Instrument identification in polyphonic music signals based on individual partials
14 Mar 2010-pp 401-404
TL;DR: A new approach to instrument identification based on individual partials is presented, which makes identification possible even when the concurrently played instrument sounds have a high degree of spectral overlapping.
Abstract: A new approach to instrument identification based on individual partials is presented. It makes identification possible even when the concurrently played instrument sounds have a high degree of spectral overlapping. A pairwise comparison scheme which emphasizes the specific differences between each pair of instruments is used for classification. Finally, the proposed method only requires a single note from each instrument to perform the classification. If more than one partial is available the resulting multiple classification decisions can be summarized to further improve instrument identification for the whole signal. Encouraging classification results have been obtained in the identification of four instruments (saxophone, piano, violin and guitar).
Citations
More filters
01 Dec 2016
TL;DR: Among the features used, after MFCC, ZCR proved to be the optimal feature for the classification of drum instrument, and the most significant feature for classifying Guitar, Violin and Drum is MFCC as it gives the better accurate results.
Abstract: Identification of musical instruments from the acoustic signal using speech signal processing methods is a challenging problem. Further, whether this identification can be carried out by a single musical note, like humans are able to do, is an interesting research issue that has several potential applications in the music industry. Attempts have been made earlier using the spectral and temporal features of the music acoustic signals. The process of identifying the musical instrument from monophonic audio recording basically involves three steps — pre-processing of music signal, extracting features from it and then classifying those. In this paper, we present an experiment-based comparative study of different features for classifying few musical instruments. The acoustic features, namely, the Mel-Frequency Cepstral Coefficients (MFCCs), Spectral Centroids (SC), Zero-Crossing Rate (ZCR) and signal energy are derived from the music acoustic signal using different speech signal processing methods. A Support Vector Machine (SVM) classifier is used with each feature for the relative comparisons. The classification results using different combinations of training by features from different music instrument and testing with another/same type of music instruments are compared. Our results indicate that the most significant feature for classifying Guitar, Violin and Drum is MFCC as it gives the better accurate results. Also, the feature which gives better accuracy results for the drum instrument is ZCR. Among the features used, after MFCC, ZCR proved to be the optimal feature for the classification of drum instrument.
11 citations
01 Oct 2017
TL;DR: The classification performance of overlapped soundtracks is effectively improved and singular spectrum analysis has been found to be an efficient way to discriminate speech/music in mixed soundtracks.
Abstract: This Soundtracks are information rich; knowledge can be extracted and discovered from them. Audio signals may be broadly divided into three classes: speech, music and events. Dedicated recognition algorithms are typically used to extract semantic data for further knowledge discovery. A preprocessor to classify and segment audio into these three classes is essential. Current practice often neglects that audio data from the real world can have any combination of these classes simultaneously. This can result in information loss, thus compromising the knowledge discovery. Singular Spectrum Analysis (SSA) is further developed to reduce the degree of overlap between classes of audio content, in order to mitigate such information losses and improve the performance of knowledge discovery. In particular the SSA method serves to mitigate the overlapping ratio between speech and music in the mixed soundtracks by generating two new soundtracks with a lower level of overlapping. Next, feature space is calculated for the output audio streams, and these are classified using random forests into either speech or music. The classification performance of overlapped soundtracks is effectively improved and singular spectrum analysis has been found to be an efficient way to discriminate speech/music in mixed soundtracks.
5 citations
15 Oct 2013
TL;DR: A new approach towards feature-based instrument recognition is presented that makes use of redundancies in the harmonic structure and temporal development of a note that is targeted at transferability towards use on polyphonic material.
Abstract: Instrument recognition is an important task in music information retrieval (MIR). Whereas the recognition of musical instruments in monophonic recordings has been studied widely, the polyphonic case still is far from being solved. A new approach towards feature-based instrument recognition is presented that makes use of redundancies in the harmonic structure and temporal development of a note. The structure of the proposed method is targeted at transferability towards use on polyphonic material. Multiple feature categories are extracted and classified separately with SVM models. In a further step, class probabilities are aggregated in a two-step combination scheme. The presented system was evaluated on a dataset of 3300 isolated single notes. Different aggregation methods are compared. As the results of the joined classification outperform individual categories, further development of the presented technique is motivated.
3 citations
Cites methods from "Instrument identification in polyph..."
...This concept has been applied by classifying instruments based on individual partials by Barbedo & Tzanetakis in [1]....
[...]
Proceedings Article•
01 Jan 2012
TL;DR: It turns out, that a selection of 505 features out of the full feature set of 1155 elements does only reduce the recognition rate of a linear SVM from 82% to 78% and with the use of a polynomial instead of alinear kernel the recognition rates with the reduced feature set can even be increased to 84%.
Abstract: A series of experiments on the automatic classification of classical guitar sounds with support vector machines has been carried out to investigate the relevance of the features and to minimise the feature set for successful classification. Features used for classification were the time series of the partial tone amplitudes, and of the MFCCs, and the energy distribution of the nontonal percussive sound that is produced in the attack phase of the tone. Furthermore the influence of sound parameters as timbre, player, fret position and string number on the recognition rate is investigated. Finally, several nonlinear kernels are compared in their classification performance. It turns out, that a selection of 505 features out of the full feature set of 1155 elements does only reduce the recognition rate of a linear SVM from 82% to 78%. With the use of a polynomial instead of a linear kernel the recognition rate with the reduced feature set can even be increased to 84%.
3 citations
Cites background from "Instrument identification in polyph..."
...Barbedo and Tsanetakis [2] published their results on the even more challenging task of instrument classification in polyphonic recordings....
[...]
