Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Direction-Aware Neural Acoustic Fields for Few-Shot Interpolation of Ambisonic Impulse Responses

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper proposes a direction-aware neural field (DANF) that learns spatially continuous first-order Ambisonic room impulse responses from finite measurements, adds an intensity-vector loss to capture direction of arrival, and shows that…

desk verdict A genuinely new application of neural fields to Ambisonic RIRs with a useful direction-aware loss, but the few-shot adaptation results rest on an early-stopping protocol that may leak the test set. read the letter →

arxiv 2505.13617 v1 pith:45PE5SIP submitted 2025-05-19 eess.AS cs.AIcs.CVcs.LGcs.SD

Christopher Ick , Gordon Wichern , Yoshiki Masuyama , François Germain , Jonathan Le Roux This is my paper · ORCID
classification eess.AScs.AIcs.CVcs.LGcs.SD
keywords spatialaudioneuralacousticfieldsAmbisonicsroomimpulseresponsesdirection-of-arrivalestimationintensityvectorfew-shotadaptationlow-rank
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

Sound in a room reaches a listener from many directions, but earlier neural-field models of room impulse responses treated sound as monaural or binaural only. This paper proposes DANF, a neural field that outputs first-order Ambisonic impulse responses, four channels that encode the pressure wave and its directional gradient, so the model can represent where sound comes from as well as what arrives. It adds an intensity-vector loss that compares the direction of energy flow in predicted and true responses; with a moderate weight this improves direction-of-arrival error without hurting standard acoustic measures, while an overly large weight sacrifices late reverberation. The paper further claims that a model pre-trained across many rooms can adapt to a new room from very few measurements, and that low-rank adaptation using under 1% of the parameters comes close to warm-start fine-tuning. If these claims hold, directional acoustic scene capture becomes practical in data-scarce settings such as virtual reality and telepresence.

What carries the argument

The central object is the first-order Ambisonic impulse response $h\in\mathbb{R}^{4\times T}$, whose four channels $w,x,y,z$ capture the omnidirectional pressure and the three Cartesian pressure-gradient components. The direction-aware machinery is the intensity vector $IV(U)(m,f)$ computed from the STFT of these channels; the paper averages it over time and frequency into a single 3D vector and applies a cosine-distance loss $L_{IV}$ between prediction and target. That loss supplies the training signal that ties the network's output to direction of arrival. The architecture follows prior neural acoustic fields: bounce points and source/listener relative positions are encoded into spatial features, multiplied with sinusoidal time encodings to form a spatio-temporal embedding, then decoded with an MLP conditioned on listener orientation. Low-rank adaptation is the few-shot mechanism: rank-$r$ matrices $B\in\mathbb{R}^{i\times r}$, $A\in\mathbb{R}^{r\times j}$ update each MLP weight as $W' = W + \frac{1}{r}BA$, so only a small parameter subset is trained for a new room.

What would settle it

Inspect the data splits: if the early-stopping holdout and the 200-point test set share indices, the reported warm-start and LoRA errors were selected on test data and are optimistic; re-running with a disjoint holdout and test set would settle whether the few-shot gains remain.

Watch

Extended reading notes

Core claim

On the paper's own terms, DANF is the first neural acoustic field to estimate Ambisonic-format room impulse responses, predicting $h\in\mathbb{R}^{4\times T}$ from source position, listener position and orientation, and a set of bounce points describing room geometry. The direction-aware component is an intensity-vector loss computed from the short-time Fourier transform of the four Ambisonic channels, averaged over time and frequency and compared to the target via cosine distance. Experiments on simulated rooms show this loss reduces direction-of-arrival error at all tested weights and, at $\lambda=10$, also improves T60, C50, and EDT metrics; at very large weights the model over-focuses on the direct sound and distorts late reverberation. For adaptation, warm-start fine-tuning and rank-3 LoRA both outperform cold-start and zero-shot baselines, with LoRA using $2.9\times 10^4$ parameters, less than 1% of the full model's $3.5\times 10^6$, while roughly matching warm-start performance.

