Why Holographic 3D Light Field Displays are Impossible, and How to Build One Anyway

In recent years, we have seen immense improvement in video quality culminating in today’s ultrahigh-definition with high dynamic range and wide color gamut. Viewers can no longer benefit from increases in resolution in flat, 2D images; they simply cannot see any more detail. Yet, both consumers and producers are looking for improved displays, including 3D displays. There have been repeated attempts to introduce stereoscopic 3D over many decades. These have either failed completely or lacked conspicuous success. Yet, people still seem fascinated by true 3D displays, such as laser-generated holograms. If high-quality true 3D displays were physically and commercially viable, it would be a transformative technology set to replace the billions of 2D displays currently in use. The consequences for the industry, both hardware and content production, would be enormous. This article seeks to address the potential for light field displays to become the next, and ultimate, display technology. In so doing, it discusses the underlying principles of light field displays and contrasts them to stereoscopic 3D with its many limitations. Producing high-quality light field displays is a very significant challenge. A huge amount of information must be conveyed to viewers so that they can see high-resolution images at different depths and from different perspectives. Light field displays are based on underlying 2D displays. Foremost among the technical challenges is the huge number of pixels required. While early commercial light field displays are already available, i they have limited spatial resolution and a very limited depth of field. The experience of viewing is something like viewing a puppet theater. Unfortunately, their conventional, century-old approach does not scale to large depths of field. This article describes how a light field display’s depth of field depends on the characteristics of the display. Based on conventional 2D sampling theory, it gives the absolute resolution of the display (that is the smallest object, say mm, that can be resolved). But viewers actually perceive angular resolution, so the analysis is adjusted accordingly. The analysis reveals the enormous number of pixels required for a large depth of field and, consequently, why existing approaches are untenable. By analyzing the image formation process from the viewer’s perspective, the article shows that many fewer pixels are required to achieve a large depth of field. Even so, more resolution is required. It is shown how this can be provided by rendering images over multiple frames, benefiting from the higher frame rates now becoming available. The analyses are complemented by an example, based on a seven-year-old display resolution, to demonstrate the viability of this approach. Considering current commercial, and near future, displays, the approach scales to larger, higher-resolution displays. Light field displays were not viable a few years ago. However, advancements in display technology and a better understanding of appropriate image rendering, presented here, imply that they have become practical and will become more so in the future.