Mel Frequency Cepstral Coefficients for Music Modeling

By Beth T. Logan. International Society for Music Information Retrieval Conference, 2000

1,381 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/55afc2d63fd410a719c3bbe9772d9bbc6bc565a6

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Song-Level Features and Support Vector Machines for Music Classification (2005, methodology): “While originally developed to decouple vocal excitation from vocal tract shape for automatic speech recognition (Oppenheim, 1969), they have found applications in other auditory domains including music retrieval (Logan, 2000; Foote, 1997).”
A Highly Robust Audio Fingerprinting System (2002, background): “Features that have been proposed are well known audio features such as Fourier coefficients [4], Mel Frequency Cepstral Coefficients (MFFC) [18], spectral flatness [2], sharpness [2], Linear Predictive Coding (LPC) coefficients [2] and others.”
Local Interpretable Model-Agnostic Explanations for Music Content Analysis (2017, background): “What does this mean in terms of the qualities of the input sound, however? Why X? Is this a real-world general principle? Or does it arise from a peculiarity of the training dataset? MFCCs were introduced for speech recognition [4], but have been argued as suitable for machine music listening [9,12]”
Automatic Musical Genre Classification of Audio Signals (2001, methodology): “The use of MFCC as features for classifying music vs speech has been explored in [22].”
Album And Artist Effects For Audio Similarity At The Scale Of The Web (2009, unknown): “[4]).”
A Large-Scale Evaluation of Acoustic and Subjective Music-Similarity Measures (2004, methodology): “D ow nloaded from http://direct.m it.edu/com j/article-pdf/28/2/63/1854081/014892604323112257.pdf by guest on 30 June 2021 65Berenzweig, Logan, Ellis, and Whitman The triangle inequality can be violated because of the multifaceted nature of similarity: for example, Michael Jackson is similar to the ”
Improving Timbre Similarity : How high’s the sky ? (2004, background): “Moreover, it has already been reported that the DCT was a satisfying approximation of the K-L transform in the case of music signals ([21].”
Hearing aid Research Data Set for Acoustic Environment Recognition (2020, background): “the mel frequency cepstral coefficients, the so called MFCCs [19]) optimized for computational resources are yet to be developed by the researchers for real-time capable applications.”
Zero-Shot Event Detection by Multimodal Distributional Semantic Embedding of Videos (2015, methodology): “We performed an experiment on a set of audio concepts trained on MFCC audio features (Davis and Mermelstein 1980; Logan and others 2000).”
Artist detection in music with Minnowmatch (2001, methodology): “The MFCC approximates an ndimensional PCA by means of the Discrete Cosine Transform, and it then scales the frequency bands along a more logical range for audio perception [7].”
Exploring automatic music annotation with "acoustically-objective" tags (2010, background): “Mel-frequency Cepstral Coefficients (MFCC) are shorttime audio representations that have been used for many speech recognition and music analysis tasks [12, 17].”
Automatic Video Classification: A Survey of the Literature (2008, methodology): “These are then transformed using a DCT or a Hadamard transform (HT).”
Multi-Label Classification of Music into Emotions (2008, methodology): “Mel Frequency Cepstral Coefﬁcients (MFCCs) are used for speech recognition and music modeling [8].”
Recommendations : Modeling Music Ratings with Temporal Dynamics and Item Taxonomy (2011, unknown): “For example, MelFrequency Cepstral Coefficients (MFCCs) are often used to generate feature vectors, which can be used to find other items with acoustic resemblance [3, 18].”
Linking Image and Text with 2-Way Nets (2016, methodology): “The acoustic data are the MFCCs [25] for the same frames, yielding a 273 dimensional vector at each point in time.”
State of the Art Report: Music Emotion Recognition: A State of the Art Review (2010, background): “MacDorman et al. examined the ability of multiple acoustic features (sonogram, spectral histogram, periodicity histogram, ﬂuctuation pattern, and mel-frequency cepstral coefﬁcients–MFCCs [51,52]) to predict pleasure and arousal ratings of music excerpts.”
Musical genre classification of audio signals (2002, methodology): “The use of MFCCs to separate music and speech has been explored in [19].”