Pith. sign in

REVIEW 4 major objections 3 minor 49 references

Separate This, and All of these Things Around It: Music Source Separation via Hyperellipsoidal Queries

T0 review · 4 major / 3 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Music source separation can be driven by hyperellipsoidal region queries: a single model extracts any stem or composite whose embedding lies inside a user-drawn region, with controllable broadness, and reports state-of-the-art results on…

desk verdict The hyperellipsoid query is a real extension of the authors' Banquet system, and the retrieval evaluation is a useful addition, but the state-of-the-art SNR claim rests on an oracle choice of the query-width parameter and a mismatched baseline. read the letter →

arxiv 2501.16171 v1 pith:2T4JGLI2 submitted 2025-01-27 eess.AS cs.IRcs.LGcs.SD

Karn N. Watcharasupat , Alexander Lerch This is my paper · ORCID
classification eess.AScs.IRcs.LGcs.SD
keywords musicsourceseparationquery-by-regionhyperellipsoidquerystem-agnosticMoisesDBPaSSTembeddingretrievalmetricsFiLMconditioning
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Music source separation has long been locked to a fixed menu of stems—vocals, drums, bass, and "other"—because the best models are trained to output exactly those four. This paper tries to break that lock: it claims a single model can extract any target, a lone instrument or an arbitrary blend of instruments, when the target is specified as a hyperellipsoid (a multidimensional ellipse) drawn in a pretrained audio-embedding space. The user sets both where the ellipse sits (the timbre wanted) and how wide it is (how much of the surrounding timbral neighborhood to include), and the model pulls out exactly the sources whose embeddings fall inside it. If the claim holds, one trained model replaces stacks of single-stem extractors, reaches instrument classes it never saw in training, and gives musicians a single continuous knob for extraction broadness. The paper reports state-of-the-art signal-to-noise ratios and retrieval scores on MoisesDB in support of that claim.

What carries the argument

The load-bearing object is the hyperellipsoid query $Q(c,K) = \{z \in \mathbb{R}^P : (z-c)^\top K^{-1}(z-c) \leq 1\}$, a Mahalanobis-distance ball specified by a center $c$ and a positive-definite spread matrix $K$. The query space is the 768-dimensional PaSST embedding, reduced by PCA to 128 dimensions (91.8\% explained variance on the training set), and the query is packed into a vector of the center plus the lower-triangular entries of $K$, which a small fully connected network maps to FiLM parameters $\gamma, \beta$ that rescale and shift the mixture embedding at the decoder bottleneck. Training queries are precomputed per clip by finding, for every possible target subset, the smallest ellipsoid enclosing the target embeddings and the largest same-center ellipsoid excluding all non-target embeddings, then sampling radii uniformly between the two. A level-matching regularizer with adaptive weighting keeps the output from collapsing to near silence, which the authors report happens without it.

What would settle it

Compute, for every stem in MoisesDB, the smallest enclosing hyperellipsoid around that stem's embeddings and count how often a non-target stem's embeddings fall inside; if pairs such as kick drum and bass guitar, which share low-frequency energy, cannot be isolated by any ellipsoid, per-pair retrieval should fall to chance, which would refute the claim that region queries can specify arbitrary targets.

Watch

Extended reading notes

Core claim

The central claim is that query-by-region works as a general formalism for music source separation: given a mixture and a hyperellipsoid in a discriminative embedding space, the model recovers exactly the sum of the sources whose embeddings lie inside the ellipsoid, regardless of how many sources or which classes the target set contains. The paper extends the point-query Banquet architecture so that the conditioning input is a full hyperellipsoid—center plus positive-definite spread matrix—mapped to FiLM parameters that adapt the mixture embedding at the bottleneck of a time-frequency masking network. Because a hyperellipsoid is the level set of a Mahalanobis distance, it is a natural geometric stand-in for a multivariate Gaussian cluster, and interpolating between a smallest enclosing ellipsoid and a largest excluding ellipsoid generates valid training queries for every source subset in each clip. On MoisesDB the system reports state-of-the-art SNR and retrieval metrics, including macro and micro average precision of 0.83 and 0.86, and it recovers long-tail instruments (organ, synth, brass, reeds, strings) where its point-query predecessor collapsed to silence.

Load-bearing premise

The load-bearing premise is that the pretrained PaSST audio-embedding space groups sounds by instrument well enough that a hyperellipsoid drawn in it always isolates a musically meaningful target set from the non-target sounds in the same mixture.

