Abstract

1. Introduction

2. Acoustic Feature Extraction

3. Sequence-Sequence Model-Based Music Melody Synthesis

4. RNN-Based Melody Synthesis

5. Experiments and Result Analysis

6. Conclusions

Data Availability

References

Abstract

     Computer music creation boasts broad application prospects. It generally relies on artificial intelligence (AI) and machine learning (ML) to generate the music score that matches the original mono-symbol score model or memorize/recognize the rhythms and beats of the music. However, there are very few music melody synthesis models based on artificial neural networks (ANNs). Some ANN-based models cannot adapt to the transposition invariance of original rhythm training set. To overcome the defect, this paper tries to develop an automatic synthesis technology of music teaching melodies based on recurrent neural network (RNN). Firstly, a strategy was proposed to extract the acoustic features from music melody. Next, the sequence-sequence model was adopted to synthetize general music melodies. After that, an RNN was established to synthetize music melody with singing melody, such as to find the suitable singing segments for the music melody in teaching scenario. The RNN can synthetize music melody with a short delay solely based on static acoustic features, eliminating the need for dynamic features. The proposed model was proved valid through experiments.

Introduction

     With the rapid development of modern computer science, many researchers have shifted their focus to computer-based algorithm composition or automatic music melody generation system. The research results on music melody synthesis and music modeling methods are being applied to various fields. The research of computer music creation aims to quantify and combine the emotional tendencies of music, with the aid of computer and mathematical algorithms. The specific tasks include aided composition, sound simulation and storage, and music analysis and creation [1, 2]. Computer music creation generally relies on artificial intelligence (AI) and machine learning (ML) to generate the music score that matches the original mono-symbol score model or memorize/recognize the rhythms and beats of the music. Despite its broad application prospects, the AI-based composition without needing lots of music knowledge rules is in the theoretical stage [3, 4].

     Speech processing has been widely applied in composition and songwriting, record production, and entertainment. Unlike simple speech synthesis, music melody synthesis has two additional processing steps: tone detection and transformation [5, 6]. Wenner et al. [7] preprocessed the musical melody synthesis corpus through automatic note segmentation and voiced/unvoiced sound recognition, constructed a high-quality music melody synthesis system, and proposed a music melody adjustment algorithm, which functions as an adaptive filter capable of detecting musical note cycles.

Conclusions

     Based on the RNN algorithm, this paper probes deep into the automatic synthesis of music teaching melodies. After extracting the acoustic features from music melodies, the authors established a sequence-sequence model for synthetizing general music melodies. To find the suitable signing segments for a given music melody in the teaching scenario, an RNN was set up to synthetize music melody with singing melody. After that, the convergence of different network models was compared through experiments, which verifies the feasibility of our model. In addition, the results of different models were compared before and after adding the singing melody, and the difference of the melody generated by our model and the original music melody was quantified accurately. Furthermore, the prediction error of phoneme time of each model configuration and that after applying the time constraint were obtained through experiments. The relevant results confirm the superiority of our model over DCNN and LSTM in modeling music melody sequence.

Abstract

Keywords

1. Introduction

2. Theoretical background

3. Research model and hypotheses

4. Research methodology

5. Results

6. Discussion

7. Limitations and further research

8. Conclusions and implications

Declarations

Appendix A. Supplementary data

References

Abstract

The music industry has undergone tremendous changes in relation to its production, distribution, and consumption habits due to the exponential development of new technologies, namely streaming platforms. The fact that sales of physical copies continue to decline significantly made it mandatory for this industry to reinvent itself by introducing music streaming services as a key part of its business development. This study aims to understand the factors that influence music consumption through streaming platforms, particularly studying the intention to adopt premium (paid) versions of a music streaming service and recommend them. An extension of the UTAUT2 model (version of the Unified Theory of Acceptance and Use of Technology, applied to the consumer side) was created. Based on data collected from 324 music streaming services users, the framework of this study was tested using structural equation modelling (SEM). Research also included in-depth semi-structured interviews in order to generate a more profound knowledge about the profile, behaviours and motivations of the new music consumer. Our findings confirm that habit, performance expectancy and price value play the most important role in influencing the intention to use a paid music streaming service. Simultaneously, new dimensions such as personalisation, attitude towards piracy and perceived freemium-premium fit arise as having an additional relevant role in adopting this type of service. The research contributes insights into music streaming services consumer behaviour, providing several theoretical and practical implications to music streaming services providers.

1. Introduction

Since the beginning of the oldest societies, music has played a fundamental role in the life of human beings, being undeniably a form of universal expression that unites old and future generations culturally and emotionally (Larsen et al., 2009, Larsen et al., 2010; Naveed et al., 2017). The importance of music in our society has led to creating an industry that includes all the concepts inherent to this thematic, such as its organisation, distribution, and profitability. This industry, made up mostly of countless record labels, has experienced golden times through sales of physical copies, thus monopolising the production and consumption of music. However, from 2001 onwards, it began to suffer the impact of the appearance of new technologies, thus initiating a digital age where the consumer has a greater capacity for decision (Arditi, 2014).

