Singer Identification Research Articles

Singer identification model using data augmentation and enhanced feature conversion with hybrid feature vector and machine learning

Analyzing songs is a problem that is being investigated to aid various operations on music access platforms. At the beginning of these problems is the identification of the person who sings the song. In this study, a singer identification application, which consists of Turkish singers and works for the Turkish language, is proposed in order to find a solution to this problem. Mel-spectrogram and octave-based spectral contrast values are extracted from the songs, and these values are combined into a hybrid feature vector. Thus, problem-specific situations such as determining the differences in the voices of the singers and reducing the effects of the year and album differences on the result are discussed. As a result of the tests and systematic evaluations, it has been shown that a certain level of success has been achieved in the determination of the singer who sings the song, and that the song is in a stable structure against the changes in the singing style and song structure. The results were analyzed in a database of 9 singers and 180 songs. An accuracy value of 89.4% was obtained using the reduction of the feature vector by PCA, the normalization of the data, and the Extra Trees classifier. Precision, recall and f-score values were 89.9%, 89.4% and 89.5%, respectively.

Literature Review on Singer Identification

Literature Review on Singer Identification

An ensemble model of CNN with Bi-LSTM for automatic singer identification

An ensemble model of CNN with Bi-LSTM for automatic singer identification

Singer and Genre recognition using deep learning

In this paper, we try to solve the two mostly occurred problems which are genre recognition and singer identification from common pieces of music. For that, we use a system that helps us to achieve our target that is based on a music information retrieval system, in which the system can process our files and the system can generate the desired results. The system is divided into different phases and for the music recognition, first, there is a need to separate the vocal features of the singer from the music clip because, with the background music, it’s slightly difficult to analyze the vocal of the singer, then extract music features, here we work on some of the music features like tempo, RMS, song duration, frequency, dynamic range, and tonality. The RNN (recurrent neural network) has been used for training the model and testing our data set. After the training of the models, the test classifier with different unknown audio files, and the results are 91% accurate for singer identification and 90% for music genre recognition. The system will identify the singer as well as the genre of the song from the same segment of the file. For the data set, the training data used 10 different Indian and Pakistani male and female singers’ audio files in Sindhi, Hindi, and Urdu language for six types of genres like rock, pop, jazz, disco, hip-hop, and classic. The songs were sung by only one singer, the group singing and live concerts music are limitations of our system.

Vocal92: Audio Dataset With a Cappella Solo Singing and Speech

Singer recognition plays a vital role in music information retrieval systems. Most songs in the singer recognition system are mixed audios of music and voice. In contrast, there is a lack of labeled a cappella solo singing data suitable for singer recognition. Text-independent singer recognition systems successfully encode audio features such as voice pitch, intensity, and timbre to achieve good performance. Most such systems are trained and evaluated using data from music with accompaniment. However, due to the influence of background music, the performance of the singer recognition model was limited. Contrarily, a powerful singer identification system can be trained and evaluated using a cappella solo singing voices with a clear and broad range of qualities. There needs to be labeled clear singing data suitable for singer recognition research. To address this issue, we present Vocal92, a multivariate a cappella solo singing and speech audio dataset spanning around 146.73 hours sourced from volunteers. Furthermore, we use two current state-of-the-art models to construct the singer recognition baseline system. In experiments, the singer recognition model developed by a cappella solo singing data performs well in both single-mode and cross-modal verification data, significantly improving related works. The dataset is accessible to everyone at https://pan.baidu.com/s/1Pn62DHfal2OOZ_5JqgGBdQ with jnz5 as the validation code. For non-commercial use, the dataset will be available free of charge at the IEEE DataPort (after acceptance).

Singer Identification by Vocal Parts Detection and Singer Classification Using LSTM Neural Networks

Abstract: Identification of singers is considered an important research area in audio signal processing. It has acquired the scientist’s intrigues in two primary branches,1) recognizing vocal parts of polyphonic music, and 2) Classifying Singer. Here, we plan to handle the two issues, simultaneously. Techniques like GMM, SVM, and HMM have previously been utilized in classifying singers. In this work, we proposed a system for singer identification using Deep Learning and Feed Forward Neural Networks, which to the best of our knowledge have not been used for this purpose before. Preprocessing involves examining large sets of audio features to extract the most efficient set for the recognition stage. Our work is divided into several stages. To begin with, the vocal parts of all music files are identified utilizing an LSTM network which can perform well for the time series information, for example, audio signals. Then, at that point, an MLP network is integrated and contrasted with an SVM classifier in order to classify the gender of the singers. At last, one more LSTM network is involved to recognize every singer ID, and contrasted with MLP network in a similar task. In each step, various classifiers are analyzed and the outcomes are looked at, which affirm the effectiveness of our technique contrasted with the best in class.

