Objective Classification of Hindustani Ragas

Table of Contents

1. INTRODUCTION

1.1 Background and motivation

1.2 Unique characteristic of Indian music

1.3 Characteristic of Music

1.4 Characteristic of Ragas

1.5 Features for classification

1.6 Objective of the book

1.7 Outline of the book

2. LITERATURE SURVEY

2.1 Introduction

3. STUDY OF ARTIST BEHAVIOUR FROM RAGA PERFORMANCE

3.1 Introduction

3.2 Mel frequency cepstral co-efficients

3.3 Statistical parametrization

3.4 Analysis and Results

3.5 Music classification

3.6 confusion matrix

3.7 Summary

4. COMPARISON BETWEEN RESTFUL AND RESTLESS RAGAS USING BASIC STATISTICAL FEATURES

4.1 Introduction

4.2 Research methodology

4.3 Experimental results

4.4 pitch transition

4.5 Analysis of transitory and non-transitory frequency movements between Notes

4.6 Summary

5. COMPARISON BETWEEN RESTFUL AND RESTLESS RAGAS USING ADVACED STATISTICAL FEATURES

5.1 Introduction

5.2 Research Methodology

5.3 Experimental Result

5.3.1 Modeling the Pilu and Todi Raga performance statistically

5.3.2 Single Exponential Smoothing for raga Todi gives a bad fit

5.3.3 Single Exponential Smoothing for raga Pilu gives a comparatively better fit.

5.4 Results and Discussion

5.5 Summary

6. STATISTICAL MODELING OF RAGA TODI PERFORMANCE

6.1 Introduction

6.1.1 Indian Classical Music

6.1.2 Musical features of raga Todi

6.2 Double exponential smoothing

6.2.1 Model equation

6.2.2 Forecasts

6.3 Experimental Result

6.4. Discussion

6.5. Summary

7. STATISTICAL STUDY OF MUSIC EMOTION

7.1 Introduction

7.2 Research methodology

7.3 Statistical parameterization

7.4 Experimental Results

7.5 Summary

8. CONCLUDING REMARKS

Research Objectives and Topics

The primary research objective of this book is to identify and extract key features from Indian classical music signals, specifically Hindustani Ragas, to enable their objective classification based on mood, artist behavior, and performance characteristics. The work explores the use of signal processing and statistical modeling to distinguish between different musical aesthetics and emotions, providing a scientific foundation for understanding these complex structures.

• Feature extraction from music signals for classification purposes.
• Study of artist behavior and stylistic differences using MFCC and Andrew plots.
• Statistical comparison of restful versus restless Ragas.
• Advanced statistical modeling (simple and double exponential smoothing) to analyze and predict Raga performance.
• Representation and classification of music emotion (sad vs. happy).

Excerpt from the Book

Summary of Chapters

1. INTRODUCTION: Provides the background and motivation for objective Raga classification, defining basic musical components and the research objectives.

2. LITERATURE SURVEY: Summarizes previous developments in music signal feature extraction, classification techniques, and Raga identification research.

3. STUDY OF ARTIST BEHAVIOUR FROM RAGA PERFORMANCE: Investigates how different artists render Ragas and uses MFCC and Andrew plots to analyze stylistic similarities and differences.

4. COMPARISON BETWEEN RESTFUL AND RESTLESS RAGAS USING BASIC STATISTICAL FEATURES: Analyzes the rhythmic and note-duration characteristics that distinguish restful and restless Raga structures.

5. COMPARISON BETWEEN RESTFUL AND RESTLESS RAGAS USING ADVACED STATISTICAL FEATURES: Employs advanced statistical modeling and smoothing techniques to differentiate between restful and restless Raga performances.

6. STATISTICAL MODELING OF RAGA TODI PERFORMANCE: Focuses on applying double exponential smoothing to model the performance of the Todi Raga in detail.

7. STATISTICAL STUDY OF MUSIC EMOTION: Explores how statistical parameters derived from musical phrases correlate with evoke happy or sad emotions in the listener.

8. CONCLUDING REMARKS: Synthesizes the main research findings and discusses potential avenues for future investigation in computational musicology.

Keywords

Indian Classical Music, Hindustani Ragas, Music Classification, Feature Extraction, MFCC, Andrew Plot, Statistical Modeling, Simple Exponential Smoothing, Double Exponential Smoothing, Music Emotion, Pitch Transition, Note Duration, Inter-Onset Interval, Gharana, Computational Musicology

Frequently Asked Questions

What is the core focus of this research?

The research focuses on the objective identification and classification of Indian Classical music, specifically Hindustani Ragas, using signal processing and statistical methods.

Which music characteristics are analyzed in this book?

The book analyzes acoustic features like pitch, timbre, rhythm, melody, and temporal features, as well as higher-level features like Raga structure, mood (restful vs. restless), and artist-specific performance nuances.

What is the main goal of the proposed classification system?

The goal is to scientifically verify aesthetic claims about Ragas and enable automated classification, which helps in organizing large music databases and enhancing content-based music retrieval.

Which scientific methods are applied?

The authors employ methods such as Mel-Frequency Cepstral Coefficients (MFCC), statistical parameterization of note sequences, projection pursuit, Andrew plots, and both simple and double exponential smoothing for time-series modeling.

What does the book cover in its analytical sections?

It covers feature extraction, comparison between restful and restless Ragas, study of artist behavior across different Gharanas (schools of music), and the statistical modeling of Raga performances.

What are the characterizing keywords of this work?

Key terms include Indian Classical Music, Hindustani Ragas, Music Classification, Feature Extraction, Statistical Modeling, and Music Emotion Recognition.

How does the book address artist-specific behavior?

The book uses Andrew plots to visualize high-dimensional audio features, showing that Raga performances by artists from the same Gharana exhibit greater similarity in their curves compared to those from different Gharanas.

What conclusions were reached regarding Todi and Bhairavi Ragas?

The statistical analysis confirms that Todi is a restful Raga, whereas Bhairavi is categorized as restless, though the authors note that the association between "sad" music and "restful" nature is not always linear.