Editorial extensions

If this is right

  • A single trained model can extract any single stem or any composite target whose sources can be enclosed by a hyperellipsoid, including instrument classes never seen in training: viola is absent from the training set yet is extracted at a median SNR of 6.1 dB.
  • Users gain a continuous broadness control: scaling the query radii toward the excluding ellipsoid widens the extraction, and the reported ROC analysis shows the effect is class-dependent, with bass guitar insensitive to the scale factor while grand piano and brass degrade markedly at the wrong setting.
  • Long-tail instruments that collapsed in the point-query predecessor are recovered: organs, synths, brass, reeds, and strings all move from zero SNR to positive median SNR, with the largest gains on exactly the classes that the fixed-stem paradigm serves worst.
  • The least-squares projection evaluation turns an audio-format separation output into per-source retrieval scores, giving query-based systems a way to separate "did it find the right sources" from "is the audio clean," and yields macro and micro average precision of 0.83 and 0.86.
  • Performance tracks the fraction of the mixture requested: median SNR and weighted mean average precision both rise as the target-to-mixture source ratio grows, so query difficulty behaves like a standard retrieval setting where the relevant proportion of the collection sets the difficulty.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the hyperellipsoid is a level set of a multivariate Gaussian, so the natural next step—one the paper lists as future work—is a Gaussian-mixture query that extracts sources from several disjoint timbral regions at once; the same FiLM conditioning machinery should carry over unchanged.
  • Editorial inference: the method's ceiling is set by the embedding space, not the decoder; swapping the PaSST query space for another pretrained or task-fine-tuned embedding and checking whether mAP tracks instrument-discriminability would isolate where the query-by-region gains come from.
  • Editorial inference: the class-dependent sensitivity to query width suggests that optimal broadness is a property of the target's timbral neighborhood; an automatic radius selector tuned on validation retrieval metrics would remove the need to sweep the scale factor at test time.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 3 minor

Summary. The paper proposes a music source separation system that generalizes query-by-example to query-by-region: the target is specified by a hyperellipsoid in the embedding space of a pretrained PaSST audio classifier, with the ellipsoid center and semi-axes controlling which instrument embeddings are included. This enables arbitrary composite targets and user-controllable broadness. The model is a FiLM-conditioned encoder-decoder trained on MoisesDB with queries precomputed from all source subsets, using enclosing/excluding ellipsoids to define valid targets. The paper reports single-source and multi-source SNR and retrieval metrics, claiming state-of-the-art performance.

Significance. If the results are taken at face value, the system would be a significant step beyond fixed-stem separation: a single model can extract arbitrary composites specified geometrically, including long-tail instruments for which prior query-based systems collapsed. The retrieval-evaluation methodology via least-squares projection is a useful contribution. The use of a pretrained, frozen PaSST embedding avoids the circularity of training the query space itself, and the training-data generation from all subsets is a systematic approach. However, the headline SOTA claim is not yet supported by the evaluation protocol.

