessions, describe frames for phases of intervention, discuss prerequisites and present hypothesized mechanisms of change. Results: The LongSTEP MT approach is resource-oriented, emphasizes parental voice and parent-infant mutual regulation, builds on family-centered care principles, and is relevant in the NICU and beyond. Essential elements include: observation and dialogue on infant and parent needs; voice as the main musical source, with parental voice as the most prominent; active parental participation; modification of music in response to infant states and cues; and integration of the family’s culture and music preferences. The music therapist facilitates and supports interaction between parents and infant. Parents learn how to adapt principles in relation to infant development across NICU hospitalization and post-discharge phases. Discussion: The LongSTEP approach is feasible in culturally diverse countries where consistent parental presence is available, but requires tailoring to local circumstances and culture, particularly in the post-discharge phase. The emphasis on parent-led infant-directed singing places a higher demand on parents than other MT approaches, and requires sufficient psychosocial and musical support for parents.

Elizabeth I, Thomas Tallis and Judith: music, resistance and the iconography of power

Abstract The use of biblical analogies hailing Elizabeth by comparing her to biblical figures has long been recognized as important for the establishment of her rule. Elizabeth was compared to the apocryphal heroine Judith throughout her reign in a positive way. Yet in the wake and aftermath of the Northern Rising, the Papal Bull and the Ridolfi plot the comparison was challenged, as were Elizabethan politics, by another Judith appearing on stage, Mary Stuart. By reading the famous motet, Spem in alium, composed by the Elizabethan court musician Thomas Tallis as a work that makes use of both Protestant and Catholic appropriations of the biblical heroine, the Essay shows the possibilities of musical communication covering a middle ground in this pivotal period of the reign. By commissioning the piece, one of the most eminent Catholic nobles, Henry Fitzalan, twelfth earl of Arundel, used a musically couched reading of the biblical story of Judith to negotiate his personal relationship with the queen after having been implicated in the Ridolfi plot. As such the motet allows new insights in the communicative interactions of Catholics and Protestants and shows the heuristic value of musical sources for understanding Elizabethan politics.

Research on DNN Methods in Music Source Separation Tools with emphasis to Spleeter

This paper tries to attempt a review on deep neural network (DNN) method in music source separation (MSS) tools with emphasis to Spleeter by Deezer, an enhanced deep learning model for music sourceseparation. It is a set of pre-trainedmodel written in python using the Tensorflow machine learning library used for musicsource separation. It was developed by Deezer, on the need to separate a given mixed music track to its constituentinstrumental or vocal tracks usually known as stems. Spleeter offers 3 pre-trainedmodels namely 2, 4, and 5 stemseparation models that are capable of separating a given mix into 2, 4, and 5 stems respectively, which can be used forvarious needs like remixing, up-mixing,music transcription, etc. This paper is the first of its kind to review on DNN methods in MSS.In this paper, we will learn about the purpose and useof Spleeter developed by Deezer as well as about the technical aspect behind this software product that includes areas like ArtificialIntelligence (AI), Machine Learning and Deep Learning, and further about Time-Frequency (TF) masking and U-NetConvolution Neural Network (CNN) which are the methodology and architecture employed in it respectively. From thereview, we learned that Spleeter by Deezer is one of the latest advancement in MSS problem that comparatively has one of the best signal to distortion ratio (SDR), signal to artifacts ratio (SAR), signal to interference ratio (SIR), and sourceimage to spatial distortion ratio (ISR) and produce a state of the art solution, and it has also paved a way togreater development in MSS problem in the future.

:Sources of Identity: Makers, Owners, and Users of Music Sources before 1600

:<i>Sources of Identity: Makers, Owners, and Users of Music Sources before 1600</i>

Efficient Robust Music Genre Classification with Depthwise Separable Convolutions and Source Separation

Given recent advances in deep music source separation, a feature representation method is proposed that combines source separation with a state-of-the-art representation learning technique that is suitably repurposed for computer audition (i.e. machine listening). A depthwise separable convolutional neural network is trained on a challenging electronic dance music (EDM) data set and its performance is compared to convolutional neural networks operating on both source separated and standard spectrograms. It is shown that source separation improves classification performance in a limited-data setting compared to the standard single spectrogram approach.

Source Separation and Depthwise Separable Convolutions for Computer Audition (Student Abstract)

Given recent advances in deep music source separation, we propose a feature representation method that combines source separation with a state-of-the-art representation learning technique that is suitably repurposed for computer audition (i.e. machine listening). We train a depthwise separable convolutional neural network on a challenging electronic dance music (EDM) data set and compare its performance to convolutional neural networks operating on both source separated and standard spectrograms. It is shown that source separation improves classification performance in a limited-data setting compared to the standard single spectrogram approach.