Load-bearing premise

The few-shot comparison assumes the 200-sample holdout used for early stopping is a separate set from the 200-point test set, so the training epoch was not chosen using the test data's error.

Editorial extensions

If this is right

  • Ambisonic RIRs predicted by DANF can be convolved with dry sources to render directional sound, so spatial audio can be synthesized at arbitrary source-listener positions without new measurements.
  • The intensity-vector loss gives a single auxiliary term that improves both direction-of-arrival and conventional room-acoustic metrics at a moderate weight, potentially replacing separate DoA post-processing.
  • Pre-training across rooms plus LoRA adaptation makes per-room customization feasible with under 1% of the model parameters, which matters in memory-limited devices.
  • Because the model is continuous over source and listener locations, a single trained field can interpolate directional impulse responses across an entire room from a finite measurement grid.
  • Even a single training example yields better T60 and C50 errors than zero-shot transfer for warm-start and LoRA models, indicating background spatial knowledge transfers across rooms.

Reading between the lines

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

  • Editorial inference: the same intensity-vector loss could serve as a self-supervised objective for unlabeled Ambisonic recordings, since it is computed from the predicted and measured signals alone without needing room geometry.
  • Editorial inference: if the early-stopping holdout and test set are genuinely disjoint, the few-shot result is stronger than if they overlap; checking the split is the first test a reader should run.
  • Editorial inference: extending the loss to higher-order Ambisonics would test whether directional accuracy scales with spherical-harmonic order, likely requiring frequency-dependent weighting because higher-order channels are quieter.
  • Editorial inference: conditioning on bounce points suggests the model could be driven by estimated room geometry from depth sensors, enabling few-shot adaptation in real rooms rather than simulated ones.
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 / 5 minor

Summary. The paper proposes DANF, a direction-aware neural field that predicts first-order Ambisonic room impulse responses as a function of source position, listener position, listener orientation, and a room-geometry context encoded through bounce points. The main contributions are: (i) the first neural-field model for Ambisonic-format RIRs, (ii) an intensity-vector loss based on the cosine distance between averaged intensity vectors of the predicted and reference RIRs, and (iii) an empirical study of few-shot adaptation to novel rooms using warm-start fine-tuning, LoRA with rank 1 and 3, cold-start training, and zero-shot prediction. Experiments are carried out on Matterport3D rooms rendered with SoundSpaces. The reported results suggest that the intensity-vector loss substantially improves direction-of-arrival estimation, that a moderate loss weight can be beneficial, and that rank-3 LoRA with far fewer parameters is competitive with full warm-start fine-tuning in low-data regimes.

Significance. If the claims hold, the paper makes a useful step beyond monaural and binaural neural acoustic fields by modeling directional sound-field information in Ambisonic form. The intensity-vector loss is a simple and sensible inductive bias, and the DoA improvements in Table 2 are consistent across all ten rooms at λ=10. The few-shot adaptation study, especially the LoRA comparison, is practically relevant for deploying neural acoustic fields in new environments. The paper does not release code or trained models, and all results are based on simulated data, so reproducibility rests entirely on the manuscript details. The explicit strengths are the architectural adaptation of INRAS to Ambisonic output, the novel IV loss, and the low-rank adaptation comparison with parameter counts.

