Song-Level Features and Support Vector Machines for Music Classification

352 citations recorded by the source corpus.

How researchers see this paper

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

FSD50K: An Open Dataset of Human-Labeled Sound Events (2020, background): “We call this the “uploader effect” (following the analogy of the “album effect” [70] or “artist effect” [71]).”
High-level event recognition in unconstrained videos (2012, methodology): “[82] used Gaussian mixture models (GMM) to describe a song as a bag of frames for classifi-”
Album And Artist Effects For Audio Similarity At The Scale Of The Web (2009, background): “A single Gaussian (G1) with full covariance represents the MFCCs of each song [5].”
Learning-Based Methods for Comparing Sequences, with Applications to Audio-to-MIDI Alignment and Matching (2016, methodology): “For our baseline, we followed the tradition of (Tzanetakis and Cook, 2002; Deng et al., 2008; Mandel and Ellis, 2005; Foster et al., 2014) and others by summarizing spectrograms by their mean and standard deviation across feature dimensions and concatenating the resulting statistics into a fixed-len”
Learning Features from Music Audio with Deep Belief Networks (2010, methodology): “Practically all audio-based music genre classiﬁcation models use different types of acoustic features to drive supervised machine learning [4, 13, 14, 23].”
Automatic genre classification of music content: a survey (2006, methodology): “Typical models include Gaussian and Gaussian mixtures (GMMs) (GMMs have been used to build song timbre models in [17], [18] and [19]).”