Self-Supervised Contrastive Learning for Singing Voices

This study introduces self-supervised contrastive learning to acquire feature representations of singing voices. To acquire robust representations in an unsupervised manner, regular self-supervised contrastive learning trains neural networks to make the feature representation of a sample close to those of its computationally transformed versions. Similarly, we employ two transformations—pitch shifting and time stretching—considering the nature of singing voices. Nevertheless, we use them reversely: we train networks to push away representations of the transformed versions. The networks then attempt to discriminate changes in vocal timbres introduced by pitch shifting without time stretching and those in singing expressions introduced by time stretching without pitch shifting. Consequently, the acquired representations become attentive to vocal timbre and singing expression. This was confirmed through a singer identification task, where we trained a classifier to learn the relationship between the feature representations to the corresponding singer labels of 500 singers. As a result, the employed transformations helped the classifier improve the classification accuracy by 9.12% (top-1 accuracy: 63.08%) compared with the case where the feature representations fed to the classifier were acquired without the transformations (top-1 accuracy: 53.96%). Furthermore, the proposed approach can be extended to acquire feature representations attentive to either vocal timbre or singing expression but not to the other by changing how the transformations are incorporated. We particularly explored the characteristics of such vocal timbre- or singing expression-oriented feature representations against song genre, singer gender, and vocal technique, and confirmed that they successfully capture different aspects of singing voices.

Positions, Channels, and Layers: Fully Generalized Non-Local Network for Singer Identification

Recently, a non-local (NL) operation has been designed as the central building block for deep-net models to capture long-range dependencies (Wang et al. 2018). Despite its excellent performance, it does not consider the interaction between positions across channels and layers, which is crucial in fine-grained classification tasks. To address the limitation, we target at singer identification (SID) task and present a fully generalized non-local (FGNL) module to help identify fine-grained vocals. Specifically, we first propose a FGNL operation, which extends the NL operation to explore the correlations between positions across channels and layers. Secondly, we further apply a depth-wise convolution with Gaussian kernel in the FGNL operation to smooth feature maps for better generalization. More, we modify the squeeze-and-excitation (SE) scheme into the FGNL module to adaptively emphasize correlated feature channels to help uncover relevant feature responses and eventually the target singer. Evaluating results on the benchmark artist20 dataset shows that the FGNL module significantly improves the accuracy of the deep-net models in SID. Codes are available at https://github.com/ian-k-1217/Fully-Generalized-Non-Local-Network.

Singer identification for Indian singers using convolutional neural networks

Singer identification is one of the important aspects of music information retrieval (MIR). In this work, traditional feature-based and trending convolutional neural network (CNN) based approaches are considered and compared for identifying singers. Two different datasets, namely artist20 and the Indian popular singers’ database with 20 singers are used in this work to evaluate proposed approaches. Cepstral features such as Mel-frequency cepstral coefficients (MFCCs) and linear prediction cepstral coefficients (LPCCs) are considered to represent timbre information. Shifted delta cepstral (SDC) features are also computed beside the cepstral coefficients to capture temporal information. In addition, chroma features are computed from 12 semitones of a musical octave, overall forming a 46-dimensional feature vector. Experiments are conducted with different feature combinations, and suitable features are selected using the genetic algorithm-based feature selection (GAFS) approach. Two different classification techniques, namely artificial neural networks (ANNs) and random forest (RF), are considered on the features mentioned above. Further, spectrograms and chromagrams of audio clips are directly fed to CNN for classification. The singer identification results obtained using CNNs seem to be better than the traditional isolated and ensemble classifiers. Average accuracy of around 75% is observed with CNN in the case of Indian popular singers database. Whereas, on artist20 dataset, the proposed configuration of feature-based approach and CNN could not give better than 60% accuracy.

Genetically related singers—acoustic feature analysis and impact on singer identification

Studies relating music with genetics have been one of the fascinating fields of research. In this study, we have attempted to answer the most curious question-how acoustically close are the genetically related singers? The present study has investigated this perception using two genetically different relations-three female sibling singers and father-son singer relation. These are famous Indian playback singers and the acoustic features are extracted using the songs of Bollywood films. Three different sets of self-developed cappella database are used for the experimentation. Positive correlations among the major musical aptitudes-pitch, vibrato, formant, and harmonic spectral envelope for both the singer relationships-revealed the genetic impact on the acoustic features. Also, the investigation of timbre spectral feature proved it a significant acoustic feature that differentiates similar voices. With Spearman's correlation coefficient, we conclude that strong acoustical association was observed between the acoustic features of genetically related singers, especially the female sibling singers. This was further validated by correlating these singers with genetically unrelated singers. A human perception test performed using cover songs indicated the genetic impact in voice similarity, while the automatic singer identification system discriminated singers more accurately than the human listeners.