major comments (4)
  1. [Section 4.1, Figure 5, Table B.II] The single-source SNR figures are computed using the clip-wise best α, i.e., an oracle selection of the query scale factor per clip. Since α is a query parameter a user must set before hearing the output, these median SNRs are an upper envelope over α rather than the performance of any deployable configuration. The abstract's claim of state-of-the-art SNR therefore rests on an oracle evaluation. Please report performance at a fixed α, or average over α with a specified selection rule, and discuss the trade-off between broadness and SNR.
  2. [Section 4.1, Table B.II] The only comparative baseline, Banquet, is evaluated under a different protocol: full-track overlap-add for Banquet versus clip-wise evaluation for the proposed system. The paper itself cautions that the comparison is only a rough gauge, yet the abstract and conclusion claim state-of-the-art performance. With a single mismatched baseline, this claim is not supported. Please add matched-protocol comparisons against at least one fixed-stem SOTA system (e.g., HTDemucs) on the same MoisesDB split, and either full-track overlap-add evaluation for the proposed system or clip-wise evaluation for the baseline.
  3. [Section 3.1] The query precomputation restricts training and evaluation to target subsets for which an enclosing hyperellipsoid excludes all non-target sources; if a non-target embedding falls inside the enclosing ellipsoid, that source is removed from the mixture. This guarantees that every query is separable in the PaSST space by construction, so the evaluation does not measure how often the query-by-region formalism fails for realistic subsets. The paper should report the fraction of subsets discarded or made infeasible, and should evaluate on all subsets (including non-separable ones) to test the underlying assumption that PaSST embeddings cluster by instrument.
  4. [Section 3.2, Table B.III] The thresholded retrieval metrics (accuracy, precision, recall, F1) require a decision threshold on the least-squares scores, and Table B.III reports a threshold per stem without stating how it was selected. If these thresholds are tuned on the test set, the metrics are optimistic. Please state the selection procedure (e.g., validation-set optimization or a fixed threshold), or report unthresholded metrics such as ROCAUC/PRAUC as primary.
minor comments (3)
  1. [Section 2.1, Eq. (8)] The vector representation q^T = [c^T tril(K)^T] is underspecified: tril(K) is not defined in the text, and the dimension of q should be stated explicitly. Please clarify the ordering of the lower-triangular entries.
  2. [Section 3.1] The notation switches between Λ = diag(λ) and r = λ^{1/2} in the definitions of inclusion/exclusion radii; please use a single consistent notation to avoid confusion.
  3. [Figure 2 caption] The caption appears truncated after the hyperplane definition; the sentence about the two-dimensional cross-section is incomplete and should be rewritten for clarity.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline SNR claim is an oracle upper bound: clip-wise best query width α is selected using ground truth, so the reported state-of-the-art SNR is forced by the selection rule.

  1. fitted input called prediction [Section 4.1, Figure 5, Table B.II]
    "For each clip, we query the system using hyperellipsoids centered at the centroid of the target embedding and with radii ˜r = αr⊥, for various α ∈ [10−3, 1]. ... Using the clip-wise best α and considering only the clips where all sources are included in the mixture, the median SNRs and root-mean-square (RMS) errors of the proposed method are shown in Figure 5, and compared against the reported results in Banquet (Q:ALL, TE+DA variant). Note again that our method was evaluated clip-wise while Banquet was evaluated over the full track."

    The reported single-source SNR is computed after selecting α per clip using the ground-truth target. This makes the reported median a maximum over ten query settings rather than the performance of any fixed, user-specifiable query. The abstract's 'state-of-the-art performance ... in terms of signal-to-noise ratios' is therefore an oracle-envelope claim, not a prediction of the model with a determined query. The simultaneous change of evaluation protocol (clip-wise here vs full-track overlap-add for Banquet) means the SOTA comparison is not established by the reported numbers.

full rationale

Apart from the oracle-α evaluation, the system's derivation is self-contained: the hyperellipsoid query space, FiLM conditioning, loss, and multi-source query construction are specified in the paper, and the PaSST embeddings come from a pretrained external model rather than being fit to the separation targets. Self-citations to Banquet for architecture, loss, and data splits are normal and are not load-bearing for the new contribution; the comparison to Banquet is a baseline, not a circular premise. The multi-source retrieval metrics and ROC curves are not oracle-selected and provide independent content. However, the headline SNR claim reduces by construction to the clip-wise best α selection, so the paper is partially circular in its central evaluation claim.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim rests mainly on the PaSST query space and the separability of source embeddings by ellipsoids, plus an unvalidated retrieval evaluation. The model itself introduces no new physical entities; the hyperellipsoid query is a geometric input representation. The free parameters above are mostly tuning choices that are not fully disclosed, and the test-time alpha selection is the most consequential for interpreting the reported performance.