Reverberation times preferred by traditionally trained versus classically trained musicians for overall impression of contemporary gugak orchestras using auralisation techniques

Gugak encompasses the traditional music of Korea to contemporary music developed based on traditional music. Gugak orchestra music performed in a concert hall is a new type of music in the 20th century that combines traditional musical instruments of Korea and the form of a Western symphony orchestra. The optimal reverberation time for gugak is controversial. In the present study, the optimal reverberation time specifically for gugak instrumental solos, traditional gugak court ensemble, and orchestra music is investigated in an auralised concert hall measuring 18,838 m3 in volume.The preferred reverberation time (mean of 500 Hz and 1 kHz) is discovered to be 1.7 s among 1.1, 1.7, 2.1, and 2.6 s for both gugak and classical musician groups, regardless of the music sources. The optimum reverberation times derived are 1.78 s for gugak musicians and 1.95 s for classical musicians from the preferred reverberation times, consistent with the optimum reverberation times as a function of room volume in traditional concert hall studies. Subtle differences in the optimum reverberation times are observed between the two musician groups; however, the results support that the optimum reverberation times for gugak are affected by the room volume rather than the musical characteristics.

A Historical Perspective from Folklorist Henry Glassie: Roots of Folk Songs in Music Education

The purpose of this study was to examine the sociocultural roots of folk songs from the perspective of renowned folklorist Henry Glassie. Dr. Henry Glassie holds the rank of Professor Emeritus at Indiana University, where he previously served on the faculty for the Department of Folklore and Ethnomusicology. Primary sources included Glassie’s archived collection of folk song transcriptions, recordings, and field notes from the Appalachian region between 1961-1967, housed at the Archives of Traditional Music at Indiana University. A total of 1,665 titles from elementary general music sources were cross-referenced with Glassie’s collection, 26 of which appeared in the archived account. Two oral history interviews with Glassie revealed that folk songs that are commonly used in elementary general music classrooms have historical roots of which teachers may be unaware. The songs in this study raise ethical and moral questions regarding their use in the general music classroom and suggest that teachers carefully research folk songs and their related meanings.

‘Remplis l’air d’alegresse pour ce seigneur chery des cieux’: Music in the Entries of Nobility and Other Dignitaries in Late Renaissance France (c.1585–c.1615)

AbstractAlthough royal entries have long been studied by scholars, those undertaken by nobles and other political figures have generally attracted less attention. This is particularly true in musicological literature, a fact undoubtedly attributable to the paucity of surviving documentation on this topic. Yet music often played an important part in these spectacles, both by underlining the dignitary’s status and by enhancing key components of the symbolism deployed. Rather than concentrating on one particular event, this article brings together information about as many entries as possible made by such figures into French cities in the late sixteenth and early seventeenth centuries. This enables us to make general observations about the types of music used in these events (such as fanfares, Te Deums and dance music) in order to draw possible connections with extant musical sources, and also to ponder why certain musical elements mattered more than others to the contemporaries describing these occasions.

Proposal of Classified Music Recommendation Model Based on Social Media

With the spread of smartphones, it has become common to always listen to music like background music, so it is necessary to create a music database that meets individual needs. By creating a music database using data from social media, music is classified in a different way from collaborative filtering, which is mainly used by existing music source providing platforms. Various other hashtags attached to posts with music titles used as hashtags are collected using web crawling, and music is classified based on the collected hashtags to reflect actual listeners’ opinions on music. It uses crawling to find posts on social media with music titles tagged as hashtags, then collects other hashtags attached to those posts. Hashtags collected by performing the same task with multiple music titles are collected, analyzed, and statistics are then classified to determine when and where the music fits. On social media, the feelings of the person who posted the post and the conditions such as the time zone, place, weather, and situation where the post was posted are reflected as hashtags. By analyzing the hashtags attached to the music title, it is possible to build a social media-based music database in which the opinions of various people are reflected as collective intelligence. It is possible to derive different results from existing collaborative filtering based on the past listening records of users of the platform used by most of the sound source providing platforms. Even if music titles are not written in complete form on social media hashtags, we plan to research them so that they can be used to build a database.

Teaching About Police Brutality Through Music

I taught 2nd year high school students at a private girl’s school in Japan about police brutality and the African American experience through songs by four different African American artists: Michael Jackson, Beyoncé, James Brown and Alicia Keys. Other resources utilized in this elective class included the film “The Hate U Give” (2018), interviews and videos on YouTube and an original interview with an African American teacher working at the school. Utilizing music and many other primary sources, this course helped students improve their English, learn about American culture, and reflect on challenging current events. While the topic was linguistically difficult for English language learners, I introduced vocabulary from the songs and other sources, prepared pre-viewing guides, and regularly checked for comprehension through discussion questions and collecting student work. The students seemed to appreciate being able to discuss a difficult, relevant topic, and their vocabulary related to the subject matter steadily improved. Students also seemed to enjoy the materials used to teach the content matter.

