Ambisonics Encoder for Wearable Array with Improved Binaural Reproduction

Boaz Rafaely; David Alon; Vladimir Tourbabin; Yhonatan Gayer; Zamir Ben-Hur

REVIEW 3 major objections 5 minor 21 references

Ambisonics Encoder for Wearable Array with Improved Binaural Reproduction

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Wearable microphone arrays can encode Ambisonics tuned for binaural playback via one mixing parameter.

desk verdict The paper's central closed-form joint solution is mathematically wrong; the useful piece is the Sec. 3.1 formulation, but the headline result reduces to an unprincipled average of two existing filters. read the letter →

arxiv 2507.04108 v1 pith:P7O2YY2J submitted 2025-07-05 eess.AS eess.SP

Yhonatan Gayer , Vladimir Tourbabin , Zamir Ben-Hur , David Alon , Boaz Rafaely This is my paper

classification eess.ASeess.SP

keywords ambisonicsbinauralreproductionwearablemicrophonearraySignalMatchinghead-relatedtransferfunctionaugmentedrealityTikhonovregularization

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

This paper tries to establish that an Ambisonics encoder for a non-ideal, head-worn microphone array can be designed to serve binaural reproduction rather than only Ambisonics channel accuracy. The proposed joint loss adds a Binaural Signal Matching error to the usual Ambisonics Signal Matching error, and the paper claims the optimal encoder is a closed-form weighted average of the two separately optimal filters, with a single parameter $\alpha$ controlling the trade-off. If correct, this gives wearable VR/AR devices a signal-independent, standard-format Ambisonics encoder whose output, once rendered with head-related transfer functions, reproduces binaural cues more faithfully without adding extra channels. The paper supports the claim with a simulation study of a five-microphone array mounted on a rigid sphere.

What carries the argument

The key object is the joint ASM-BSM filter matrix of Eq. (23). The argument is made tractable by a vectorization step: the ASM filter matrix is flattened into a long vector $c^{\mathrm{flat}}_{nm}$, and the binaural rendering path is written with a block-diagonal HRTF matrix $H^{l,r}_{nm}$ (Eqs. 16-18), so that the binaural error becomes a quadratic function of the same vector. Each sub-problem then has a Tikhonov-regularized closed-form solution, and the paper claims the joint solution is the $\alpha$-weighted sum of those two solutions.

What would settle it

For the same five-microphone array, HRTFs, noise level, and $\alpha = 0.5$, solve Eq. (22) directly by numerical least squares over the vectorized filter coefficients and compare the result with the right-hand side of Eq. (23); any material difference in the filter entries or in the resulting ASM and BSM errors would falsify the central formula. A single-frequency check, e.g., at 1 kHz, would already be informative.

Watch

Extended reading notes

Core claim

The central claim is that the minimizer of the combined loss in Eq. (22), a weighted sum of the Ambisonics reconstruction error over all channels and the binaural reproduction error, is exactly the same weighted linear combination of the two individually optimal filter matrices: $C^{\mathrm{joint}}_{nm} = \alpha C^{\mathrm{ASM}}_{nm} + (1-\alpha) C^{\mathrm{BSM}}_{nm}$ (Eq. 23). The paper therefore asserts that a single parameter $\alpha$ sweeps continuously from pure Ambisonics matching to pure binaural matching. In simulation with five microphones arranged on a rigid sphere like a glasses-mounted array, the balanced choice $\alpha = 0.5$ lowers binaural error substantially relative to pure ASM while keeping Ambisonics error closer to the pure-ASM case than pure BSM does.

Load-bearing premise

The load-bearing premise is that the minimizer of the weighted sum of the ASM and BSM errors equals the weighted sum of the two individually optimal filters, a step taken between Eqs. (22) and (23); that identity is not generally true for least-squares problems with different operators and is not proven in the paper.

Editorial extensions

If this is right

With Eq. (23), a wearable-array encoder can position itself anywhere on the ASM-BSM trade-off by choosing one scalar $\alpha$.
At $\alpha = 0.5$, the simulated head-worn array achieves binaural errors close to direct BSM matching while improving Ambisonics channel accuracy over pure BSM, so the method fits into standard Ambisonics-based playback pipelines.
Because the encoded format remains standard Ambisonics, binaural rendering still uses ordinary HRTF decoders, and no extra residual channels or HRTF-specific transmission format is required.
The closed-form solution avoids iterative optimization: filters are computed directly from the array geometry, the HRTFs, and the noise-to-signal ratio.