free parameters (6)
  • D (query embedding PCA dimension) = 128
    Chosen by hand to match the bottleneck dimension; PCA retains 91.8% variance. Affects the geometry of the query space and the size of the FiLM network input.
  • epsilon (pseudo-inverse threshold) = not specified
    Used in Eq. (7) to compute the truncated inverse of K; small threshold chosen by the authors but no value or sensitivity analysis is given.
  • delta (single-source covariance floor) = not specified, small
    Set in the query precomputation when |Z|=1 so that K=delta*I; its value is not reported.
  • lambda0, delta-lambda, Lmin (regularization weighting) = not specified
    Parameters of the adaptive level-matching regularization in Eq. (10); no values or tuning procedure are given.
  • alpha (query scale factor in single-source evaluation) = swept over [1e-3, 1], clip-wise best selected
    The reported single-source SNR uses the best alpha per clip, an oracle choice that can inflate results versus a fixed deployment rule.
  • Retrieval metric thresholds = per-class thresholds reported in Table B.III
    The thresholds used to compute accuracy/precision/recall/F1 are listed but the selection rule (e.g., chosen on test set to maximize F1) is not stated.
assumptions (4)
  • domain assumption The mixture is a linear instantaneous sum of the ground-truth sources (Eq. 3).
    The system and evaluation assume no reverberation or nonlinear effects; this is standard in the MSS literature.
  • domain assumption The PaSST embedding space is sufficiently discriminative that a hyperellipsoid region corresponds to a musically meaningful set of sources.
    Invoked in Section 2.1 where PaSST is chosen as the query space; the entire query-by-region concept depends on this.
  • domain assumption The least-squares projection of the model output onto the known source signals yields valid retrieval scores for source presence.
    Section 3.2 defines the retrieval metrics via this projection but provides no validation against human judgments or an alternative retrieval method; correlated sources could make the projection ill-conditioned.
  • domain assumption For every training subset, a valid hyperellipsoid exists that contains all target embeddings and excludes all non-target embeddings.
    The precomputed query generation in Section 3.1 assumes such separability; subsets where the covariance-based enclosing and excluding ellipsoids overlap are discarded or handled by edge-case rules, biasing training data toward separable targets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Separate This, and All of these Things Around It: Music Source Separation via Hyperellipsoidal Queries." pith.science (2026). https://pith.science/paper/2T4JGLI2

@misc{pith2026250116171,
  author       = {Pith},
  title        = {Pith review of: Separate This, and All of these Things Around It: Music Source Separation via Hyperellipsoidal Queries},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2T4JGLI2}},
  note         = {Machine review of arXiv:2501.16171}
}
read the original abstract

Music source separation is an audio-to-audio retrieval task of extracting one or more constituent components, or composites thereof, from a musical audio mixture. Each of these constituent components is often referred to as a "stem" in literature. Historically, music source separation has been dominated by a stem-based paradigm, leading to most state-of-the-art systems being either a collection of single-stem extraction models, or a tightly coupled system with a fixed, difficult-to-modify, set of supported stems. Combined with the limited data availability, advances in music source separation have thus been mostly limited to the "VDBO" set of stems: \textit{vocals}, \textit{drum}, \textit{bass}, and the catch-all \textit{others}. Recent work in music source separation has begun to challenge the fixed-stem paradigm, moving towards models able to extract any musical sound as long as this target type of sound could be specified to the model as an additional query input. We generalize this idea to a \textit{query-by-region} source separation system, specifying the target based on the query regardless of how many sound sources or which sound classes are contained within it. To do so, we propose the use of hyperellipsoidal regions as queries to allow for an intuitive yet easily parametrizable approach to specifying both the target (location) as well as its spread. Evaluation of the proposed system on the MoisesDB dataset demonstrated state-of-the-art performance of the proposed system both in terms of signal-to-noise ratios and retrieval metrics.

Figures

Figures reproduced from arXiv: 2501.16171 by the authors.

Figure 1
Figure 1. Overview of the Proposed System region query apart from the aforementioned hyperbolic system [Petermann et al., 2023]. In addition, given the difficulty of evaluating the retrieval performance of query-based MSS systems with a large num￾ber of stems and corresponding large timbral diversity, we also introduce a method to evaluate our audio-to-audio MSS system as a retrieval system. This is done by using simple least… view at source ↗
Figure 2
Figure 2. A two-dimensional cross-section of a hyperellipsoid in [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Simplified two-dimensional representation of some possible [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Single-source queries: SNR (dB) distribution by target “stem” over query scale factors, [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Single-source queries: Plot of the median SNRs (dB) and [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Single-source queries: ROC for lead bass guitar, grand [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 8
Figure 8. Figure 8: Multi-source queries: Weighted mean average precision by [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 48 canonical work pages

  1. [1]

    Zero-shot Audio Source Separation through Query-based Learning from Weakly-labeled Data

    Ke Chen, Xingjian Du, Bilei Zhu, Zejun Ma, Taylor Berg-Kirkpatrick , and Shlomo Dubnov. Zero-shot Audio Source Separation through Query-based Learning from Weakly-labeled Data . In Proceedings of the 36th AAAI Conference on Artificial Intelligence , Online, February 2022. AAAI

    work page 2022

  2. [2]

