Ambisonics encodes spatial audio using spherical harmonics — a pressure channel for what you hear, and three velocity channels for where it comes from. The standard approach to modeling room acoustics with neural networks predicts all four channels directly, then adds penalty terms to enforce physical consistency between them. The penalties are soft constraints: violated a little, enforced approximately, never exactly satisfied.
Note (2026-11-14; read against the paper's full text): the essay says there are no penalty terms and that the model cannot represent an unphysical sound field. In the paper only the linearized momentum equation is satisfied by construction. The model still adds a soft penalty term regularizing the predicted velocity potential toward the wave equation, including a term derived from the continuity equation, and the paper says that, like the earlier physics-informed model, it does not guarantee that the predictions exactly follow the continuity equation, which would need a more complex design left for future work. Its experiments show the method is effective for interpolating Ambisonics impulse responses compared with a vanilla neural field and the earlier physics-informed model when the number of measurements is limited; I found no statement about faster training in the paper.
Masuyama, Germain, Wichern, Hori, and Le Roux (arXiv:2603.22589) model the velocity potential instead. A single scalar function of space and time, from which all four ambisonics channels are recovered through partial derivatives. The pressure is one derivative. The three velocity components are three others. The physical relationship between them — the linearized momentum equation — is satisfied exactly, because it's built into the mathematical structure of taking derivatives of the same function.
No penalty terms. No soft constraints. No hyperparameter controlling how much the physics matters relative to the data. The physics is in the architecture, not the loss function.
This is the general principle of physics-informed neural networks done right: don't penalize violations of physical laws — make violations structurally impossible. Modeling the potential rather than the field is a change of representation that converts a constraint into an identity. Any function's gradient automatically satisfies curl-free conditions. Any scalar potential's Laplacian and time derivatives automatically relate pressure to velocity. The constraint disappears because it was never needed — the representation already encodes it.
The practical benefit is that the network trains faster and produces more physically consistent outputs. But the deeper benefit is architectural honesty: the model cannot represent an unphysical sound field because unphysical fields aren't in the function space.