Unsupervised learning of low-level audio features for music similarity estimation
نویسندگان
چکیده
While there is an enormous amount of music data available, the field of music analysis almost exclusively uses manually designed features. In this work we learn features from music data in a completely unsupervised way and evaluate them on a musical genre classification task. We achieve results very close to state-of-the-art performance which relies on highly hand-tuned feature extractors.
منابع مشابه
Unsupervised Audio Feature Extraction for Music Similarity Estimation
Fostered by the constant advancement of digital technologies, both catalogs of music distributors and personal music collections have grown to sizes that call for automated methods to manage them. In this context, music similarity estimation plays an important role: It can be used to recommend music based on examples, to organize a collection into groups, or to generate well-sounding playlists....
متن کاملMirex 2012 Submission Audio Classification Using High-dimensional Representations Learned on Standard Audio Features
We present a training/test framework for audio classification using learned feature representations. In contentbased music information retrieval tasks, standard audio features such as MFCC and chroma are typically used to represent the music content. As an alternative, there is increasing interest in learning feature representations from data using unsupervised learning algorithms. In the previ...
متن کاملI-Vectors for Timbre-Based Music Similarity and Music Artist Classification
In this paper, we present a novel approach to extract songlevel descriptors built from frame-level timbral features such as Mel-frequency cepstral coefficient (MFCC). These descriptors are called identity vectors or i-vectors and are the results of a factor analysis procedure applied on framelevel features. The i-vectors provide a low-dimensional and fixed-length representation for each song an...
متن کاملLearning the Similarity of Audio Music in Bag-of-frames Representation from Tagged Music Data
Due to the cold-start problem, measuring the similarity between two pieces of audio music based on their low-level acoustic features is critical to many Music Information Retrieval (MIR) systems. In this paper, we apply the bag-offrames (BOF) approach to represent low-level acoustic features of a song and exploit music tags to help improve the performance of the audio-based music similarity com...
متن کاملApplying the Independent Component Analysis for the estimation of an upper bound of the expressiveness of basic audio features commonly used in algorithms for computing music similarity Research Proposal
Many different methods for automatic retrieval of music similarity have been proposed. There are a number of low-level descriptors that are widely used as a basis for comparison algorithms. Although the different authors usually have different sets of songs to evaluate the algorithms and there usually is no exhaustive user evaluation, which makes the comparison between the different algorithms ...
متن کامل