    LaSAFT : Latent Source Attentive Frequency Transformation For Conditioned Source Separation

    Woosung Choi, Minseok Kim, Jaehwa Chung, and Soonyoung Jung. LaSAFT : Latent Source Attentive Frequency Transformation For Conditioned Source Separation . In Proceedings of the 2021 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , pages 171--175. IEEE, June 2021

    work page 2021

  3. [3]

    Demucs: Deep Extractor for Music Sources with extra unlabeled data remixed, 2019

    Alexandre D \'e fossez, Nicolas Usunier, L \'e on Bottou, and Francis Bach. Demucs: Deep Extractor for Music Sources with extra unlabeled data remixed, 2019

    work page 2019

  4. [4]

    Music Source Separation in the Waveform Domain , November 2019

    Alexandre D \'e fossez, Nicolas Usunier, L \'e on Bottou, and Francis Bach. Music Source Separation in the Waveform Domain , November 2019

    work page 2019

  5. [5]

    Hybrid Spectrogram and Waveform Source Separation

    Alexandre D \'e fossez. Hybrid Spectrogram and Waveform Source Separation . In Proceedings of the 2021 Music Demixing Workshop at the 22nd International Society for Music Information Retrieval Conference ( ISMIR ) , Online, November 2021. ISMIR

    work page 2021

  6. [6]

    The Sound Demixing Challenge 2023 -- Music Demixing Track

    Giorgio Fabbro, Stefan Uhlich, Chieh-Hsin Lai, Woosung Choi, Marco Mart \'i nez-Ram \'i rez , Weihsiang Liao, Igor Gadelha, Geraldo Ramos, Eddie Hsu, Hugo Rodrigues, Fabian-Robert St \"o ter, Alexandre D \'e fossez, Yi Luo, Jianwei Yu, Dipam Chakraborty, Sharada Mohanty, Roman Solovyev, Alexander Stempkovskiy, Tatiana Habruseva, Nabarun Goswami, Tatsuya H...

    work page 2023

  7. [7]

    One- Shot Conditional Audio Filtering of Arbitrary Sounds

    Beat Gfeller, Dominik Roblek, and Marco Tagliasacchi. One- Shot Conditional Audio Filtering of Arbitrary Sounds . In Proceedings of the 2021 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , pages 501--505, Toronto, Canada, June 2021. IEEE

    work page 2021

  8. [8]

    Spleeter: A fast and efficient music source separation tool with pre-trained models

    Romain Hennequin, Anis Khlif, Felix Voituret, and Manuel Moussallam. Spleeter: A fast and efficient music source separation tool with pre-trained models. Journal of Open Source Software , 5(50):2154, June 2020

    work page 2020

Show all 49 references
  1. [9]

    LightSAFT : Lightweight Latent Source Aware Frequency Transform for Source Separation

    Yeong-Seok Jeong, Jinsung Kim, Woosung Choi, Jaehwa Chung, and Soonyoung Jung. LightSAFT : Lightweight Latent Source Aware Frequency Transform for Source Separation . In Proceedings of the 2021 Music Demixing Workshop at the 22nd International Society for Music Information Ret...

    2021

  2. [10]

    KUIELab-MDX-Net : A Two-Stream Neural Network for Music Demixing

    Minseok Kim, Woosung Choi, Jaehwa Chung, Daewon Lee, and Soonyoung Jung. KUIELab-MDX-Net : A Two-Stream Neural Network for Music Demixing . In Proceedings of the 2021 Music Demixing Workshop at the 22nd International Society for Music Information Retrieval Conference ( ISMIR )...

    2021

  3. [11]

    Berg, Wan-Yen Lo, Piotr Doll \'a r, and Ross Girshick

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C. Berg, Wan-Yen Lo, Piotr Doll \'a r, and Ross Girshick. Segment Anything , April 2023

    2023

  4. [12]

    Decoupling Magnitude and Phase Estimation with Deep ResUNet for Music Source Separation

    Qiuqiang Kong, Yin Cao, Haohe Liu, Keunwoo Choi, and Yuxuan Wang. Decoupling Magnitude and Phase Estimation with Deep ResUNet for Music Source Separation . In Proceedings of the 22nd International Society for Music Information Retrieval Conference ( ISMIR ) , pages 342--349, O...