In the light of this event, the space for this industry as we knew it has become limited, and a reinvention of it was mandatory (Warr and Goode, 2011). The decrease in the volume of revenues, mainly due to the lower number of sales of physical copies (Sinclair and Tinson, 2017), led the main record labels to modernise. In particular, the growth of streaming services has revolutionised consuming music, as the number of users of these services keeps increasing (IFPI, 2021). It is known that since 2010, the number of users of the Spotify streaming platform has increased from 15 to 100 million worldwide (Aguiar and Waldfogel, 2018).

1- Introduction

2- Methodology, audience, and scope

3- Related ontologies and data models

4- Knowledge acquisition

5- Specification

6- Ontology description

7- Implementation and maintenance

8- Evaluation

9- Conclusions

References

Abstract

The Internet of Musical Things (IoMusT) is an emerging research area consisting of the extension of the Internet of Things paradigm to the music domain. Interoperability represents a central issue within this domain, where heterogeneous objects dedicated to the production and/or reception of musical content (Musical Things) are envisioned to communicate between each other. This paper proposes an ontology for the representation of the knowledge related to IoMusT ecosystems to facilitate interoperability between Musical Things. There was no previous comprehensive data model for the IoMusT domain, however the new ontology relates to existing ontologies, including the SOSA Ontology for the representation of sensors and actuators and the Music Ontology focusing on the production and consumption of music. This paper documents the design of the ontology and its evaluation with respect to specific requirements gathered from an extensive literature review, which was based on scenarios involving IoMusT stakeholders, such as performers and audience members. The IoMusT Ontology can be accessed at: https://w3id.org/iomust#.

Introduction

The Internet of Musical Things (IoMusT) is an emerging research area consisting of the extension of the Internet of Things paradigm to the musical domain. This field is positioned at the confluence of music technology, the Internet of Things, human-computer interaction, and artificial intelligence. IoMusT relates to the networks of computing devices embedded in physical objects (Musical Things) dedicated to the production and/or reception of musical content. Considering the computer science perspective, Turchet and colleagues defined a Musical Thing as “a computing device capable of sensing, acquiring, processing, or actuating, and exchanging data serving a musical purpose”. The IoMusT was then defined as “the ensemble of interfaces, protocols and representations of music-related information that enable services and applications serving a musical purpose based on interactions between humans and Musical Things or between Musical Things themselves, in physical and/or digital realms. Music-related information refers to data sensed and processed by a Musical Thing, and/or exchanged with a human or with another Musical Thing” [1].

2- Deep-learning models

3- Data sets and tasks

4- Practical guide

5- Applications

6- Limitations and future challenges

References

From feature engineering to end-to-end learning

Humans classify or annotate music based on diverse characteristics extracted from the audio signals. For example, a heavily distorted electric guitar sound with growling vocals is a good indication of metal music. Swing rhythms, syncopation, and chromatic comping by polyphonic instruments (e.g., piano or guitars) are obvious cues that the music is jazz. Translating these acoustic and musical features into numerical representations that computers can interpret is the essence of music classification and tagging. This usually involves a series of computation steps that convert audio content into a time–frequency representation, extract discriminative features, summarize them over time, and repeat the feature extraction and summarization until the proper category for the music can be determined. The way of improving each feature extraction step to achieve the best performance has evolved with advances in learning algorithms from hand engineering based on domain knowledge to end-to-end learning. Humphrey et al. [9] explained the transition in a unified deep architecture model where multiple blocks of affine transformation, nonlinear function, and optional pooling operation are pipelined. Figure 1 illustrates four different feature representation approaches in their framework. In reviewing the evolution of such approaches, we first separate them into two classes: feature engineering and feature learning.

Deciding which artist or song to sign and promote has always been a challenge for recording companies, especially when it comes to innovative newcomer singers without any chart history. However, the specifics of a creative industry such as the hedonic nature of music, socio-network effects, and ever fastening fashion cycles in combination with digitalization have made the recording industry even more competitive and these initial decisions even more crucial. With respect to the ongoing digital transformation and shift in power from organizations to consumers, we leverage digitally mediated wisdom of the crowd to build a forecasting model for better understanding chart success. Therefore, we draw on the literature of hedonic and experiential goods to investigate the relationship between crowd evaluations based on listening experience and popular music chart success. We track 150 song positions in reported music charts and also evaluate these songs via the crowd. Our model indicates that the wisdom of the crowd can improve forecasting chart success by almost 30% relatively to factors that have been earlier identified in the literature. However, this forecasting relevance is bound to certain conditions, namely the composition of the crowd, the underlying chart and market mechanisms, and the novelty of the musical material. In sum we find that crowd-based mechanisms are especially suited to forecast the performance of novel songs from unknown artists, which makes them a powerful yet very affordable decision support instrument for very uncertain contexts with limited historical data available. These findings can support recording companies to address the challenge of signing newcomers and thereby further enable the innovation system of the industry.