Reading between the lines

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

The linear-combination step in Eq. (23) is asserted rather than proven; for general quadratic least-squares problems the minimizer of a weighted sum is not the weighted sum of minimizers, so a direct numerical check of Eq. (23) is the natural next test.
If that step fails, the plotted $\alpha = 0.5$ curve would not be the true optimum of the joint loss, and the reported trade-off could change under a numerically optimized joint filter.
The same vectorized formulation extends naturally to higher Ambisonics orders, multiple listener head orientations, or individualized HRTFs, because each addition contributes another quadratic term to the same loss.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Here's my read. The paper targets a real problem: for low-order Ambisonics from a wearable array, the standard ASM encoder is far from optimal for binaural playback. The idea of adding a binaural-matching term to the encoder loss is worth pursuing.

The useful contribution is Sec. 3.1, where the authors express the binaural error of the ASM encoder as a function of the encoder coefficients, vectorize it, and derive a BSM-constrained encoder via Tikhonov. That's clean and novel enough to be a building block.

But the central claim, Eq. (23), is false. The minimizer of a weighted sum of two least-squares costs is not generally the weighted sum of the minimizers. The two operators here (steering V and the BSM operator involving H_l^nm V) are different, so no special cancellation applies. The scalar counterexample in the stress-test note (v=1, y=2, h=1, α=0.5) is enough: Eq. (23) predicts 1.5, the true optimum is 1.2. The paper asserts (23) without proof, and I don't think a proof is possible under the stated assumptions.

This makes the experiments in Sec. 4 less informative than they look. They evaluate the interpolated filter, not the minimizer of (22). The trade-off in Fig. 2 is exactly what you'd get from any convex combination of the ASM and BSM filters; it doesn't validate the joint optimization. So the headline result reduces to a weighted average of two existing filters, and the only real novelty left is the Sec. 3.1 derivation.

That said, the paper is well-organized, the experimental setup is sensible, and the authors are clearly familiar with the literature. The flaw is serious but localized. I'd send this to peer review rather than desk-reject, because the underlying problem is important and the Sec. 3.1 machinery could support a corrected version. The reviewer should demand a correct derivation of the joint minimizer (or a proof of the interpolation property, which I doubt exists) and experiments comparing the true optima against the interpolated filter.

Recommendation: engage with it, but not as-is. I wouldn't cite it until the joint solution is fixed and validated.

Referee Report

3 major / 5 minor

Summary. The paper proposes a method for designing a first-order Ambisonics encoder for a wearable microphone array by jointly minimizing an Ambisonics signal matching (ASM) error and a binaural signal matching (BSM) error. The central claim, stated in Eq. (23), is that the minimizer of the weighted sum in Eq. (22) is the weighted sum of the individual minimizers in Eqs. (8) and (19), leading to a closed-form filter C^joint_nm = α C^ASM_nm + (1−α) C^BSM_nm. The paper then presents a simulation study with a five-microphone array on a rigid sphere, showing error trade-offs for α = 1, 0, and 0.5, and concludes that the joint design enables lower binaural error at the cost of higher Ambisonics error.

Significance. Should Eq. (23) hold, the paper would offer a simple, parameter-free closed-form encoder with a tunable trade-off, which would be a useful contribution to binaural rendering from head-worn arrays. The paper also carefully formulates ASM and BSM in a common framework and uses clearly defined error measures. However, the key mathematical claim is incorrect: the minimizer of a weighted sum of two least-squares problems with different operators is not generally the weighted sum of the minimizers. Since the experiments evaluate exactly the interpolation formula from Eq. (23) rather than the solution of the joint loss, the reported results do not support the proposed optimization. The contribution therefore reduces to a heuristic convex combination of existing filters, which does not deliver the claimed joint optimization.

major comments (3)