major comments (4)
  1. [§4.3 (Table 1)] The early-stopping protocol for low-data warm-start fine-tuning is ambiguous and potentially test-set-contaminating. Section 4.3 states that each target environment C' has 200 points in its test set and that the Warm-Start models with fewer than 80 training examples were early-stopped on a 200-sample holdout set, but the paper nowhere states that this holdout is disjoint from the test set. If the same 200 points are used both for early stopping and for final evaluation, then the training epoch is selected using test data, and the Warm-Start errors (and any comparison against them, including the LoRA equivalence claim in Table 1) are optimistically biased. Please state explicitly whether the holdout is disjoint from the test set; if it is not, rerun the affected experiments with a separate validation split and update Table 1 and the associated claims.
  2. [§5.1 (Table 2)] The claim that the intensity-vector loss at λ=10 improves 'both room acoustic metrics as well as directional performance' is not supported by the per-room results in Table 2. Averaging over the ten rooms, T60 error drops from 1.59% to 0.99%, but C50 error increases from 2.58 dB to 2.80 dB and EDT error increases from 56.8 ms to 59.6 ms; only 5 of 10 rooms improve in C50 and 6 of 10 improve in EDT. The direction-of-arrival improvement is clear and consistent, but the claim of simultaneous acoustic-metric gains at λ=10 should either be removed or qualified, unless Figure 3, which is described as showing the same single-room setup, uses a different aggregation or different data. Please reconcile Table 2 with the text and Figure 3.
  3. [§5.1 (Fig. 3)] The λ-sweep results are presented as point curves without error bars, repeated runs, or significance tests, despite the sweep being used to draw the load-bearing conclusion that λ=10 is a sweet spot. With only ten rooms and a single training run per configuration, the apparent improvements at λ=10 in T60, C50, and EDT could be driven by one or two rooms, and the per-room Table 2 already shows non-uniform behavior. Please provide per-room variance or significance information, or at minimum state the number of runs per condition, to support the 'improves both room acoustic metrics' conclusion.
  4. [§4.3, §5.2 (Table 1)] The few-shot comparisons in Table 1 are based on single runs and do not report the pretraining data, the number of rooms used for pretraining, the architecture hyperparameters, or the variability across training examples. Since the 1-example and 80-example regimes are highly sensitive to initialization and example selection, the equivalence between Warm-Start and LoRA(3) should be supported by multiple fine-tuning seeds (or bootstrap over training subsets). Without this, the 'near equivalent performance' claim in Section 5.2 is not yet quantitatively established.
minor comments (5)
  1. [Eq. (5)] The definition of the spectral convergence term L_sc appears to contain a typo: the numerator should be |||H_hat| - |H|||_2, with a hat on the predicted magnitude, not |||H| - |H|||_2.
  2. [Fig. 3 caption] The x-axis label reads 'log10 10', which appears to be a formatting artifact; it should likely read 'log10 λ'.
  3. [§4.3] There is a typo in 'matricies' (should be 'matrices'), and the author list formatting in the title page shows an unusual space in 'Franc ¸ois'.
  4. [§4.1] The dataset description says 100 unique rooms are available from five buildings, but the single-environment experiments use only 10 rooms with no explanation of how those rooms were selected; please state the selection criterion and list or characterize the room volumes.
  5. [§2] Equation (1) uses W, X, Y, Z for the STFT components but the preceding definition uses lowercase w, x, y, z for the FOA channels; please add a sentence clarifying the correspondence.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the intensity-vector loss is a supervised objective compared against ground-truth RIRs, and the few-shot adaptation claims rest on held-out evaluations rather than on fitted inputs or self-cited uniqueness theorems.

full rationale

The paper contains no derivation-from-input circularity. The proposed direction-aware loss in Eq. (6) is a supervised term comparing the intensity vector of the predicted RIR, I(hat h), with the intensity vector of the ground-truth RIR, I(h); it is not a fitted parameter renamed as a prediction. Although the DoA metric is related to the same intensity-vector representation, the model must still generalize to held-out 200-point test sets, and the reported improvements in T60, C50, and EDT at lambda=10 are not the same quantity as the cosine-distance loss, so those findings are not forced by construction. The architecture follows the external INRAS work [17] and the adaptation protocol follows [20], with no load-bearing self-citation and no uniqueness theorem imported from the authors. The early-stopping holdout described in Section 4.3 is not explicitly stated to be disjoint from the 200-point test set; if it were the same set, the Warm-Start few-shot numbers would be optimistically biased, and the lambda sweep in Figure 3 is test-set model selection. However, these are experimental-validity and statistical concerns, not circular reductions, and the paper does not state the identity that would be needed to make them a fitted-input-called-prediction circularity. Overall, the central contributions are empirical and architectural, and they do not reduce to their inputs by definition.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new physical entities, forces, or particles. Its free parameters are the loss weight λ, the LoRA rank, and unstated architecture dimensions. The main domain assumptions are that simulated RIRs are trustworthy, the averaged intensity vector tracks direct-sound DoA, and bounce points capture the environment.

