SPIE Fireside Chat with Tomas Sluka

Big thanks to Bernard Kress for the great opportunity to share what we do at CREAL with the most relevant audience at the prime online webinar on AR hardware #spiexr !

It is super encouraging to see the conclusion that “real” 3D display solutions such as Light-Field (LF) and Computer Generated Holography (CGH), are high above the existing flat stereo displays – at least in the context of visual comfort and 3D realism. Indeed, they provide literally a new dimension to image quality.

I even agree that the CGH is in principle an even more complete 3D display solution than LF and cheer for every progress in it.

Here, I would like to take the liberty to elaborate more on the practical side of the differences between LF and CGH. Because there is a reason why we do LF and not CGH. In short: Both concepts ultimately serve the same purpose very similarly well, but our LF does it more practically than CGH from the product point of view. Indeed, it works pretty well already now. I, therefore, see the LF display as the “better one bird in the hand than another one in the sky” (yes, not two). 

A more technical explanation for those interested:
CGH, in its core principle, shapes light waves as the real 3D world does – it is a very sound technology – while LF is already in its core principle “only” an approximation of light waves. In practice, however, CGH is a digital approximation, too, but this time, its technical realization causes significantly more difficulties to make it work sufficiently well compared to LF. In consequence, the difference between the physics of the two concepts has practically no impact on the achievable experience, while technical realization has a huge impact. For instance, LF already provides high-quality real 3D imagery (as far as human vision is concerned) and at practically no extra computing cost, while CGH still needs some time to get there. Indeed, there are big chances that CGH will resort to displaying less “real” 3D, such as only a few focal planes. The image will be certainly equally good enough, but the other qualities will be different, such as the cost, form-factor, robustness, brightness, field-of-view, eye-box, power efficiency, computational efficiency, or framerate, and, mainly, the time of market readiness. LF is potentially at least good enough in all those qualities, while it is exceptionally good in some of the most important ones, such as in the efficiency and framerate. CGH possibly too, but why struggle and wait for the same, then :-)”

  • – Tomas Sluka