[Section 3.2, Eq. (23)] The statement that the solution to (22) is a linear combination of the solutions to (8) and (19) is not true in general. For quadratic costs with different linear operators, the minimizer of α||A_1 c − b_1||^2 + (1−α)||A_2 c − b_2||^2 is (α A_1^H A_1 + (1−α) A_2^H A_2)^{-1} (α A_1^H b_1 + (1−α) A_2^H b_2), which does not equal α (A_1^H A_1)^{-1} A_1^H b_1 + (1−α) (A_2^H A_2)^{-1} A_2^H b_2 in general. A simple scalar counterexample: minimize 0.5||c−2||^2 + 0.5||2c−2||^2; the individual minimizers are 2 and 1, but the joint minimizer is c=1.2, not 1.5. Thus Eq. (23) is false as written, and the claimed 'closed-form solution' to the joint problem is unsupported. Since this is the central derivation, the paper's core contribution fails.
[Section 4, Eqs. (24)-(25)] The experimental evaluation computes the errors for the filter from Eq. (23) only, i.e., the interpolated average of the ASM and BSM filters. It never computes the actual minimizer of the joint loss (22). The observed trade-off between ξ_ASM and ξ_BSM as α varies is therefore a property of the ad-hoc average, not a validation that joint optimization improves binaural reproduction. The conclusion that 'joint ASM-BSM design enables the optimization of ASM filters that are also tailored for BSM' is not supported by the data.
[Section 3.2, Eq. (22)] The objective in Eq. (22) is not well-defined as written. ε_BSM_nm is introduced in Eq. (19) with a subscript nm, but the expression on the right-hand side of (19) is independent of n and m; it is a single binaural error for the full vector c_nm,flat. Summing this same quantity over all n,m in (22) weights the binaural term by (N_a+1)^2 or is otherwise inconsistent. The authors should define the joint loss explicitly with the proper weighting and clarify whether the BSM term is one global error or a per-channel decomposition.

minor comments (5)

[Section 2.1] The phrase 'ri is a radius in meters ∀1 ≤ i ≤ Q' should read '∀1 ≤ i ≤ M'.
[Equation (19)] There are malformed norm expressions such as '∥hl,r]T ∥' that should be corrected to ∥hl,r∥^2.
[Throughout] Several typographical errors need correction: 'Ambisoniocs', 'Binural', 'be can be generalized', 'the from of a long vector', 'rearenged'. A careful proofread is required.
[Introduction / Conclusion] The paper motivates the work by comparison to residual channels in [9], but no comparison or discussion with that approach is included in the experiments.
[Section 4.4] Figure 2 is referenced but not included in the submitted text; the results are described qualitatively without numerical values or confidence intervals.

Circularity Check

1 steps flagged · score 6.0 of 10

The 'joint ASM-BSM solution' in Eq. (23) is not derived from the joint loss; it is a convex combination of the two known filters, so the reported trade-off evaluates that combination rather than the claimed minimization.

renaming known result [Section 3.2, Eqs. (22)-(23); used in Section 4.3, Eq. (25)]
"The solution to the error in (22) is a linear combination of the solutions to (15) and (8), expressed as: Cjoint_nm = α CASM_nm + (1 − α) CBSM_nm (23) where Cjoint_nm represents the joint ASM-BSM filter."

Equation (22) defines the joint objective to be minimized; Equation (23) simply asserts that the minimizer is α times the ASM filter plus (1−α) times the BSM filter. No minimization is performed: Cjoint is constructed as a weighted average of the two already-known filters given in Eq. (10) and Eq. (20). For sums of least-squares terms with different operators this equality is not generally true, so the 'solution' is not a consequence of Eq. (22). In Section 4.3 the performance measures explicitly use 'Cjoint_nm is computed as in Eq. (23)', so the experimental trade-off between ASM error and BSM error is a property of that convex combination, i.e., of the input filters, not a validation that Eq. (22) was minimized.

full rationale

The paper contains no data fitting: α is a user-chosen weight and the evaluation uses independent normalized-error measures, so the usual fitted-input circularity does not arise. The central issue is structural: the claimed 'solution' in Eq. (23) is a convex combination of the existing ASM and BSM filters, presented as if it were derived from the joint loss in Eq. (22). Because Section 4 then evaluates exactly that combination, the reported behavior—lower ASM error with higher α and lower BSM error with lower α—is built into the construction rather than being a consequence of solving the joint optimization. This makes the main 'joint ASM-BSM design' a renaming of a weighted interpolation rather than a derived minimizer. The self-citations, including [9], are motivational and background, not load-bearing for the derivation. No other circularity was found.

