Improving Timbre Similarity : How high’s the sky ?

By Jean-Julien Aucouturier, François Pachet. 2004

425 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/55b2dffe1611470744911e78eda92686b19cfd75

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

State of the Art Report: Music Emotion Recognition: A State of the Art Review (2010, background): “…data (e.g., social factors, such as “Christmas music”, or songs considered to be “one hit wonders”) are not revealed in the audio, and results using only acoustic (spectral) features for Music-IR tasks have led many to believe there is an upper limit to performance using these features alone [71].”
Detection and Classification of Acoustic Scenes and Events (2015, background): “The “bag-of-frames” approach to audio classiﬁcation (see Section II) is relatively simple, and has been criticized for the assumptions it incurs [41].”
Song-Level Features and Support Vector Machines for Music Classification (2005, methodology): “They show promising results on 6-way genre classification tasks, with nearly 83% classification accuracy for their best system.”
A Simple Method to Determine if a Music Information Retrieval System is a “Horse” (2014, background): “Many researchers have observed a “performance limit” to MIR systems, which has been called a “glass ceiling,” and attributed to “signal only approaches” and the “semantic gap” [45]–[47], the lack considering the involvement of humans [44], or the inappropriateness of standard features and machine le”
A Survey of Audio-Based Music Classification and Annotation (2011, methodology): “A comparative study conducted by Aucouturier and Pachet [86] have shown that the performance of a simple Gaussian model for genre classification is highly competitive with those of GMM and -means clustering.”
Acoustic Scene Classification: Classifying environments from the sounds they produce (2014, methodology): “” Aucoutourier [3] studied the performance of a class of algorithms for audio timbre similarity, which followed a method similar to the ASC baseline.”
The music information retrieval evaluation exchange (2005-2007): A window into music information retrieval research (2008, background): “Toward Shattering the ‘‘Glass Ceiling’’ In 2004, Aucouturier and Pachet [10] published a paper wherein they noted that use of naı̈ve(2) audio-based timbral feature sets to perform MIR similarity tasks has real limitations in terms of successfully identifying musically similar pieces of audio.”