free parameters (3)
  • λ (IV loss weight) = 10
    Swept over {10^γ | γ=-4,...,4}; λ=10 is reported as best in Fig. 3 based on test-set metrics, so it is a hyperparameter selected on the test set.
  • LoRA rank r = r=3 and r=1
    Two ranks are evaluated without justification or a selection rule; the choice affects few-shot performance.
  • Architecture dimensions (D, K, MLP widths) = not reported
    Spatial embedding dimension, number of bounce points, and MLP sizes are required to reimplement but are not stated.
assumptions (3)
  • domain assumption SoundSpaces-simulated Ambisonic RIRs are treated as ground truth.
    All training and evaluation rely on simulated RIRs from SoundSpaces 2.0, with no validation against measured impulse responses.
  • domain assumption The time-frequency averaged intensity vector represents direction-of-arrival (direct sound).
    The IV loss assumes averaging the intensity vector over all time and frequency corresponds to the direct-sound DoA, citing [24,25,26]. This may be poor in highly reverberant rooms.
  • domain assumption Uniformly sampled bounce points from the room mesh fully characterize the acoustic environment.
    Borrowed from INRAS [17]; room material properties and finer geometry are not modeled except through learned features derived from these points.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Direction-Aware Neural Acoustic Fields for Few-Shot Interpolation of Ambisonic Impulse Responses." pith.science (2026). https://pith.science/paper/45PE5SIP

@misc{pith2026250513617,
  author       = {Pith},
  title        = {Pith review of: Direction-Aware Neural Acoustic Fields for Few-Shot Interpolation of Ambisonic Impulse Responses},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/45PE5SIP}},
  note         = {Machine review of arXiv:2505.13617}
}
read the original abstract

The characteristics of a sound field are intrinsically linked to the geometric and spatial properties of the environment surrounding a sound source and a listener. The physics of sound propagation is captured in a time-domain signal known as a room impulse response (RIR). Prior work using neural fields (NFs) has allowed learning spatially-continuous representations of RIRs from finite RIR measurements. However, previous NF-based methods have focused on monaural omnidirectional or at most binaural listeners, which does not precisely capture the directional characteristics of a real sound field at a single point. We propose a direction-aware neural field (DANF) that more explicitly incorporates the directional information by Ambisonic-format RIRs. While DANF inherently captures spatial relations between sources and listeners, we further propose a direction-aware loss. In addition, we investigate the ability of DANF to adapt to new rooms in various ways including low-rank adaptation.

Figures

Figures reproduced from arXiv: 2505.13617 by the authors.

Figure 1
Figure 1. Overview of the proposed DANF framework. The input is the environmental context in the form of K bounce points ck ∈ R 3 , as well as the source and listener locations s, l ∈ R 3 , all of which are concatenated into spatial feature C˜ (denoted by operator ⊕) used to create a spatial-temporal encoding E ∈ R 3K×T . For the impulse generation module, the spatio-temporal embedding E and the orientation of the listener θ … view at source ↗
Figure 2
Figure 2. Single-room performance for different values of λ (row) and metrics (columns). Each metric is normalized to [0, 1] for ease of visualization. We initially pre-train a model over a wide range of environ￾ments Cpretrain. We then freeze all trainable parameters of the model and establish a new environment C ′ ∈/ Cpretrain with a corresponding set of training and test data. Each environment C ′ has 200 points in its tes… view at source ↗
Figure 3
Figure 3. Model performance on single rooms with varying in￾tensity vector loss weight λ, with the dashed line representing no intensity vector loss, or λ = 0. showed a better correlation in the azimuth of the sound source. The larger weight, λ = 104 , further improved the DoA perfor￾mance but deteriorated the distribution of other acoustic mea￾sures. This result suggests that the predicted RIRs focus on the direct sound and … view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Low-Rank Adaptation of Deep Prior Neural Networks For Room Impulse Response Reconstruction

    eess.AS 2025-07 conditional novelty 5.0 of 10

    LoRA-adapted Deep Prior networks reconstruct room impulse responses as accurately as full fine-tuning while training only about 30% of the network parameters.

