Opensmile: the munich versatile and fast open-source audio feature extractor

By Florian Eyben, Martin Wöllmer, Björn Schuller. 2010

3,249 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

AVEC 2013: the continuous audio/visual emotion and depression recognition challenge (2013, methodology): “In this Challenge, as was the case for AVEC 2012 and AVEC 2011, an extended set of features with respect to the INTERSPEECH 2009 Emotion Challenge (384 features) [19] and INTERSPEECH 2010 Paralinguistic Challenge (1 582 features) [20] is given to the participants, again using the freely available op”
Emotion Recognition from Speech Using Wav2vec 2.0 Embeddings (2021, methodology): “The eGeMAPS features were extracted using opensmilepython [33].”
Semi-Supervised Active Learning for Sound Classification in Hybrid Learning Environments (2016, methodology): “All features and functionals were extracted with the OpenSMILE toolkit [40].”
Knowledge-based Systems (2009, positive): “All the Machine-Learning processes (steps 2 to 4) were implemented within the Weka platform (University of Waikato, New Zealand, GNU General Public License [31]), while the feature extraction was performed using OpenSMILE (Audeering GmbH, Gilching, Germany [32]).”
Context-Dependent Sentiment Analysis in User-Generated Videos (2017, methodology): “(Metallinou et al., 2008) and (Eyben et al., 2010a) fused audio and textual modalities for emotion recognition.”
Conversational Memory Network for Emotion Recognition in Dyadic Dialogue Videos (2018, methodology): “To extract audio features we use openSMILE (Eyben et al., 2010).”
Missing Modality Imagination Network for Emotion Recognition with Uncertain Missing Modalities (2021, methodology): “Acoustic features: OpenSMILE toolkit (Eyben et al., 2010) with the configuration of “IS13 ComParE” is used to extract frame-level features, which have similar performance with the IS10 utterance-level acoustic features used in (Liang et al.”
A Survey on perceived speaker traits: Personality, likability, pathology, and the first challenge (2015, methodology): “The classification, performed by feeding 6 552 features available in openSMILE (Eyben et al., 2010) to BoosTexter (a boosting-based system for text categorisation), achieved an accuracy of 95 % for Extroversion and 63 % for Conscientiousness (no significant results for the other traits).”
AVEC 2011-The First International Audio/Visual Emotion Challenge (2011, methodology): “In this Challenge, an extended set of features with respect to the INTERSPEECH 2009 Emotion Challenge (384 features) [13] and INTERSPEECH 2010 Paralinguistic Challenge (1 582 features) [14] is given to the participants, again using the freely available open-source Emotion and Affect Recognition (ope”
ICON: Interactive Conversational Memory Network for Multimodal Emotion Detection (2018, methodology): “To extract audio features, we ﬁrst format the audio of each utterance-video as a 16-bit PCM WAV ﬁle and use the open-sourced software openSMILE (Eyben et al., 2010).”
Multimodal Sentiment Analysis using Hierarchical Fusion with Context Modeling (2018, methodology): “We use openSMILE [49], which is capable of automatic pitch and voice intensity extraction, for audio feature extraction.”
Smartphone-Based Recognition of States and State Changes in Bipolar Disorder Patients (2015, methodology): “Voice features: The open-source “openSmile” toolbox [34] is used to extract the acoustic features.”
Recurrent Neural Networks for Emotion Recognition in Video (2015, methodology): “It uses 1582 features extracted with the open-source Emotion and Aﬀect Recognition (openEAR) [12] toolkit which uses openSMILE [11] as backend.”
Alzheimer's Dementia Recognition through Spontaneous Speech: The ADReSS Challenge (2020, methodology): “Several statistical functions are applied to these features, resulting in a total of 988 features for every speech segment [11].”
Categorical and dimensional affect analysis in continuous input: Current trends and future directions (2013, background): “The openSMILE library is written in C++ and enables extraction of large audio feature spaces in real-time [164].”