    2021

  5. [13]

    Plumbley

    Qiuqiang Kong, Ke Chen, Haohe Liu, Xingjian Du, Taylor Berg-Kirkpatrick , Shlomo Dubnov, and Mark D. Plumbley. Universal Source Separation with Weakly Labelled Data , May 2023

    2023

  6. [14]

    Efficient Training of Audio Transformers with Patchout

    Khaled Koutini, Jan Schl \"u ter, Hamid Eghbal-zadeh , and Gerhard Widmer. Efficient Training of Audio Transformers with Patchout . In Proceedings of the 23rd Annual Conference of the International Speech Communication Association ( Interspeech ) , pages 2753--2757, Incheon, K...

    2022

  7. [15]

    Jonathan Le Roux, Scott Wisdom, Hakan Erdogan, and John R. Hershey. SDR - Half-baked or Well Done ? In Proceedings of the 2019 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , pages 626--630, Brighton, United Kingdom, 2019. IEEE

    2019

  8. [16]

    Audio query-based music source separation

    Jie Hwan Lee, Hyeong-Seok Choi, and Kyogu Lee. Audio query-based music source separation. In Proceedings of the 20th International Society for Music Information Retrieval Conference ( ISMIR ) , Delft, Netherlands, 2019. ISMIR

    2019

  9. [17]

    Creating a Multitrack Classical Music Performance Dataset for Multimodal Music Analysis : Challenges , Insights , and Applications

    Bochen Li, Xinzhao Liu, Karthik Dinesh, Zhiyao Duan, and Gaurav Sharma. Creating a Multitrack Classical Music Performance Dataset for Multimodal Music Analysis : Challenges , Insights , and Applications . IEEE Transactions on Multimedia , 21(2):522--535, February 2019

    2019

  10. [18]

    A Unified Model for Zero-shot Music Source Separation , Transcription and Synthesis

    Liwei Lin, Qiuqiang Kong, Junyan Jiang, and Gus Xia. A Unified Model for Zero-shot Music Source Separation , Transcription and Synthesis . In Proceedings of the 22nd International Society for Music Information Retrieval Conference ( ISMIR ) , pages 381--388, Online, August 2021. ISMIR

    2021

  11. [19]

    Plumbley, and Wenwu Wang

    Xubo Liu, Haohe Liu, Qiuqiang Kong, Xinhao Mei, Jinzheng Zhao, Qiushi Huang, Mark D. Plumbley, and Wenwu Wang. Separate What You Describe : Language-Queried Audio Source Separation . In Proceedings of the 23rd Annual Conference of the International Speech Communication Associa...

    2022

  12. [20]

    Plumbley, and Wenwu Wang

    Xubo Liu, Qiuqiang Kong, Yan Zhao, Haohe Liu, Yi Yuan, Yuzhuo Liu, Rui Xia, Yuxuan Wang, Mark D. Plumbley, and Wenwu Wang. Separate Anything You Describe , October 2023

    2023

  13. [21]

    The 2016 Signal Separation Evaluation Campaign

    Antoine Liutkus, Fabian-Robert St \"o ter, Zafar Rafii, Daichi Kitamura, Bertrand Rivet, Nobutaka Ito, Nobutaka Ono, and Julie Fontecave. The 2016 Signal Separation Evaluation Campaign . In Proceedings of the 13th International Conference on Latent Variable Analysis and Signal...

    2016

  14. [22]

    Decoupled Weight Decay Regularization

    Ilya Loshchilov and Frank Hutter. Decoupled Weight Decay Regularization . In Proceedings of the 2019 International Conference on Learning Representations ( ICLR ) , May 2019

    2019

  15. [23]

    Music Source Separation with Band-Split RoPE Transformer

    Wei-Tsung Lu, Ju-Chiang Wang, Qiuqiang Kong, and Yun-Ning Hung. Music Source Separation with Band-Split RoPE Transformer . In Proceedings of the 2024 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , pages 481--485, Seoul, Korea, Republic o...

    2024

  16. [24]

    Music Source Separation With Band-Split RNN

    Yi Luo and Jianwei Yu. Music Source Separation With Band-Split RNN . IEEE/ACM Transactions on Audio, Speech, and Language Processing , 31:1893--1901, 2023

    1901

  17. [25]

    Source Separation By Steering Pretrained Music Models

    Ethan Manilow, Patrick O'Reilly, Prem Seetharaman, and Bryan Pardo. Source Separation By Steering Pretrained Music Models . In Proceedings of the 2022 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , pages 126--130, Singapore, Singapore, M...