Reference graph

Works this paper leans on

35 extracted references · 31 canonical work pages · cited by 1 Pith paper

  1. [1]

    Introduction Inferring how an acoustic signal emitted from a sound source is received by a listener involves modeling the physically complex propagation of that signal through their shared environment [1]. Even when the source and listener are in direct line-of-sight, re- flections of the sound signal at various surfaces and objects add significant contri...

    work page

  2. [2]

    It is typ- ically defined for given locations of the source and listener in a particular acoustic environment [1]

    Ambisonics Room Impulse Responses The transfer function characterizing the physical process of sound propagation is commonly referred to as RIR. It is typ- ically defined for given locations of the source and listener in a particular acoustic environment [1]. As an RIR can be treated as a linear filter, by convolving it with an anechoic sound source, we c...

    arXiv 2025

  3. [3]

    bounce points

    Methods 3.1. Neural Acoustic Fields Prior approaches in neural acoustic fields establish the problem of estimating an RIRh∈R N×T [17–19]. Prior work has fo- cused on the monaural or binaural cases, and while some have allowed for different listener orientationsθ∈[0,2π). Their training methods and evaluations did not consider inter-channel directional feat...

    work page

  4. [4]

    Dataset We utilize Matterport3D-RGB [27] to define the acoustic en- vironment for each of our rooms

    Experiments 4.1. Dataset We utilize Matterport3D-RGB [27] to define the acoustic en- vironment for each of our rooms. Because each Matterport3D building incorporates multiple connected rooms, we utilize the region metadata to slice each building into several individual rooms. We sample 5 buildings, giving us 100 unique rooms at varying volumes. Within eac...

    work page

  5. [5]

    bounce points

    Results 5.1. Spatial Improvements We can see that the IV loss significantly improves DANF’s abil- ity to accurately reconstruct the directional characteristics of Ambisonic-RIRs. Figure 2 shows the correlations between the acoustic measures of the oracle and the rendered RIRs. We can see the effect of adding the IV loss as the estimates of the DoA azimuth...

    work page

  6. [6]

    To our knowledge, DANF is the first NF model estimating Ambisonic-format RIRs

    Conclusion In this work, we presented DANF, a novel neural acoustic field model for estimating direction-aware RIRs. To our knowledge, DANF is the first NF model estimating Ambisonic-format RIRs. By utilizing a directional intensity vector loss, DANF is able to accurately capture not only the acoustic properties of the envi- ronment, but also the directio...

    work page

  7. [7]

    Roginska and P

    A. Roginska and P. Geluso,Immersive sound: The Art and Sci- ence of Binaural and Multi-channel Audio, 1st ed. Focal Press, 2018

    work page 2018

  8. [8]

    Binaural ren- dering of ambisonic signals by head-related impulse response time alignment and a diffuseness constraint,

    M. Zaunschirm, C. Sch ¨orkhuber, and R. H¨oldrich, “Binaural ren- dering of ambisonic signals by head-related impulse response time alignment and a diffuseness constraint,”The Journal of the Acoustical Society of America, vol. 143, no. 6, pp. 3616–3627, 06 2018

    work page 2018

Show all 35 references
  1. [9]

    Towards generating am- bisonics using audio-visual cue for virtual reality,

    A. Rana, C. Ozcinar, and A. Smolic, “Towards generating am- bisonics using audio-visual cue for virtual reality,” inProc. ICASSP, 2019, pp. 2012–2016

    2019

  2. [10]

    Image method for efficiently simulat- ing small-room acoustics,

    J. B. Allen and D. Berkley, “Image method for efficiently simulat- ing small-room acoustics,”J. Acoust. Soc. Am., vol. 65, no. 4, pp. 943–950, 04 1979

    1979

  3. [11]

