mlsanityData to 7 Oct 2026

On loss functions and evaluation metrics for music source separation

arXiv 2202.07968 ↗First posted 16 Feb 2022cs.SD cs.LG eess.AScited by 17

Not on the map. Why?

Abstract (arXiv)

We investigate which loss functions provide better separations via benchmarking an extensive set of those for music source separation. To that end, we first survey the most representative audio source separation losses we identified, to later consistently benchmark them in a controlled experimental setup. We also explore using such losses as evaluation metrics, via cross-correlating them with the results of a subjective test. Based on the observation that the standard signal-to-distortion ratio metric can be misleading in some scenarios, we study alternative evaluation metrics based on the considered losses.

Key facts. Cited by 17 papers in the map (data through 7 Oct 2026). Builds on 20 of its 39 references that are in the map. First posted on arXiv 16 Feb 2022.

Builds on

20 of 39 references.

References · 20

  1. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,451
  2. 2016Improved Techniques for Training GANscited by 2,938
  3. 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,678
  4. 2015Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networkscited by 2,048
  5. 2016Least Squares Generative Adversarial Networkscited by 1,000
  6. 2016CNN Architectures for Large-Scale Audio Classificationcited by 701
  7. 2018SDR - half-baked or well done?cited by 420
  8. 2018Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separationcited by 400
  9. 2019Parallel WaveGAN: A fast waveform generation model based on generative adversarial networks with multi-resolution spectrogramcited by 337
  10. 2017SEGAN: Speech Enhancement Generative Adversarial Networkcited by 165
  11. 2018Wave-U-Net: A Multi-Scale Neural Network for End-to-End Audio Source Separationcited by 125
  12. 2019Music Source Separation in the Waveform Domaincited by 98
8 more
  1. 2020Upsampling artifacts in neural audio synthesiscited by 28
  2. 2020D3Net: Densely connected multidilated DenseNet for music source separationcited by 24
  3. 2018Fast Spectrogram Inversion using Multi-head Convolutional Neural Networkscited by 21
  4. 2020An empirical study of Conv-TasNetcited by 14
  5. 2019Demystifying TasNet: A Dissecting Approachcited by 12
  6. 2017Adversarial Semi-Supervised Audio Source Separation applied to Singing Voice Extractioncited by 11
  7. 2021On permutation invariant training for speech source separationcited by 4
  8. 2019Spectrogram Feature Losses for Music Source Separationcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 2; 2023: 3; 2024: 4; 2025: 5; 2026: 3.

17 papers.

Sort:

By area:

  1. 2023Multi-Source Diffusion Models for Simultaneous Music Generation and Separationcited by 63
  2. 2025Fast Text-to-Audio Generation with Adversarial Post-Trainingcited by 14
  3. 2024Simultaneous Music Separation and Generation Using Multi-Track Latent Diffusion Modelscited by 13
  4. 2023Latent Autoregressive Source Separationcited by 12
  5. 2022Music Separation Enhancement with Generative Modelingcited by 8
  6. 2022Adversarial Permutation Invariant Training for Universal Sound Separationcited by 8
  7. 2025MGE-LDM: Joint Latent Diffusion for Simultaneous Music Generation and Source Extractioncited by 4
  8. 2025Efficient and Fast Generative-Based Singing Voice Separation using a Latent Diffusion Modelcited by 2
  9. 2023Distribution Preserving Source Separation With Time Frequency Predictive Modelscited by 1
  10. 2024Improving Music Source Separation with Diffusion and Consistency Refinementcited by 1
  11. 2025StutterZero and StutterFormer: End-to-End Speech Conversion for Stuttering Transcription and Correctioncited by 1
  12. 2024Spectral Mapping of Singing Voices: U-Net-Assisted Vocal Segmentationcited by 0
5 more
  1. 2024Source Separation of Multi-source Raw Music using a Residual Quantized Variational Autoencodercited by 0
  2. 2025Solving Copyright Infringement on Short Video Platforms: Novel Datasets and an Audio Restoration Deep Learning Pipelinecited by 0
  3. 2026Music-Source-Separation-Training (MSST): A Unified Framework for Training and Evaluating Music Demixing Modelscited by 0
  4. 2026Iterative Audio Separation with Mixture Consistency via MIMO Model Extensioncited by 0
  5. 2026Prism-SQA: An Interpretable and Adaptable Neural Framework for Surface Electromyography Quality Assessmentcited by 0