Introduction

It has always been crucial for music companies and labels to sign the right newcomer artists and select the best songs for a release in order to ensure a constant revenue stream from music product sales (Cameron, 2016; Dewan and Ramaprasad, 2014; Ordanini, 2006). These decisions are by no way perfect even today, which is reflected in revenues that are distributed according to the power law: revenues of few commercial successes compensate for the losses of many failures (Cameron, 2016; Strobl and Tucker, 2000). The reasons for this high number of failures and their challenges for firm survival can be found in the complex and uncertain market characteristics of the music publishing industry as well as in ever changing technologies. First, the consequences of the digital transformation, such as music piracy, lower per unit revenues of digital content, and new business models have diminished the total profitability of the industry (Aguiar and Waldfogel, 2015; Benner and Waldfogel, 2016; Dobusch and Schüßler, 2014; Hiller, 2016; Lam and Tan, 2001). Second, the selection criteria of the recording companies are based on informal heuristics since the hedonic nature of music prevents an objective measurement of its quality. Third, direct and large-scale consumer engagement has been expensive. In order to increase the chances of reaching the consumer, the products are therefore developed and bundled according to the expected consumer taste formulated by the media gatekeepers (Hirsch, 1972; Ordanini, 2006). Fourth, consumer taste is susceptible to fashion cycles, socio-network effects and decision difficulties due to increasing choice of consumption, which result in diversity and uncertainty about expected demand.

Different studies have proved the importance of language and vocabularies in structuring shared perceptions among stakeholders within an industry. Little is known about the discursive work involved in the evolution of these vocabularies. The use of corpus linguistics provides insights in the process through which an industry stabilises itself and how incumbents engage in discursive work to maintain their interests. We explore the jolt the Internet provoked in the French recorded music industry between 1988 and 2008. We argue that one major explanation for the industry inertia is that the technology of the Internet and the new distribution channels it opened up were framed by discourses, characterized by a relatively stable vocabulary which repeated established words such as right and artistic work. The discursive work of incumbent actors supports the stabilization of this vocabulary. When new words appeared, they were incorporated into existing vocabulary. This dual process of repetition and incorporation was facilitated by the fact that many of the words central to the debate were relatively empty and ambiguous floating signifiers. This paper represents a promising avenue to better account for the place of words in the institutional work and more specifically institutional maintenance.

Introduction

Institutions are usually defined as relatively taken-for-granted and enduring sets of practices (Greenwood, Oliver, Sahlin, & Suddaby, 2008). In recent years, researchers have begun to emphasize the fact that even deeply institutionalized practices and social relations require significant deliberative and conscious work to keep them running (Angus, 1993; Zilber, 2002, 2009). In that vein, there has been an emerging interest in the notion of institutional maintenance, referring to efforts ‘supporting, repairing and recreating’ (Lawrence & Suddaby, 2006 p230) existing institutions. This recognizes the agentic side of institutional stability. One way researchers have tried to get grips with institutional maintenance is by considering the role of language. Zilber (2009), for example, analysed how narratives which establish fixed meanings, are key to maintaining institutions. Others have looked further at the component parts of stories, such as the very words they are made up of are an important aspect of maintaining institutions. For instance institutional vocabularies (Suddaby & Greenwood, 2005) and key words (Jones & Livne-Tarandach, 2008) in articulating and manipulating institutions. While we know language more generally and the words we use in particular are an important part of stabilizing institutions, we do not really understand exactly how words evolve in an institutional field and what is the relationship between discourses at a micro level and diffused words constituting institutional vocabularies. Therefore, in this paper we explore the dynamics of words associated with institutional maintenance. This implies two interrelated levels of discourse analysis: a micro level associated with discursive work and a macro level which refers to an institutionalized vocabulary (that is, words systematically repeated among industrial actors).