    Pyroomacoustics: A python package for audio room simulation and array processing algorithms,

    R. Scheibler, E. Bezzam, and I. Dokmanic, “Pyroomacoustics: A python package for audio room simulation and array processing algorithms,” inProc. ICASSP, Apr. 2018, p. 351–355

    2018

  4. [12]

    SoundSpaces: Audio- visual navigaton in 3D environments,

    C. Chen, U. Jain, C. Schissler, S. V . A. Gari, Z. Al-Halah, V . K. Ithapu, P. Robinson, and K. Grauman, “SoundSpaces: Audio- visual navigaton in 3D environments,” inProc. ECCV, 2020

    2020

  5. [13]

    A review of finite-element methods for time- harmonic acoustics,

    L. L. Thompson, “A review of finite-element methods for time- harmonic acoustics,”J. Acoust. Soc. Am., vol. 119, no. 3, pp. 1315–1330, 03 2006

    2006

  6. [14]

    Source and lis- tener directivity for interactive wave-based sound propagation,

    R. Mehra, L. Antani, S. Kim, and D. Manocha, “Source and lis- tener directivity for interactive wave-based sound propagation,” IEEE Trans. Vis. Comput. Graph., vol. 20, no. 4, pp. 495–503, 2014

    2014

  7. [15]

    An efficient GPU-based time domain solver for the acoustic wave equation,

    R. Mehra, N. Raghuvanshi, L. Savioja, M. C. Lin, and D. Manocha, “An efficient GPU-based time domain solver for the acoustic wave equation,”Appl. Acoust., vol. 73, no. 2, pp. 83–94, 2012

    2012

  8. [16]

    Analysis of mutiple-shepers radiation and scattering problems by using a null-field integral equation approach,

    J.-T. Chen, Y .-T. Lee, and Y .-J. Lin, “Analysis of mutiple-shepers radiation and scattering problems by using a null-field integral equation approach,”Appl. Acoust., vol. 71, no. 8, pp. 690–700, 2010

    2010

  9. [17]

    Parametric wave field coding for precomputed sound propagation,

    N. Raghuvanshi and J. Snyder, “Parametric wave field coding for precomputed sound propagation,”ACM Trans. Graph., vol. 33, no. 4, Jul. 2014

    2014

  10. [18]

    Low frequency inter- polation of room impulse responses using compressed sensing,

    R. Mignot, G. Chardon, and L. Daudet, “Low frequency inter- polation of room impulse responses using compressed sensing,” IEEE/ACM Trans. Audio, Speech, Lang. Process., vol. 22, no. 1, pp. 205–216, 2014

    2014

  11. [19]

    Parametric directional coding for precomputed sound propagation,

    N. Raghuvanshi and J. Snyder, “Parametric directional coding for precomputed sound propagation,”ACM Trans. Graph., vol. 37, no. 4, Jul. 2018

    2018

  12. [20]

    Spatial extrapolation of early room impulse responses in local area using sparse equivalent sources and image source method,

    I. Tsunokuni, K. Kurokawa, H. Matsuhashi, Y . Ikeda, and N. Os- aka, “Spatial extrapolation of early room impulse responses in local area using sparse equivalent sources and image source method,”Appl. Acoust., vol. 179, p. 108027, 2021

    2021

  13. [21]

    NeRF: Representing scenes as neural ra- diance fields for view synthesis,

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ra- mamoorthi, and R. Ng, “NeRF: Representing scenes as neural ra- diance fields for view synthesis,” inProc. ECCV, 2020

    2020

  14. [22]

    Learning neural acoustic fields,

    A. Luo, Y . Du, M. Tarr, J. Tenenbaum, A. Torralba, and C. Gan, “Learning neural acoustic fields,” inProc. NeurIPS, vol. 35, 2022, pp. 3165–3177

    2022

  15. [23]

    INRAS: Implicit neural rep- resentation for audio scenes,

    K. Su, M. Chen, and E. Shlizerman, “INRAS: Implicit neural rep- resentation for audio scenes,” inProc. NeurIPS, vol. 35, 2022, pp. 8144–8158