01 Jan 2019
TL;DR: The proposed work relies on Chroma-based low-dimensional feature vector to categorize String instruments, which is an octave independent estimation of strength of all possible notes in Western 12 note scale at different points of time.
Abstract: Identification of instrument type from acoustic signal is a challenging issue. It is also an interesting and popular research area having several promising applications in music industry. Researchers have already been able to classify instruments into several broad categories like String, Woodwind, Percussion, and Keyboard etc. using acoustic features like Mel-Frequency Cepstral Coefficients (MFCCs), Zero Crossing Rate (ZCR) etc. MFCC has been found to be excessively used. In this work an alternative acoustic feature of MFCC has been proposed. Chroma is an octave independent estimation of strength of all possible notes in Western 12 note scale at different points of time. Sound envelope originated by a note reflects the signature of an instrument and this can be used to stratify String instruments into various categories. The proposed work relies on Chroma-based low-dimensional feature vector to categorize String instruments. For classification purpose, simple and popular classifiers like Neural Network, k-NN, Naive Bayes’ have been exercised.
1 citations
References
More filters
TL;DR: In this paper, a set of features is evaluated for recognition of musical instruments out of monophonic musical signals, aiming to achieve a compact representation, the adopted features regard only spectral characteristics of sound and are limited in number.
Abstract: A set of features is evaluated for recognition of musical instruments out of monophonic musical signals. Aiming to achieve a compact representation, the adopted features regard only spectral characteristics of sound and are limited in number. On top of these descriptors, various classification methods are implemented and tested. Over a dataset of 1007 tones from 27 musical instruments, support vectormachines and quadratic discriminant analysis show comparable results with success rates close to 70% of successful classifications. Canonical discriminant analysis never had momentous results, while nearest neighbours performed on average among the employed classifiers. Strings have been the most misclassified instrument family, while very satisfactory results have been obtained with brass and woodwinds. The most relevant features are demonstrated to be the inharmonicity, the spectral centroid, and the energy contained in the first partial.
120 citations
03 Oct 2001
TL;DR: Over a dataset of 1007 tones from 27 musical instruments and without employing any hierarchical structure, quadratic discriminant analysis shows the lowest error rate and outperforming all the other classification methods.
Abstract: A set of features is evaluated for musical instrument recognition out of monophonic musical signals. Aiming to achieve a compact representation, the adopted features regard only spectral characteristics of sound and are limited in number. On top of these descriptors, various classification methods are implemented and tested. Over a dataset of 1007 tones from 27 musical instruments and without employing any hierarchical structure, quadratic discriminant analysis shows the lowest error rate (7.19% for the individual instrument and 3.13% for instrument families), outperforming all the other classification methods (canonical discriminant analysis, support vector machines, nearest neighbours). The most relevant features are demonstrated to be the inharmonicity, the spectral centroid and the energy contained in the first partial.
120 citations
"Instrument identification in polyph..." refers background in this paper
...Index Terms— instrument identification, polyphonic musical signals, pairwise comparison, partial-based classification...
[...]
TL;DR: A new solution to the problem of feature variations caused by the overlapping of sounds in instrument identification in polyphonic music by weighting features based on how much they are affected by overlapping, which improves instrument identification using musical context.
Abstract: We provide a new solution to the problem of feature variations caused by the overlapping of sounds in instrument identification in polyphonic music. When multiple instruments simultaneously play, partials (harmonic components) of their sounds overlap and interfere, which makes the acoustic features different from those of monophonic sounds. To cope with this, we weight features based on how much they are affected by overlapping. First, we quantitatively evaluate the influence of overlapping on each feature as the ratio of the within-class variance to the between-class variance in the distribution of training data obtained from polyphonic sounds. Then, we generate feature axes using a weighted mixture that minimizes the influence via linear discriminant analysis. In addition, we improve instrument identification using musical context. Experimental results showed that the recognition rates using both feature weighting and musical context were 84.1% for duo, 77.6% for trio, and 72.3% for quartet; those without using either were 53.4, 49.6, and 46.5%, respectively.
91 citations
"Instrument identification in polyph..." refers background in this paper
...Index Terms— instrument identification, polyphonic musical signals, pairwise comparison, partial-based classification...
[...]
08 Nov 2004
TL;DR: The frequency envelope distribution (FED) algorithm is presented, which was introduced to musical duet separation and Artificial neural networks (ANNs) are employed as a decision system and they are trained with a set of feature vectors extracted from musical sounds recorded at the Multimedia Systems Department.
Abstract: The aim of this paper is to present solutions related to identifying musical data. These are discussed mainly on the basis of experiments carried out at the Multimedia Systems Department, Gdansk University of Technology, Gdansk, Poland. The topics presented in this paper include automatic recognition of musical instruments and separation of duet sounds. The classification process is shown as a three-layer process consisting of pitch extraction, parametrization, and pattern recognition. These three stages are discussed on the basis of experimental examples. Artificial neural networks (ANNs) are employed as a decision system and they are trained with a set of feature vectors (FVs) extracted from musical sounds recorded at the Multimedia Systems Department. The frequency envelope distribution (FED) algorithm is presented, which was introduced to musical duet separation. For the purpose of checking the efficiency of the FED algorithm, ANNs are also used. They are tested on FVs derived from musical sounds after the separation process is performed. The experimental results are shown and discussed.
88 citations
"Instrument identification in polyph..." refers background in this paper
...Index Terms— instrument identification, polyphonic musical signals, pairwise comparison, partial-based classification...
[...]
10 Oct 2001
72 citations
"Instrument identification in polyph..." refers background in this paper
...In those cases, normally the solo instrument has to be strongly dominant, so the signals characteristics are quasi-monophonic....
[...]