Assumptions & free parameters 1 free parameters · 3 assumptions · 0 invented entities

The central claim rests on an unproved and mathematically false linearity assumption about least-squares solutions, plus standard diffuse-field and rigid-sphere modeling assumptions.

free parameters (1)

alpha (balancing factor) = 0, 0.5, 1 (hand-chosen)
Controls the trade-off between ASM and BSM objectives in Eq. (22). Values are hand-picked for the experiments, not fitted to data.

assumptions (3)

ad hoc to paper The minimizer of a weighted sum of two quadratic error functions is the weighted sum of the individual minimizers (Eqs. 22 to 23).
This unproved and generally false assertion is the basis for the closed-form solution in Eq. (23). It is not a standard least-squares property.
domain assumption Diffuse sound field and white, uncorrelated microphone noise (Rs = sigma_s^2 I, Rn = sigma_n^2 I).
Used in deriving the signal-independent filters in Secs. 2.3, 2.4, and the proposed method.
domain assumption Rigid-sphere head model with microphones and ears on the surface approximates a wearable array and real HRTFs.
The simulation in Sec. 4 relies on this model; real HRTFs and head rotations are left to future work.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Ambisonics Encoder for Wearable Array with Improved Binaural Reproduction." pith.science (2026). https://pith.science/paper/P7O2YY2J

@misc{pith2026250704108,
  author       = {Pith},
  title        = {Pith review of: Ambisonics Encoder for Wearable Array with Improved Binaural Reproduction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P7O2YY2J}},
  note         = {Machine review of arXiv:2507.04108}
}

read the original abstract

Ambisonics Signal Matching (ASM) is a recently proposed signal-independent approach to encoding Ambisonic signal from wearable microphone arrays, enabling efficient and standardized spatial sound reproduction. However, reproduction accuracy is currently limited due to the non-ideal layout of the microphones. This research introduces an enhanced ASM encoder that reformulates the loss function by integrating a Binaural Signal Matching (BSM) term into the optimization framework. The aim of this reformulation is to improve the accuracy of binaural reproduction when integrating the Ambisonic signal with Head-Related Transfer Functions (HRTFs), making the encoded Ambisonic signal better suited for binaural reproduction. This paper first presents the mathematical formulation developed to align the ASM and BSM objectives in a single loss function, followed by a simulation study with a simulated microphone array mounted on a rigid sphere representing a head-mounted wearable array. The analysis shows that improved binaural reproduction with the encoded Ambisonic signal can be achieved using this joint ASM-BSM optimization, thereby enabling higher-quality binaural playback for virtual and augmented reality applications based on Ambisonics.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

21 extracted references · 19 canonical work pages

[1]

Ambisonics Encoder for Wearable Array with Improved Binaural Reproduction

INTRODUCTION Binaural reproduction is essential for creating immer- sive auditory experiences in virtual reality (VR) and aug- mented reality (AR), enabling spatial perception by en- abling real-world acoustic cues [1]. Ambisonics [2] has emerged as a powerful method for spatial audio represen- tation due to its ability to encode sound fields independent ...

work page Pith review arXiv 2025
[2]

2.1 Signal Model Consider an array consisting ofM omnidirectional micro- phones, where each microphone is located at spherical co- ordinates (ri, θi, ϕi) for 1 ≤ i ≤ M

BACKGROUND This section provides the mathematical foundation for the signal model, presents the Ambisonics representation de- rived from this model, and outlines the formulations for encoding Ambisonics and rendering binaural signals us- ing both spherical and arbitrary microphone arrays. 2.1 Signal Model Consider an array consisting ofM omnidirectional m...

work page 2025
[3]

PROPOSED METHOD This paper introduces a unified approach for Ambisonics encoding by jointly minimizing ASM and BSM errors us- ing a single loss function. Unlike existing methods that optimize these errors separately, the proposed method bal- ances both objectives within a unified framework to com- pute an Ambisonic filter that encodes Ambisonics opti- miz...

work page 2025
[4]

The model is tested with different α values, allowing us to analyze the trade-off between the ASM and BSM solutions, high- lighting the benefits of their combination

EXPERIMENT In this experiment, the joint ASM-BSM model is evalu- ated by measuring the ASM and BSM errors. The model is tested with different α values, allowing us to analyze the trade-off between the ASM and BSM solutions, high- lighting the benefits of their combination. 4.1 Setup Steering functions and HRTF were simulated in Matlab based on a spherical...

work page 2025
[5]

Matching for Ambisonics in this case typically leads to a high binaural error, signifi- cantly higher than what can usually be achieved by BSM