When one thinks of popular music heritage, one thinks, perhaps, of celebrities being inducted into the Rock & Roll Hall of Fame, of The Cavern in Liverpool, museum exhibitions about popular music icons, such as David Bowie or Kylie Minogue, or even Joe Coreé’s burning of £5 m worth of punk heritage on the basis that “punk was never meant to be nostalgic” (BBC, 2016). Although there are many recent examples of this kind of music heritage, research into how and why music consumption communities remember their heritage is scarce. And yet it is important that scholars should focus attention on this. It has to do, after all, with the sustainability and continuity of a community over the passage of time. Whilst consumption communities of various forms have been of significant interest to consumer researchers and marketers (e.g. Cova and Cova, 2002; Schouten and McAlexander, 1995; Muňiz and O’Guinn, 2001), much of the literature has focused on defining the characteristics of the different types of communities and the dyadic relationship between the group and the product/brand. There has been less research on specific issues such as how communities are sustained through the construction of their history and heritage. In this paper, we aim to contribute insights into consumption communities’ past by answering the research question: RQ1. How do musical communities remember their past? Our inquiry sits within a tradition of research that seeks to answer the question: “how do societies remember?” (e.g. Connerton, 1991). We take a broadly social and cultural perspective to answering this question, whereas previous work on cultural memory has tended to focus at the level of the state, the nation, or the large organisation, in this case, we focus on a smaller group with a specific, musical, offer. The heritage literature, to the extent that it has focused on music, has focused more on music celebrities and the mainstream, whereas this paper focuses on an independent band. Finally, the collective consumption literature has largely ignored social and cultural memory. This paper seeks to begin to close all of these gaps by designing an inquiry which generates empirical data to develop an understanding of this nexus of issues. We draw upon insights from theories of memory in culture, heritage, and consumption communities to frame an exploration of the phenomenon. Our empirical inquiry focuses on how independent English rock band New Model Army (henceforth “NMA”) and its fans set up a free-of-charge exhibition of its art and artefacts in collaboration with the local museum service in the metropolitan county of West Yorkshire, England. The event was entitled One Family – One Tribe: The Art & Artefacts of New Model Army. This was a collection of the band’s art and artefacts that reflected 25 years of the band’s visual culture, including original cover artwork, lyrics notebooks, stage clothes, instruments, merchandise, photographs, audience souvenirs, and press cuttings. We make use of data from interviews, photographs, video, documents, and observations to document this case.

دوره ی قبل از مدرسه زمانی که مغز در سریع ترین حالت خودش پیشرفت میکند، یکی از بهترین زمینه ها برای توسعه ی ادراکی، زبانی، حرکتی، و احساسی کودک را فراهم میکند. ازین رو، کودکان مخصوصا در شش سال اولیه بسیار سریع رشد میکنند که به این دوره، دوره ی پیش از مدرسه گفته میشود که افراد در این دوره با سرعت بسیاری با توسعه و تحولات ادراکی و رشدی رو به رو میشوند. ازین رو، یک کودک در این دوره پتانسیل این را دارد که عضوی مولد و موثر در جامعه شود. تحصیلات مهم ترین روندی است که در شکل گیری، جهت دهی، تغییر و توسعه ی افراد و جوامع نقش دارد. تحصیلات در دوره ی فعلی در قالب های کاری شکل میگیرد که شامل سه سوژه ی اصلی با نام های علم، هنر و فن پیگیری میشود. مشابه علوم دیگر، موسیقی نیز یکی از شاخه های مهم هنر است که کیفیت های کلامی و شنوایی دارد. این هنر نقش بسیار مهمی در تحصیلات پیش از مدرسه دارد. در این زمینه، برنامه های پیش از مدرسه در سال 2013 در ترکیه آماده سازی شده و از نظر ارزیابی محتوایی که به صورت مستقیم در مورد تحصیلات موسیقی بوده است، مورد مطالعه قرار گرفته است.

1. مقدمه
تحصیلات را میتوان به عنوان روندی تعریف کرد که توسعه ی افراد از نظر ذهنی، احساسی، فیزیکی، و ابعاد اجتماعی، حمایت میکند و به آن ها کمک میکند تا با آگاهی رشد کرده و بتوانند به صورت هماهنگ با جامعه زندگی کنند. تحصیلات دو کارکرد مهم در زندگی فرد و جامعه دارد. به عنوان یک فرد، افرادی تربیت میشوند که نسبت به زندگی دیدگاه مثبتی داشته باشند و به عنوان جامعه، شهروند هایی تربیت میشوند که نسبت به ارزش های جامعه اهمیت میدهند و سعی میکنند در توسعه ی جامعه نقش داشته باشند. تحصیلات، با استفاده از برنامه های مشخص، به گونه ای برای کودکان طراحی شده است تا بتوانند فلسفه و طبیعت جامعه ی آزاد را درک کند. بر اساس گفته های آکای، که یک مدرس موسیقی است، موسیقی به افراد کمک میکند که به صورت کیفی رشد کنند که این موضوع به نفع جامعه است. این یکی از ویژگی های جدا نشدنی از هر نوع از تحصیلات است. اما، تحصیلات به عنوان یک کلمه، نشان دهنده ی یک کلیت میباشد. به صورت جزئی، توسعه ی ابعاد هنری، اجتماعی، فرهنگی و علمی در تحصیلات میتواند مد نظر باشد.