    2022

  18. [26]

    Conditioned- U-Net : Introducing a Control Mechanism in the U-Net for Multiple Source Separations

    Gabriel Meseguer-Brocal and Geoffroy Peeters. Conditioned- U-Net : Introducing a Control Mechanism in the U-Net for Multiple Source Separations . In Proceedings of the 20th International Society for Music Information Retrieval Conference ( ISMIR ) , pages 159--165, Delft, Neth...

    2019

  19. [27]

    Music Demixing Challenge 2021

    Yuki Mitsufuji, Giorgio Fabbro, Stefan Uhlich, Fabian-Robert St \"o ter, Alexandre D \'e fossez, Minseok Kim, Woosung Choi, Chin-Yun Yu, and Kin-Wai Cheuk. Music Demixing Challenge 2021. Frontiers in Signal Processing , 1:808395, January 2022

    2021

  20. [28]

    MoisesDB : A Dataset for Source Separation Beyond 4- Stems

    Igor Pereira, Felipe Ara \'u jo, Filip Korzeniowski, and Richard Vogl. MoisesDB : A Dataset for Source Separation Beyond 4- Stems . In Proceedings of the 24th International Society for Music Information Retrieval Conference ( ISMIR ) , pages 619--626, Milan, Italy, 2023

    2023

  21. [29]

    FiLM : Visual Reasoning with a General Conditioning Layer

    Ethan Perez, Florian Strub, Harm de Vries , Vincent Dumoulin, and Aaron Courville. FiLM : Visual Reasoning with a General Conditioning Layer . In Proceedings of the 32nd AAAI Conference on Artificial Intelligence , New Orleans, LA, USA, December 2017. AAAI

    2017

  22. [30]

    Hyperbolic Audio Source Separation

    Darius Petermann, Gordon Wichern, Aswin Subramanian, and Jonathan Le Roux. Hyperbolic Audio Source Separation . In Proceedings of the 2023 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , Rhodes Island, Greece, June 2023. IEEE

    2023

  23. [31]

    Hybrid Transformers for Music Source Separation

    Simon Rouard, Francisco Massa, and Alexandre D \'e fossez. Hybrid Transformers for Music Source Separation . In Proceedings of the 2023 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , Rhodes Island, Greece, May 2023. IEEE

    2023

  24. [32]

    Tim Salimans and Diederik P. Kingma. Weight normalization: A simple reparameterization to accelerate training of deep neural networks. In Proceedings of the 30th International Conference on Neural Information Processing Systems , NIPS '16, pages 901--909, Red Hook, NY, USA, De...

    2016

  25. [33]

    SDR - Medium Rare With Fast Computations

    Robin Scheibler. SDR - Medium Rare With Fast Computations . In Proceedings of the 2022 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , pages 701--705, Singapore, Singapore, 2022. IEEE

    2022

  26. [34]

    Conditioned Source Separation for Music Instrument Performances

    Olga Slizovskaia, Gloria Haro, and Emilia G \'o mez. Conditioned Source Separation for Music Instrument Performances . IEEE/ACM Transactions on Audio, Speech, and Language Processing , 29:2083--2095, 2021

    2021

  27. [35]

    Wave- U-Net : A Multi-Scale Neural Network for End-to-End Audio Source Separation

    Daniel Stoller, Sebastian Ewert, and Simon Dixon. Wave- U-Net : A Multi-Scale Neural Network for End-to-End Audio Source Separation . In Proceedings of the 19th International Society for Music Information Retrieval Conference ( ISMIR ) , pages 334--340, Paris, France, 2018. ISMIR

    2018

  28. [36]

    The 2018 Signal Separation Evaluation Campaign

    Fabian Robert St \"o ter, Antoine Liutkus, and Nobutaka Ito. The 2018 Signal Separation Evaluation Campaign . In Proceedings of the 14th International Conference on Latent Variable Analysis and Signal Separation , pages 293--305, Guildford, United Kingdom, 2018. Springer Inter...