Singer Gender Classification using Feature-based and Spectrograms with Deep Convolutional Neural Network

The task of music information retrieval (MIR) is gaining much importance since the digital cloud is growing sparklingly. An important attribute of MIR is the singer-id, which helps effectively during the recommendation process. It is highly difficult to identify a singer in the case of music as the number of signers available in the digital cloud is high. The process of identifying the gender of a singer may simplify the task of singer identification and also helps with the recommendation. Hence, an effort has been made to detect the gender information of a singer. Two different datasets have been considered. Of which, one is collected from Indian cine industries having 20 different singer details of four regional languages. The other dataset is standard Artist20. Various spectral, temporal, and pitch related features have been used to obtain better accuracy. The features considered for this task are Mel-frequency cepstral coefficients (MFCCs), pitch, velocity, and acceleration of MFCCs. The experimentation has been done on various combinations of the mentioned features with the support of artificial neural networks (ANNs) and random forest (RF). Further, the genetic algorithm-based feature selection (GAFS) has been used to select the suitable features out of the best combination obtained. Moreover, we have also utilized the recent popular convolutional neural networks (CNNs) with the support of spectrograms to obtain better accuracy over the traditional feature vector. Average accuracy of 91.70% is obtained for both the Indian and Western clips, which is an improved accuracy of 3% over hand engineering features.

Singer Identification using Autocorrelation Method

songs are the compositions embedding voice and different instrument’s sound. Different human emotions can be created by playing the appropriate song .autocorrelation algorithm is used here to find out singer identification. In the first experiment three singers with three hindi songs (vocal) are taken as data set. Tempo is used as musical features. Then autocorrelation is proposed on concerning a total of three singers. Using bartlett test we have found the most significant autocorrelation values of those songs of three singers. In second experiment three singers with one hindi song (vocal) are taken as data set. Here rms is used as musical features. Then autocorrelation is proposed on concerning those three singers. Using bartlett test we have found the insignificant autocorrelation values of the song of three singers. The first experiment is used to identify the singers for each song. Here three singers identify their own identification test giving most significant values of their songs .the second experiment gives the insignificant value. The insignificance values of musical features of three singers does not give the singer’s identification test.

Research on Singing Voice Detection Based on a Long-Term Recurrent Convolutional Network with Vocal Separation and Temporal Smoothing

Singing voice detection or vocal detection is a classification task that determines whether a given audio segment contains singing voices. This task plays a very important role in vocal-related music information retrieval tasks, such as singer identification. Although humans can easily distinguish between singing and nonsinging parts, it is still very difficult for machines to do so. Most existing methods focus on audio feature engineering with classifiers, which rely on the experience of the algorithm designer. In recent years, deep learning has been widely used in computer hearing. To extract essential features that reflect the audio content and characterize the vocal context in the time domain, this study adopted a long-term recurrent convolutional network (LRCN) to realize vocal detection. The convolutional layer in LRCN functions in feature extraction, and the long short-term memory (LSTM) layer can learn the time sequence relationship. The preprocessing of singing voices and accompaniment separation and the postprocessing of time-domain smoothing were combined to form a complete system. Experiments on five public datasets investigated the impacts of the different features for the fusion, frame size, and block size on LRCN temporal relationship learning, and the effects of preprocessing and postprocessing on performance, and the results confirm that the proposed singing voice detection algorithm reached the state-of-the-art level on public datasets.

Combined Evidence of MFCC and CRP Features Using Machine Learning Algorithms for Singer Identification

Singer identification is a challenging task in music information retrieval because of the combined instrumental music with the singing voice. The previous approaches focus on identification of singers based on individual features extracted from the music clips. The objective of this work is to combine Mel Frequency Cepstral Coefficients (MFCC) and Chroma DCT-reduced Pitch (CRP) features for singer identification system (SID) using machine learning techniques. The proposed system has mainly two phases. In the feature extraction phase, MFCC, [Formula: see text]MFCC, [Formula: see text]MFCC and CRP features are extracted from the music clips. In the identification phase, extracted features are trained with Bidirectional Long Short-Term Memory (BLSTM)-based Recurrent Neural Networks (RNN) and Convolution Neural Networks (CNN) and tested to identify different singer classes. The identification accuracy and Equal Error Rate (EER) are used as performance measures. Further, the experiments also demonstrate the effectiveness of score level fusion of MFCC and CRP feature in the singer identification system. Also, the experimental results are compared with the baseline system using support vector machines (SVM).

