Subjective and Objective Quality Assessment of Audio Source Separation

By Valentin Emiya, Emmanuel Vincent, Niklas Harlander, Volker Hohmann. IEEE Transactions on Audio, Speech, and Language Processing, 2011

317 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Signal Processing

View Paper: https://www.semanticscholar.org/paper/1a72a7155d50f9b2e0e3c78244de2eb6b940d331

How researchers see this paper

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

In Cognitive Neuroscience

In Signal Processing

Citations in the corpus

The Conversation: Deep Audio-Visual Speech Enhancement (2018, methodology): “We evaluate the enhancement performance of the model in terms of perceptual speech quality using the blind source separation criteria described in [42] (we use the implementation provided by [43]).”
Can we Automatically Transform Speech Recorded on Common Consumer Devices in Real-World Environments into Professional Production Quality Speech?—A Dataset, Insights, and Challenges (2015, background): “It could therefore be beneficial to jointly consider denoising and dynamic range compression (rather than considering them as parts of independent subproblems) to attempt to reduce this issue.”
Nonnegative signal factorization with learnt instrument models for sound source separation in close-microphone recordings (2013, methodology): “Further information about this metrics can be found in [55].”
Equalization matching of speech recordings in real-world environments (2016, background): “Traditional objective metrics for speech enhancement [9] and audio source separation [21, 22] are not well suited to measure equalization matching performance due to the nature of the task.”
Direction of Arrival Based Spatial Covariance Model for Blind Sound Source Separation (2014, unknown): “The evaluation is based on objective separation quality metrics proposed in [30], [31] and perceptually motivated metrics [32].”
A Consolidated Perspective on Multimicrophone Speech Enhancement and Source Separation (2017, unknown): “SiSEC Authors of submission Method Average of SDR metric [38] Average of OPS metric [302]”
Crowdsourced Pairwise-Comparison for Source Separation Evaluation (2018, background): “As in our previous study [1], we compare our quality scores to the results of a lab-based listening test that followed the MUSHRA recommendation and was conducted by the developers of PEASS [6] to train the PEASS objective scoring models.”
Singing Voice Separation with Deep U-Net Convolutional Networks (2017, unknown): “Emiya et al. introduced a protocol for the subjective evaluation of source separation algorithms [7].”
Harp-Net: Hyper-Autoencoded Reconstruction Propagation for Scalable Neural Audio Coding (2021, unknown): “While SNR values are not completely reliable, their correlation with subjective score can sometime be found high [20].”
Objective Assessment of Perceptual Audio Quality Using ViSQOLAudio (2017, methodology): “PEMO-Q was tested with the default settings supplied in version 2.0 of the PEASS Toolkit [31], as optimized in [32].”
SDR – Half-baked or Well Done? (2018, methodology): “While the merits of various objective measures such as PESQ [12], Loizou’s composite measure [13], PEMO-Q [14], PEASS [15], or STOI [16], could be debated and compared, we are concerned here with an issue with the way the widely relied upon BSS eval toolbox [17] has been used.”