CONCLUSION AND FUTURE WORK A framework for designing an Ambisonics encoder for a wearable array was presented. Matching for Ambisonics in this case typically leads to a high binaural error, signifi- cantly higher than what can usually be achieved by BSM. The joint ASM-BSM design enables the optimization of ASM filters that are also tailored for BSM. These...

work page
[6]

Spatial audio signal pro- cessing for binaural reproduction of recorded acoustic scenes - review and challenges,

B. Rafaely, V . Tourbabin, E. Habets, Z. Ben-Hur, H. Lee, H. Gamper, L. Arbel, L. Birnie, T. Abhaya- pala, and P. Samarasinghe, “Spatial audio signal pro- cessing for binaural reproduction of recorded acoustic scenes - review and challenges,” vol. 6, 10 2022

work page 2022
[7]

Ambisonics: A Practical 3D Audio Theory for Recording, Studio Production, Sound Reinforcement, and Virtual Reality,

F. Zotter and M. Frank, “Ambisonics: A Practical 3D Audio Theory for Recording, Studio Production, Sound Reinforcement, and Virtual Reality,” 01 2019

work page 2019
[8]

Design of a fast broadband individual head-related transfer function measurement system,

B. Masiero, M. Pollow, and J. Fels, “Design of a fast broadband individual head-related transfer function measurement system,” vol. 97, pp. 136–136, 2011

work page 2011

Show all 21 references

[9]

Plane-wave decomposition of the sound field on a sphere by spherical convolution,

B. Rafaely, “Plane-wave decomposition of the sound field on a sphere by spherical convolution,” vol. 116, pp. 2149–2157, 1 2004

2004
[10]

Fundamentals of spherical array process- ing,

B. Rafaely, “Fundamentals of spherical array process- ing,” vol. 8, Springer, 2015

2015
[11]

Meyer and G

J. Meyer and G. W. Elko, “Analysis of the high- frequency extension for spherical eigenbeamforming 11th Convention of the European Acoustics Association M´alaga, Spain • 23rd – 26th June 2025 • microphone arrays.,” vol. 127, pp. 1979–1979, Acous- tical Society of America, 2010

2025
[12]

Analysis and design of spherical micro- phone arrays,

B. Rafaely, “Analysis and design of spherical micro- phone arrays,” vol. 13, pp. 135–143, 2005

2005
[13]

Binaural Reproduction From Microphone Array Sig- nals Incorporating Head-Tracking,

L. Madmoni, J. Donley, V . Tourbabin, and B. Rafaely, “Binaural Reproduction From Microphone Array Sig- nals Incorporating Head-Tracking,” in 2021 Immer- sive and 3D Audio: from Architecture to Automotive (I3DA), pp. 1–5, 2021

2021
[14]

Ambisonics encoding for arbitrary micro- phone arrays incorporating residual channels for bin- aural reproduction,

Y . Gayer, V . Tourbabin, Z. Ben-Hur, J. Donley, and B. Rafaely, “Ambisonics encoding for arbitrary micro- phone arrays incorporating residual channels for bin- aural reproduction,” in 2024 IEEE International Con- ference on Acoustics, Speech, and Signal Processing Workshops (...

2024
[15]

Design and analysis of binaural sig- nal matching with arbitrary microphone arrays,

L. Madmoni, Z. Ben-Hur, J. Donley, V . Tourbabin, and B. Rafaely, “Design and analysis of binaural sig- nal matching with arbitrary microphone arrays,”arXiv preprint arXiv:2408.03581, 2024

2024 arXiv
[16]

Ambisonics Capture using Microphones on Head-worn Device of Arbitrary Ge- ometry,

A. Bastine, L. Birnie, T. D. Abhayapala, P. Samaras- inghe, and V . Tourbabin, “Ambisonics Capture using Microphones on Head-worn Device of Arbitrary Ge- ometry,” in 2022 30th European Signal Processing Conference (EUSIPCO), pp. 309–313, 2022

2022
[17]

Room acoustics mea- surements by microphone arrays,

I. Balmages and B. Rafaely, “Room acoustics mea- surements by microphone arrays,” in 2004 23rd IEEE Convention of Electrical and Electronics Engineers in Israel, pp. 420–423, 2004

2004
[18]

Analysis and design of spherical micro- phone arrays,

B. Rafaely, “Analysis and design of spherical micro- phone arrays,” IEEE Transactions on speech and au- dio processing, vol. 13, no. 1, pp. 135–143, 2004

