librosa: Audio and Music Signal Analysis in Python
By Brian McFee, Colin Raffel, Dawen Liang, D. Ellis, Matt McVicar, Eric Battenberg, Oriol Nieto. 2015
3,082 citations recorded by the source corpus.
This document describes version 0.4.0 of librosa: a Python pack- age for audio and music signal processing. At a high level, librosa provides implementations of a variety of common functions used throughout the field of music information retrieval. In this document, a brief overview of the library's functionality is provided, along with explanations of the design goals, software development practices, and notational conventions.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Speech & Audio, 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 Artificial Intelligence
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Dacheng Tao
- Trevor Darrell
- Ming-Hsuan Yang
- Xiaogang Wang
- Jitendra Malik
- Andrew Zisserman
- Bernt Schiele
In Speech & Audio
- Junichi Yamagishi
- Shinji Watanabe
- Lior Wolf
- Haizhou Li
- Sakriani Watiasri Sakti
- Satoshi Nakamura
- Andros Tjandra
- Wei Ping
In Signal Processing
- Junichi Yamagishi
- Xavier Serra
- Tuomas Virtanen
- Haizhou Li
- Lior Wolf
- Eduardo Fonseca
- Shinji Watanabe
- Justin Salamon
Citations in the corpus
- VQVAE Unsupervised Unit Discovery and Multi-scale Code2Spec Inverter for Zerospeech Challenge 2019 (2019, methodology): “Both MFCC and Mel-spectrogram are generated by the Li-brosa package [23].”
- Speech-to-Speech Translation Between Untranscribed Unknown Languages (2019, methodology): “We extracted both the MFCC and the linear magnitude spectrogram with Librosa [33] library.”
- PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition (2019, methodology): “We use the torchlibrosa,1 a PyTorch implementation of functions of librosa [46] to build log mel spectrogram extraction into PANNs.”
- Meta-StyleSpeech : Multi-Speaker Adaptive Text-to-Speech Generation (2021, methodology): “We downsampled an audio to 16kHz and trimmed leading and trailing silence using Librosa (McFee et al., 2015).”
- RAD-TTS: Parallel Flow-Based TTS with Robust Alignment Learning and Diverse Synthesis (2021, methodology): “We use a sampling rate of 22 050 Hz and mel-spectrograms with 80 bins using Librosa’s (McFee et al., 2015) mel filter defaults.”
- Balanced Multimodal Learning via On-the-fly Gradient Modulation (2022, methodology): “The whole audio data is transformed into a spectrogram of size 257×1,004 by librosa [28] using a window with length of 512 and overlap of 353.”
- Local Interpretable Model-Agnostic Explanations for Music Content Analysis (2017, methodology): “It uses a set of 30 Mel-filters to extract 30 MFCC coefficients (including the 0) and their first-order derivatives from each audio frame using Librosa [14].”
- Towards Achieving Robust Universal Neural Vocoding (2018, methodology): “The mel-spectrograms used for conditioning the network were extracted using Librosa library [25], with 80 coefficients and frequencies ranging from 50 Hz to 12 kHz.”
- Does Audio Deepfake Detection Generalize? (2022, positive): “We use Python, librosa [31] and scipy [32].”
- JDI-T: Jointly trained Duration Informed Transformer for Text-To-Speech without Explicit Alignment (2020, methodology): “The acoustic feature is 80dimensional mel-spectrogram extracted from audio with sampling rate of 22,050 Hz using the Librosa library [19].”
- Improved Prosody from Learned F0 Codebook Representations for VQ-VAE Speech Waveform Reconstruction (2020, methodology): “The silence segments were also trimmed in advance using Librosa [22].”
- Open-Source Practices for Music Signal Processing Research: Recommendations for Transparent, Sustainable, and Reproducible Audio Research (2019, background): “Many open source libraries exist for audio feature extraction, such as Essentia [23], librosa [24], aubio [25], Madmom [26], or Marsyas [27].”
- A Universal Music Translation Network (2018, methodology): “The automatic assessment employed the pitch tracker of the librosa package [24].”
- Multi-label Few-shot Learning for Sound Event Recognition (2019, methodology): “And, 128-bin log mel-spectrogram from raw audio is extracted with the librosa library [21] as the input feature to the neural networks.”
- Exploiting time-frequency patterns with LSTM-RNNs for low-bitrate audio restoration (2019, methodology): “The feature extraction and music reconstruction are implemented by the librosa Python package [35].”
- AI Choreographer: Music Conditioned 3D Dance Generation with AIST++ (2021, methodology): “We use the publicly available audio processing toolbox Librosa [60] to extract the music features including: 1-dim envelope , 20-dim MFCC , 12-dim chroma , 1-dim one-hot peaks and 1-dim one-hot beats , resulting in a 35-dim music feature.”