Virtual Computer Systems in AI-Powered Music Analysis: A Comparative Study for Genre Classification and Musicological Investigations

Journal of Information Systems Engineering & Management, Год журнала: 2023, Номер 8(4), С. 23395 - 23395

Опубликована: Дек. 13, 2023

The convergence of artificial intelligence (AI) and music analysis in recent years has altered how humans perceive analyze music. purpose this study was to investigate the effectiveness virtual computer systems for AI-powered analysis, as well they affect musicological insights genre classification. goal project uncover hidden patterns inside musical compositions while improving our understanding features underlying structures by fusing cutting-edge AI algorithms with possibilities virtualization technology. A quantitative design controlled experiments using standardized datasets used. Musical various styles were chosen, relevant such melody, rhythm, harmony retrieved. Metrics performance evaluation included categorization accuracy, precision, recall, F1-score, efficacy indicators investigations. findings shed light on innovative AI-driven analysis. Across a range genres, accurate classification achieved, demonstrating accuracy models identifying subtle traits. Deeper knowledge works aided discovery complex melodic motifs, chord progressions, rhythmic through research. By highlighting synergies between techniques systems, contributes expanding landscape It demonstrates AI's potential automating hard activities, complementing investigations, providing that supplement human expertise. demonstrated but it also highlighted its shortcomings due biases training data, model overfitting, resource restrictions systems. These limitations highlight necessity constant improvement awareness when incorporating into musicology.

Язык: Английский

---

A Comprehensive Survey of Music Genre Classification Using Audio Files

International Journal of Enhanced Research In Science Technology & Engineering, Год журнала: 2023, Номер 12(06), С. 183 - 192

Опубликована: Янв. 1, 2023

This survey extensively studies music genre classification, a critical task in information retrieval, to automatically categorize audio recordings into various genres. It provides comprehensive review of approaches, methodologies, and recent advancements classification from data. Scholars practitioners the field will find this study be valuable resource as it covers aspects discipline, including feature extraction, methods, dataset exploration, evaluation metrics, developments. The aims enhance understanding foster further research progress by critically evaluating state-of-the-art techniques discussed papers, discussing their strengths limitations, providing overview field.

Язык: Английский

---

Low Complexity Deep Learning Framework for Greek Orthodox Church Hymns Classification

Applied Sciences, Год журнала: 2023, Номер 13(15), С. 8638 - 8638

Опубликована: Июль 27, 2023

The Byzantine religious tradition includes Greek Orthodox Church hymns, which significantly differ from other cultures’ music. Since the deep learning revolution, audio and music signal processing are often approached as computer vision problems. This work trains scratch three different novel convolutional neural networks on a hymns dataset to perform classification for mobile applications. data first transformed into Mel-spectrograms then fed input model. To study in more detail our models’ performance, two state-of-the-art (SOTA) models were trained same dataset. Our approach outperforms SOTA both terms of accuracy their characteristics. Additional statistical analysis was conducted validate results obtained.

Язык: Английский

---

Effective acoustic parameters for automatic classification of performed and synthesized Guzheng music

EURASIP Journal on Audio Speech and Music Processing, Год журнала: 2023, Номер 2023(1)

Опубликована: Дек. 1, 2023

Abstract This study focuses on exploring the acoustic differences between synthesized Guzheng pieces and real performances, with aim of improving quality music. A dataset consideration generalizability multiple sources genres is constructed as basis analysis. Classification accuracy up to 93.30% a single feature put forward fact that although in subjective perception evaluation are recognized by human listeners, there very significant difference performed With features compensating each other, combination only three can achieve nearly perfect classification 99.73%, essential two related spectral flux an auxiliary MFCC. The conclusion this work points out potential future improvement direction algorithms properties.

Язык: Английский

---