2004
[19]

Parametric Ambisonic Encoding of Arbi- trary Microphone Arrays,

L. McCormack, A. Politis, R. Gonzalez, T. Lokki, and V . Pulkki, “Parametric Ambisonic Encoding of Arbi- trary Microphone Arrays,” vol. 30, pp. 2062–2075, 2022

2022
[20]

Tikhonov regularization and total least squares,

G. H. Golub, P. C. Hansen, and D. P. O’Leary, “Tikhonov regularization and total least squares,” vol. 21, pp. 185–194, SIAM, 1999

1999
[21]

Easycom: An augmented reality dataset to support algorithms for easy communication in noisy environments,

J. Donley, V . Tourbabin, J.-S. Lee, M. Broyles, H. Jiang, J. Shen, M. Pantic, V . K. Ithapu, and R. Mehra, “Easycom: An augmented reality dataset to support algorithms for easy communication in noisy environments,” arXiv preprint arXiv:2107.04174 , 2021. Figure 1 : An illustr...

2021 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [1]

Ambisonics Encoder for Wearable Array with Improved Binaural Reproduction

INTRODUCTION Binaural reproduction is essential for creating immer- sive auditory experiences in virtual reality (VR) and aug- mented reality (AR), enabling spatial perception by en- abling real-world acoustic cues [1]. Ambisonics [2] has emerged as a powerful method for spatial audio represen- tation due to its ability to encode sound fields independent ...

work page Pith review arXiv 2025

[2] [2]

2.1 Signal Model Consider an array consisting ofM omnidirectional micro- phones, where each microphone is located at spherical co- ordinates (ri, θi, ϕi) for 1 ≤ i ≤ M

BACKGROUND This section provides the mathematical foundation for the signal model, presents the Ambisonics representation de- rived from this model, and outlines the formulations for encoding Ambisonics and rendering binaural signals us- ing both spherical and arbitrary microphone arrays. 2.1 Signal Model Consider an array consisting ofM omnidirectional m...

work page 2025

[3] [3]

PROPOSED METHOD This paper introduces a unified approach for Ambisonics encoding by jointly minimizing ASM and BSM errors us- ing a single loss function. Unlike existing methods that optimize these errors separately, the proposed method bal- ances both objectives within a unified framework to com- pute an Ambisonic filter that encodes Ambisonics opti- miz...

work page 2025

[4] [4]

The model is tested with different α values, allowing us to analyze the trade-off between the ASM and BSM solutions, high- lighting the benefits of their combination

EXPERIMENT In this experiment, the joint ASM-BSM model is evalu- ated by measuring the ASM and BSM errors. The model is tested with different α values, allowing us to analyze the trade-off between the ASM and BSM solutions, high- lighting the benefits of their combination. 4.1 Setup Steering functions and HRTF were simulated in Matlab based on a spherical...

work page 2025

[5] [5]

Matching for Ambisonics in this case typically leads to a high binaural error, signifi- cantly higher than what can usually be achieved by BSM

CONCLUSION AND FUTURE WORK A framework for designing an Ambisonics encoder for a wearable array was presented. Matching for Ambisonics in this case typically leads to a high binaural error, signifi- cantly higher than what can usually be achieved by BSM. The joint ASM-BSM design enables the optimization of ASM filters that are also tailored for BSM. These...

work page

[6] [6]

Spatial audio signal pro- cessing for binaural reproduction of recorded acoustic scenes - review and challenges,

B. Rafaely, V . Tourbabin, E. Habets, Z. Ben-Hur, H. Lee, H. Gamper, L. Arbel, L. Birnie, T. Abhaya- pala, and P. Samarasinghe, “Spatial audio signal pro- cessing for binaural reproduction of recorded acoustic scenes - review and challenges,” vol. 6, 10 2022

work page 2022

[7] [7]

Ambisonics: A Practical 3D Audio Theory for Recording, Studio Production, Sound Reinforcement, and Virtual Reality,

F. Zotter and M. Frank, “Ambisonics: A Practical 3D Audio Theory for Recording, Studio Production, Sound Reinforcement, and Virtual Reality,” 01 2019

work page 2019

[8] [8]

Design of a fast broadband individual head-related transfer function measurement system,

B. Masiero, M. Pollow, and J. Fels, “Design of a fast broadband individual head-related transfer function measurement system,” vol. 97, pp. 136–136, 2011

work page 2011