    2022

  16. [24]

    A V-NeRF: Learning neural fields for real-world audio-visual scene synthe- sis,

    S. Liang, C. Huang, Y . Tian, A. Kumar, and C. Xu, “A V-NeRF: Learning neural fields for real-world audio-visual scene synthe- sis,” inProc. NeurIPS, 2023

    2023

  17. [25]

    Neural acoustic context field: Rendering realistic room impulse response with neural fields,

    ——, “Neural acoustic context field: Rendering realistic room impulse response with neural fields,” inProc. ICCV, 2023

    2023

  18. [26]

    Real acoustic fields: An audio-visual room acoustics dataset and benchmark,

    Z. Chen, I. D. Gebru, C. Richardt, A. Kumar, W. Laney, A. Owens, and A. Richard, “Real acoustic fields: An audio-visual room acoustics dataset and benchmark,” inProc. CVPR, 2024

    2024

  19. [27]

    Implicit neural represen- tation with physics-informed neural networks for the reconstruc- tion of the early part of room impulse responses,

    M. Pezzoli, F. Antonacci, and A. Sarti, “Implicit neural represen- tation with physics-informed neural networks for the reconstruc- tion of the early part of room impulse responses,”arXiv preprint arXiv:2306.11509, 2023

    2023 arXiv

  20. [28]

    Room impulse response reconstruction with physics-informed deep learning,

    X. Karakonstantis, D. Caviedes-Nozal, A. Richard, and E. Fernandez-Grande, “Room impulse response reconstruction with physics-informed deep learning,”J. Acoust. Soc. Am., vol. 155, no. 2, pp. 1048–1059, 2024

    2024

  21. [29]

    Physics-informed machine learning for sound field estima- tion: Fundamentals, state of the art, and challenges,

    S. Koyama, J. G. C. Ribeiro, T. Nakamura, N. Ueno, and M. Pez- zoli, “Physics-informed machine learning for sound field estima- tion: Fundamentals, state of the art, and challenges,”IEEE Signal Process. Mag., vol. 41, no. 6, pp. 60–71, 2024

    2024

  22. [30]

    CRNN-based joint azimuth and elevation localization with the ambisonics in- tensity vector,

    L. Perotin, R. Serizel, E. Vincent, and A. Gu ´erin, “CRNN-based joint azimuth and elevation localization with the ambisonics in- tensity vector,” inProc. IWAENC, 2018, pp. 241–245

    2018

  23. [31]

    Three-dimensional sound source localization using B-format signals,

    H. Khaddour, J. Schimmel, and M. Trzos, “Three-dimensional sound source localization using B-format signals,”IJATES 2, vol. 2, Mar. 2013

    2013

  24. [32]

    Im- proved localization of sound sources using multi-band processing of ambisonic components,

    D. Charalampos, K. George, A. Konstantinos, and P. George, “Im- proved localization of sound sources using multi-band processing of ambisonic components,”J. Audio Eng. Soc., no. 7691, May 2009

    2009

  25. [33]

    Matterport3D: Learn- ing from RGB-D data in indoor environments,

    A. Chang, A. Dai, T. Funkhouser, M. Halber, M. Niessner, M. Savva, S. Song, A. Zeng, and Y . Zhang, “Matterport3D: Learn- ing from RGB-D data in indoor environments,” inProc. 3DV, 2017

    2017

  26. [34]

    SoundSpaces 2.0: A simulation platform for visual-acoustic learning,

    C. Chen, C. Schissler, S. Garg, P. Kobernik, A. Clegg, P. Calamia, D. Batra, P. W. Robinson, and K. Grauman, “SoundSpaces 2.0: A simulation platform for visual-acoustic learning,” inProc. NeurIPS Datasets and Benchmarks Track, 2022

    2022

  27. [35]

    LoRA: Low-rank adaptation of large lan- guage models,

    E. J. Hu, yelong shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen, “LoRA: Low-rank adaptation of large lan- guage models,” inProc. ICLR, 2022

    2022

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.