An Effective Framework for Speech and Music Segregation

Speech and music segregation from a single channel is a challenging task due to background interference and intermingled signals of voice and music channels. It is of immense importance due to its utility in wide range of applications such as music information retrieval, singer identification, lyrics recognition and alignment. This paper presents an effective method for speech and music segregation. Considering the repeating nature of music, we first detect the local repeating structures in the signal using a locally defined window for each segment. After detecting the repeating structure, we extract them and perform separation using a soft time-frequency mask. We apply an ideal binary mask to enhance the speech and music intelligibility. We evaluated the proposed method on the mixtures set at -5 dB, 0 dB, 5 dB from Multimedia Information Retrieval-1000 clips (MIR-1K) dataset. Experimental results demonstrate that the proposed method for speech and music segregation outperforms the existing state-of-the-art methods in terms of Global-Normalized-Signal-to-Distortion Ratio (GNSDR) values

A Longitudinal Ageing Analysis of Vocal Parameters of Singing Voice of Female Playback Singer

Age-related changes to the vocal structure affect the singing ability of the singer. We present a longitudinal study of vocal ageing of a female professional playback singer having more than six decades of singing span (covering singer age from 19 to 80 years). The ageing analysis is performed on six vocal parameters like – fundamental frequency (F0), vibrato, formants and spectral features like spectral roll-off and centroid. Statistical variations in these vocal parameters over the entire singing span of the singer are discussed in the paper. Significant effects noted with the ageing voice were - decrease in F0, decreased vocal range, reduction in vibrato rate, increase in vibrato extent, decrease in F2 & F4 formants and rapid change in the spectral features. This investigation also studied the effect of ageing on singing voice quality through the measurement of singing power ratio (SPR). Increase in SPR measures was observed with ageing voice. The study of impact of vocal ageing with longitudinal data on singer identification (SID) is scare. The SID experimentation performed with 350 cappella songs covering entire singing span of the singer, showed a clear impact that change in acoustical parameters with ageing affected the performance of singer identification systems.

A Novel Singing Voice Separation Method Based on a Learnable Decomposition Technique

In this paper, a new monaural singing voice separation algorithm is presented. This field of signal processing provides important information in many areas dealing with voice recognition, data retrieval, and singer identification. The proposed approach includes a sparse and low-rank decomposition model using spectrogram of the singing voice signals. The vocal and non-vocal parts of a singing voice signal are investigated as sparse and low-rank components, respectively. An alternating optimization algorithm is applied to decompose the singing voice frames using the sparse representation technique over the vocal and non-vocal dictionaries. Also, a novel voice activity detector is presented based upon the energy of the sparse coefficients to learn atoms related to the non-vocal data in the training step. In the test phase, the learned non-vocal atoms of the music instrumental part are updated according to the non-vocal components captured from the test signal using domain adaptation technique. The proposed dictionary learning process includes two coherence measures: atom–data coherence and mutual coherence to provide a learning procedure with low reconstruction error along with a proper separation in the test step. The simulation results using different measures show that the proposed method leads to significantly better results in comparison with the earlier methods in this context and the traditional procedures.

A deep learning method for Chinese singer identification

As a subfield of Multimedia Information Retrieval (MIR), Singer IDentification (SID) is still in the research phase. On one hand, SID cannot easily achieve high accuracy because the singing voice is difficult to model and always disturbed by the background instrumental music. On the other hand, the performance of conventional machine learning methods is limited by the scale of the training dataset. This study proposes a new deep learning approach based on Long Short-Term Memory (LSTM) and Mel-Frequency Cepstral Coefficient (MFCC) features to identify the singer of a song in large datasets. The results of this study indicate that LSTM can be used to build a representation of the relationships between different MFCC frames. The experimental results show that the proposed method achieves better accuracy for Chinese SID in the MIR-1K dataset than the traditional approaches.

Robust singer identification of Indian playback singers

Singing voice analysis has been a topic of research to assist several applications in the domain of music information retrieval system. One such major area is singer identification (SID). There has been enormous increase in production of movies and songs in Bollywood industry over the last 50 decades. Surveying this extensive dataset of singers, the paper presents singer identification system for Indian playback singers. Four acoustic features namely—formants, harmonic spectral envelope, vibrato, and timbre—that uniquely describe the singer are extracted from the singing voice segments. Using the combination of these multiple acoustic features, we address the major challenges in SID like the variations in singer’s voice, testing of multilingual songs, and the album effect. Systematic evaluation shows the SID is robust against the variations in singer’s singing style and structure of songs and is effective in identifying the cover songs and singers. The results are investigated on in-house cappella database consisting of 26 singers and 550 songs. By performing dimension reduction of the feature vector and using Support Vector Machine classifier, we achieved an accuracy of 86% using fourfold cross validation process. In addition, performance comparison of the proposed work with other existing approaches reveals the superiority in terms of volume of dataset and song duration.

Singer Identification in Indian Hindi Songs: A Review

Singer Identification in Indian Hindi Songs: A Review