Music Source Separation Using ASPP Based on Coupled U-Net Model

Noise has established itself as one of the factors that interfere with modern human life, and various noise canceling techniques have been studied to prevent noise. While the old era's noise-canceling technique focused on the physical soundproofing technique, multiple studies have been conducted on the active noise canceling technique that removes only the activated noise in the current era. Active noise canceling (ANC) or digital noise-canceling technology is based on the sound source separation method. This leads to sound source separation technology, which refers to the technology to separate individual sound signals from mixture sounds. Most of the source separation technologies focus on improving speech, not noise reduction. This technology makes it possible to obtain desired sound information more accurately and further improves noise-canceling technology by eliminating unwanted sound information. To provide deeper capability and more enhanced sound separation than the existing structure, we are focused on coupled U-Net model and Atrous spatial pyramid pooling technique (ASPP). This paper presents the music source separation method that combined Coupled U-Net structure with Atrous spatial pyramid pooling technique. To prove the proposed source separation method, we compared GNSDR, GSIR, and GSAR using MIR-1K, a data set that can evaluate the performance of the music source separation. Performance results show that the proposed source separation method overcame other methods' disadvantages and strengthened the feature map.

Parallel multichannel blind source separation using a spatial covariance model and nonnegative matrix factorization

In this paper, we present a multichannel nonnegative matrix factorization (MNMF) system for the task of source separation. We propose a novel signal model using spatial covariance matrices (SCM) where the mixing filter encodes the spatial information and the source variances are modeled using a NMF structure. Moreover, the proposed model is initialized with the estimated source direction of arrival (DoA) in order to mitigate the strong sensitivity to parameter initialization. The proposed system has been evaluated for the task of music source separation using a multichannel classical chamber music dataset showing that it is possible to reach real time in the tested scenarios by combining multi-core architectures with parallel and high-performance techniques.

Design of online music teaching platform based on cognitive wireless sensor network and interactive design

With the development of science and technology in recent years, wireless sensor networks also have many new development directions. The online music learning platform can not only realize various interactions, but also meet the various needs and behaviors of users. There are already many courses suitable for online learning, and students can interact with computers. In this article, we mainly compare the advantages and disadvantages of each music platform, and choose an open music source platform that can better meet people’s needs to build a corresponding environment. We mainly adopt the B/S architecture and the corresponding development framework for music system design. Our goal is to develop a platform that can teach online. By analyzing typical algorithms in cognitive sensor networks, relevant routing algorithms are used to balance the energy consumption of network nodes. At the same time, we optimized the algorithm, added an energy threshold, and changed the weight of the cluster head. It can not only ensure the reasonable distribution of channels, but also ensure the balance of energy consumption, making the network more stable. At the same time, we also integrate the energy level on the basis of the existing radio network algorithm, and propose a new routing algorithm.

Robotic Musicianship Based on Least Squares and Sequence Generative Adversarial Networks

Robotic musicianship research aims to the configuration of robots can analyze, reason, and generate music autonomous. The goal of this research is to achieve the inspiring and meaningful musical interactions between humans and artificially creative robots. This research presents a new model of automatic music generation which is based on least squares and generative adversarial networks (GANs). This research specifies classical piano as the music source and uses the sensors as the context-awareness technology to sense and receive the input audio in human-robot interaction. Therefore, this research applies sequence generation adversarial network (SeqGAN) techniques that are better able to address discrete issues in generating samples of classical piano melodies and datasets. Modifying the SeqGAN approach, this research presented the Least Squares SeqGAN (LS-SeqGAN) method to create melody units on different chords and generates a set of music pieces as testing dataset. In this research, we implement the original method and use the least squares method to stabilize the training of GANs. The performance evaluation shows that proposed LS-SeqGAN method can fulfill the need both of music quality and creativity. It offers a robust infrastructure for the human-robotic interaction that can be used to promote the related robotic applications.

Conditional LSTM-GAN for Melody Generation from Lyrics

Melody generation from lyrics has been a challenging research issue in the field of artificial intelligence and music, which enables us to learn and discover latent relationships between interesting lyrics and accompanying melodies. Unfortunately, the limited availability of a paired lyrics–melody dataset with alignment information has hindered the research progress. To address this problem, we create a large dataset consisting of 12,197 MIDI songs each with paired lyrics and melody alignment through leveraging different music sources where alignment relationship between syllables and music attributes is extracted. Most importantly, we propose a novel deep generative model, conditional Long Short-Term Memory (LSTM)–Generative Adversarial Network for melody generation from lyrics, which contains a deep LSTM generator and a deep LSTM discriminator both conditioned on lyrics. In particular, lyrics-conditioned melody and alignment relationship between syllables of given lyrics and notes of predicted melody are generated simultaneously. Extensive experimental results have proved the effectiveness of our proposed lyrics-to-melody generative model, where plausible and tuneful sequences can be inferred from lyrics.