Classification of Indian Classical Carnatic Music Based on Raga Using Deep Learning

John, Siji; Sinith, M. S.; Sudheesh, R. S.; Lalu, P P

doi:10.1109/raics51191.2020.9332482

Cited by 13 publications

(14 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results obtained for the model in terms of accuracy was of 75% and Fmeasure of 73 % In [11] developed model obtained results based on the pattern recognition among the obtained pitches an accuracy 94% and Error value of 6. In [12] developed a model that utilized various classifiers for raga recognition distinguished lowered more information and resulted in over-fitting issues. Thus, the results obtained in terms of accuracy was of 95 % and Error value of 5.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Automated Transcription for Raga Recognition and Classification in Indian Classical Music Using Machine Learning

Gowrishankar¹,

Nagappa²

2021

Atlantis Highlights in Computer Sciences

View full text Add to dashboard Cite

Raga recognition is only possible by trained musician to understand the notes based on the lead voice but a beginner is unable to decode the notes. This is significant for current scenarios in developing an automated note transcription system in Indian Classical Music (ICM). In the present research, various properties of raga and the machine learning techniques that are used for identifying the raga by a machine rather than a human or music expert are surveyed. The previously developed automatic raga recognition techniques using Carnatic and Hindustani Music, the main drawbacks and the improvements required are discussed. The present research work discusses about the future proposed models for automatic raga recognition using pitch detection algorithm, finding Tuning Offset, and Note Segmentation process. The proposed model will obtain better accuracy more than 96 % when compared to the existing CNN, GMM that obtained accuracy of 94 % and 95 %.

show abstract

Section: Discussionmentioning

confidence: 99%

“…John et al [12] performed classification of ICM based data on automatic Raga recognition using a Deep Learning model through audio signal processing. The model classified the music signal that managed the audio dataset for music therapy.…”

Section: Resolution Of Pitch-classesmentioning

confidence: 99%

Automated Transcription for Raga Recognition and Classification in Indian Classical Music Using Machine Learning

Gowrishankar¹,

Nagappa²

2021

Atlantis Highlights in Computer Sciences

View full text Add to dashboard Cite

show abstract

“…The Indian Music dataset consists of two datasets that are known as Carnatic Music Dataset (CMD) and Hindustani music dataset (HMD) [15]. The dataset comprises of full-length audio recordings having their respective raga labels and are useful for evaluating the approaches for automatic raga recognition in Indian art music.…”

Section: Data Collectionmentioning

confidence: 99%

“…Sinith [2] developed the Discrete Pitch Contour (DPC) based Hidden Markov Model (HMM) on a table derived using Fibonacci series failed to consider when the same notes in different ragas performed DPC that showed deviations and lowered the performance. Similarly, Siji John [15] utilized CNN for automatic raga classification that faced problem during raga recognition. However, the most challenging task was faced during pattern recognition using pitch contour based algorithm to obtain pitch sequences present in Parsel-mouth library.…”

Section: Comparative Analysismentioning

confidence: 99%

Automatic Music Mood Classification using Multi-class Support Vector Machine based on Hybrid Spectral Features

Krishnaiah¹,

Parameshachari²

2021

IJIES

View full text Add to dashboard Cite

Automatic Raga identification is an important step in computational musicology as it is helpful for indexing Indian music, classifying, recommending tunes, etc. However, automatic raga recognition is a difficult task as it requires guidance from the expertise training and each of the ragas is identified based on the characteristic phrase and pitch collection. However, the identification or extraction of appropriate musical sample features was difficult to enhance the success rate of the classifier. The feature identification and extraction were considered as the major issue in the existing models. To overcome such an issue, the proposed hybrid spectral feature extraction technique extracted and combined the spectral features of audios such as Spread, centroid, skewness, Mel Frequency Cepstrum Coefficients (MFCCs) and Linear Predictor Coefficients (LPCs) reduced the dimensionality complexity enhanced the success rate of the Multi-Support Vector Machine (MSVM) classifier for mood classification. The CompMusic dataset is utilized for this research work, where 4 types of ragas namely Sindhu Bhairavi, Darbari, Saveri, and Sri Raga, are considered for the mood classification of ragas. From these ragas, tone based features are extracted based on the amplitude of each raga samples and are fed to the MSVM classifier for mood classification of 4 different types are Sympathy, Serious, Peace, and Sad. The classification results for the proposed hybrid spectral feature extraction obtained better accuracy of 97.53% when compared to the existing Hidden Markov Model (HMM) obtained accuracy as 95.3%, Convolution Neural Network (CNN) model as 94%, and Gaussian Mixture Model (GMM) as 95 %.

show abstract

“…In modern days, deep learning is not only solving the problem of computer vision but also dealing with sequencing and time series problems [11][12]. Researchers have attained tremendous success in music genre classification by taking features like MFCCs, Spectrogram, and Scalograms from the audio songs and feeding them into neural networks [13]- [15]. Besides Convolutional Neural Network (CNN) [15], VGG-16 model [16], and ResNet-50, RNN-LSTM is also used to classify music genres by taking MFCCs features of audio files and performed slightly better when compared with CNN and DNN [9] [4].…”

Section: Introductionmentioning

confidence: 99%

Classification of Indian Classical Music (Hindustani Music) Genres through MFCCs Features using RNN-LSTM Model

Bisht

Negi

Singh

2022

Preprint

View full text Add to dashboard Cite

Music has been considered an inseparable part of our culture and tradition. In this work, we created a dataset with six Hindustani music genres: Abhang, Bhajan, Thumri, Tappa, Ghazal, and Kajri, each of which contains 100 songs in wave(.wav) format. To classify the Hindustani music genres, we employ the mel frequency ceptral coefficients features, which contain timbral information, and the Recurrent Neural Network-Long Short Term Memory. Our best three models achieved an average accuracy of 86% when trained on various feature sets with MFCC values of 18, 26, and 39. Furthermore, we use uniform manifold approximation and projection to transform and visualise higher-dimensional feature set data into two-dimensional space. Based on the results, we can infer that Hindustani music has more intricate melodies than western music, and feeding 18 MFCC features to the deep neural network is the optimum strategy to obtain better accuracy. Increasing the hop length from 512 to 1024 reduces the input dimension size, which facilitates the RNN-LSTM model. As a result, the performance of the RNN-LSTM models has been slightly improved. Our RNN-LSTM models’ test set accuracy decreased by 5% when we took 5 segments. Additionally, we evaluated and compared our model to six genres of the GTZAN dataset and achieved 90% accuracy.

show abstract

Classification of Indian Classical Carnatic Music Based on Raga Using Deep Learning

Cited by 13 publications

References 12 publications

Automated Transcription for Raga Recognition and Classification in Indian Classical Music Using Machine Learning

Automated Transcription for Raga Recognition and Classification in Indian Classical Music Using Machine Learning

Automatic Music Mood Classification using Multi-class Support Vector Machine based on Hybrid Spectral Features

Classification of Indian Classical Music (Hindustani Music) Genres through MFCCs Features using RNN-LSTM Model

Contact Info

Product

Resources

About