    2018

  29. [37]

    Open- Unmix - A Reference Implementation for Music Source Separation

    Fabian-Robert St \"o ter, Stefan Uhlich, Antoine Liutkus, and Yuki Mitsufuji. Open- Unmix - A Reference Implementation for Music Source Separation . Journal of Open Source Software , 4(41):1667, September 2019

    2019

  30. [38]

    Multi- Scale multi-band densenets for audio source separation

    Naoya Takahashi and Yuki Mitsufuji. Multi- Scale multi-band densenets for audio source separation. In Proceedings of the 2017 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics ( WASPAA ) , pages 21--25, New Paltz, NY, October 2017. IEEE

    2017

  31. [39]

    D3Net : Densely connected multidilated DenseNet for music source separation, March 2021

    Naoya Takahashi and Yuki Mitsufuji. D3Net : Densely connected multidilated DenseNet for music source separation, March 2021

    2021

  32. [40]

    MMDenseLSTM : An Efficient Combination of Convolutional and Recurrent Neural Networks for Audio Source Separation

    Naoya Takahashi, Nabarun Goswami, and Yuki Mitsufuji. MMDenseLSTM : An Efficient Combination of Convolutional and Recurrent Neural Networks for Audio Source Separation . In Proceedings of the 16th International Workshop on Acoustic Signal Enhancement , pages 106--110, Tokyo, J...

    2018

  33. [41]

    Improving music source separation based on deep neural networks through data augmentation and network blending

    Stefan Uhlich, Marcello Porcu, Franck Giron, Michael Enenkl, Thomas Kemp, Naoya Takahashi, and Yuki Mitsufuji. Improving music source separation based on deep neural networks through data augmentation and network blending. In Proceedings of the 2017 IEEE International Conferen...

    2017

  34. [42]

    Emmanuel Vincent, Hiroshi Sawada, Pau Bofill, Shoji Makino, and Justinian P. Rosca. First stereo audio source separation evaluation campaign: Data , algorithms and results. In Proceedings of the 2007 International Conference on Independent Component Analysis and Signal Separat...

    2007

  35. [43]

    Emmanuel Vincent, Shoko Araki, Fabian Theis, Guido Nolte, Pau Bofill, Hiroshi Sawada, Alexey Ozerov, Vikrham Gowreesunker, Dominik Lutter, and Ngoc Q.K. Duong. The Signal Separation Evaluation Campaign (2007-2010): Achievements and Remaining Challenges . Signal Processing , 92...

    2007

  36. [44]

    Bittner, and Juan Pablo Bello

    Yu Wang, Daniel Stoller, Rachel M. Bittner, and Juan Pablo Bello. Few- Shot Musical Source Separation . In Proceedings of the 2022 IEEE International Conference on Acoustics , Speech and Signal Processing ( ICASSP ) , pages 121--125, Singapore, Singapore, May 2022. IEEE

    2022

  37. [45]

    Watcharasupat and Alexander Lerch

    Karn N. Watcharasupat and Alexander Lerch. A Stem-Agnostic Single-Decoder System for Music Source Separation Beyond Four Stems . In Proceedings of the 25th International Society for Music Information Retrieval Conference ( ISMIR ) , San Francisco, CA, USA, June 2024. ISMIR

    2024

  38. [46]

    Watcharasupat, Chih-Wei Wu, Yiwei Ding, Iroro Orife, Aaron J

    Karn N. Watcharasupat, Chih-Wei Wu, Yiwei Ding, Iroro Orife, Aaron J. Hipple, Phillip A. Williams, Scott Kramer, Alexander Lerch, and William Wolcott. A Generalized Bandsplit Neural Network for Cinematic Audio Source separation. IEEE Open Journal of Signal Processing , 5:73--81, 2023

    2023

  39. [47]

    Watcharasupat, Chih-Wei Wu, and Iroro Orife

    Karn N. Watcharasupat, Chih-Wei Wu, and Iroro Orife. Facing the Music : Tackling Singing Voice Separation in Cinematic Audio Source Separation . In Late- Breaking Demo Session of the 25th International Society for Music Information Retrieval Conference , San Francisco, CA, USA...

    2024

  40. [48]

    Implicit regularization and convergence for weight normalization

    Xiaoxia Wu, Edgar Dobriban, Tongzheng Ren, Shanshan Wu, Zhiyuan Li, Suriya Gunasekar, Rachel Ward, and Qiang Liu. Implicit regularization and convergence for weight normalization. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin, editors, Advances in Neural In...

    2020

  41. [49]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence '...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.