Large Étendue 3D Holographic Display with Content-adaptive Dynamic Fourier Modulation

TL;DR

A newetendue expansion approach is described that combines multiple coherent sources with content-adaptive amplitude modulation of the hologram spectrum in the Fourier plane and devise a pupil-aware gradient-descent-based computer-generated holography algorithm that is supervised by a large-baseline target light field.

Abstract

Emerging holographic display technology offers unique capabilities for next-generation virtual reality systems. Current holographic near-eye displays, however, only support a small étendue, which results in a direct tradeoff between achievable field of view and eyebox size. Étendue expansion has recently been explored, but existing approaches are either fundamentally limited in the image quality that can be achieved or they require extremely high-speed spatial light modulators. We describe a new étendue expansion approach that combines multiple coherent sources with content-adaptive amplitude modulation of the hologram spectrum in the Fourier plane. To generate time-multiplexed phase and amplitude patterns for our spatial light modulators, we devise a pupil-aware gradient-descent-based computer-generated holography algorithm that is supervised by a large-baseline target light field. Compared with relevant baseline approaches, our method demonstrates significant improvements in image quality and étendue in simulation and with an experimental holographic display prototype.

Figures (6)

• Figure 1: System Architecture. Multiple mutually incoherent sources illuminate a fast phase-only SLM, creating a high-étendue backlight. An additional amplitude display is placed at the Fourier plane to remove ghost image artifacts created by HDOs and the multisource illumination. The phase and amplitude patterns are optimized jointly for a target light field in a content adaptive manner. We illustrate our design following the style of the schematic in grace2023multi for easier comparison.
• Figure 2: Tradeoff between 2D field of view (FoV) and eyebox size. Each white line represents the fixed étendue of a holographic display system illuminated by a grid of $\alpha \times \alpha$ sources with different $\alpha$ values. We show the étendue of the systems in log-scaled color maps. As the number of sources increase, the étendue of the system also increases.
• Figure 3: Illustration of baseline display configurations. Single-source configurations (I, II) trade image quality off for étendue expansion factor (II). Multi-source approaches that use all sources simultaneously (III, IV) benefit from a high-étendue "backlight" but operate within limited effective degrees of freedom, which also makes it challenging to achieve a high image quality. Steered illumination approaches (VI, VII) sequentially illuminate the system from different directions and require high-speed SLMs. Without a Fourier mask, the image quality achieved by these systems is also limited by high diffraction orders (VI). We illustrate our design following the style of the schematic in grace2023multi for easier comparison.
• Figure 4: Photograph of our multi-source holographic display prototype. The propagation path is illustrated in red and components are labeled.
• Figure 5: Comparison of hardware configurations using simulated reconstruction. Here, we compare different étendue-expanded holographic display configurations, including the conventional setup with light-field supervision (1st row), phase masks kuo2020mask (2nd row), multisource (3rd row), multisource with random Fourier mask jo2022binary (4th row), steering (5th row), steering with filter (6th row), and ours with 1-frame and 6-frame time multiplexing (multisource with content-adaptive dynamic Fourier modulation, 7--8th rows). For each configuration, we present the central view (1st column) and insets at two focal slices (rear and front) and two different viewpoints (left and right) in the next four columns. Quantitative evaluations are included as PSNR (dB)/SSIM on the left. Note that all methods use the same, large-baseline target light field for supervision, which degrades the quality of the single-source configuration (I) because it simply does not support such a large eyebox.