Machine Learning Applied to Music/Audio Signal Processing

Over the past two decades, the utilization of machine learning in audio and music signal processing has dramatically increased [...]

Audio-Tactile Rendering: A Review on Technology and Methods to Convey Musical Information through the Sense of Touch

Tactile rendering has been implemented in digital musical instruments (DMIs) to offer the musician haptic feedback that enhances his/her music playing experience. Recently, this implementation has expanded to the development of sensory substitution systems known as haptic music players (HMPs) to give the opportunity of experiencing music through touch to the hearing impaired. These devices may also be conceived as vibrotactile music players to enrich music listening activities. In this review, technology and methods to render musical information by means of vibrotactile stimuli are systematically studied. The methodology used to find out relevant literature is first outlined, and a preliminary classification of musical haptics is proposed. A comparison between different technologies and methods for vibrotactile rendering is performed to later organize the information according to the type of HMP. Limitations and advantages are highlighted to find out opportunities for future research. Likewise, methods for music audio-tactile rendering (ATR) are analyzed and, finally, strategies to compose for the sense of touch are summarized. This review is intended for researchers in the fields of haptics, assistive technologies, music, psychology, and human–computer interaction as well as artists that may make use of it as a reference to develop upcoming research on HMPs and ATR.

Harmonic Classification with Enhancing Music Using Deep Learning Techniques

Automatic extraction of features from harmonic information of music audio is considered in this paper. Automatically obtaining of relevant information is necessary not just for analysis but also for the commercial issue such as music program of tutoring and generating of lead sheet. Two aspects of harmony are considered, chord and global key, facing the issue of the extraction problem by the algorithm of machine learning. Contribution here is to recognize chords in the music by the feature extraction method (voiced models) that performd better than manually one. The modelling carried out chord sequence, getting from frame-by-frame basis, which is known in recognition of the chord system. Technique of machine learning such the convolutional neural network (CNN) will systematically extract the chord sequence to achieve the superiority context model. Then, traditional classification is used to create the key classifier which is better than others or manually one. Datasets used to evaluate the proposed model show good achievement results compared with existing one.

Singing Voice Detection in Opera Recordings: A Case Study on Robustness and Generalization

Automatically detecting the presence of singing in music audio recordings is a central task within music information retrieval. While modern machine-learning systems produce high-quality results on this task, the reported experiments are usually limited to popular music and the trained systems often overfit to confounding factors. In this paper, we aim to gain a deeper understanding of such machine-learning methods and investigate their robustness in a challenging opera scenario. To this end, we compare two state-of-the-art methods for singing voice detection based on supervised learning: A traditional approach relying on hand-crafted features with a random forest classifier, as well as a deep-learning approach relying on convolutional neural networks. To evaluate these algorithms, we make use of a cross-version dataset comprising 16 recorded performances (versions) of Richard Wagner’s four-opera cycle Der Ring des Nibelungen. This scenario allows us to systematically investigate generalization to unseen versions, musical works, or both. In particular, we study the trained systems’ robustness depending on the acoustic and musical variety, as well as the overall size of the training dataset. Our experiments show that both systems can robustly detect singing voice in opera recordings even when trained on relatively small datasets with little variety.

MUSIC RADIO RECORDINGS AS OBJECTS OF THE ARCHIVAL STORAGE

The purpose of the article is to characterize the peculiarities of organizing the storage of musical audio recordings in the repositories of radio stations. The methodology is based on the use of general scientific and special methods. The universal nature of music as a special language determines the internationality of music art. From the moment of birth and during the past years verbal report of the unwritten conditions, that are connected to music sound record accumulation in different cultural institutions, has changed not once. The transformations of this relationship were caused by different reasons – social, political, technical, and technological nature. In current high-technology conditions integration properties inherent in the holistic process of creation, circulation, and spread of information, accumulated in music compositions for radio. Today the formation of the new qualitative communication area with the rapid growth of the sound messages streams level occurs. The article is dedicated to the scientific problem of the preservation and use of music audio recordings of radio companies as objects of archival storage. Archived musical radio records are defined as a special cluster of the communication area. The socio-communicative approach is the methodological basis of the publication. The scientific novelty. The main directions of ensuring the preservation, restoration, and restoration of music audio recordings of archival audio recordings are substantiated, as well as the peculiarities of digitization and use of sound documents of this type, are revealed. The specifics of the formation of the respective collections are considered on the example of the BBC Archive Center and The British Library Sound Archive as leading foreign institutions where music records are stored. It is concluded that digital technologies have led to a change in the culture of music consumption and, accordingly, have transformed the processes of storing music archival recordings in the repositories of radio companies, which have acquired specific properties. Examining music radio recordings as archival objects have shown that in the age of the digital revolution, the music industry around the world has undergone significant changes. Both the revenue structure and the cost structure of record labels and music radio companies have fundamentally changed. Conclusions. Digital technologies have led to a change in the culture of music consumption (there has been a change in the ideology of authorship for music products) and the emergence of digital recording technologies that use artificial intelligence, digital workstations, etc. In this regard, the specifics of the organization of storage of music archival recordings were transformed, in particular in the phono repositories of radio companies that have acquired specific properties of the service grade (reprint of archival music recordings that were previously specially recorded for radio stations).