Geometric Dynamics and Semantic Communication for Retinal-Resolution XR Video Transmission: A Technical and Systems-Level Analysis

This work presents a technical and systems-level analysis of extended reality (XR) video transmission under retinal-resolution constraints, focusing on the interaction between geometric representation, semantic abstraction, and causal latency limits. Rather than proposing a new codec or deployment-ready system, the contribution reframes the XR transmission problem as a geometric capacity allocation problem. The analysis demonstrates that conventional Euclidean representations, even when combined with state-of-the-art codecs, are structurally misaligned with the hierarchical and non-uniform nature of human visual perception at high angular resolutions. The paper introduces the concept of dynamic geometric error reallocation, showing that the primary benefit of non-Euclidean geometry lies in controlling where reconstruction error occurs, rather than minimizing global distortion metrics. This perspective exposes fundamental limitations of traditional quality measures (e.g., PSNR, MSE) in XR regimes. A clear separation is established between signal geometry (pixel-level spatial organization) and semantic geometry (hierarchical and object-level representations), resolving common ambiguities in prior work. The analysis further demonstrates why hybrid geometric architectures—combining Euclidean residual channels with non-Euclidean structural representations—are a mathematical necessity rather than an engineering compromise. Crucially, the work treats motion-to-photon latency as a first-class causal constraint. It shows that bandwidth-optimal solutions, including semantic communication and generative reconstruction, can be physically infeasible under strict latency budgets, shifting the bottleneck from transmission to reconstruction. This leads to the formulation of a fundamental decoder-latency paradox for semantic XR systems. The paper consolidates experimental evidence from controlled geometric transformations to support these claims, without relying on proprietary implementations or system-specific optimizations. The result is a coherent framework that clarifies when geometric and semantic approaches are theoretically justified, when they fail, and how they should be evaluated. This work is intended as a conceptual and analytical foundation for future research in XR, video transmission, and semantic communication, particularly in bandwidth-constrained or latency-tolerant scenarios such as satellite links and non-interactive streaming.