Abstract

We explored the earliest example of multispectral imaging, developed in the late 19th century by Gabriel Lippmann, who received the 1908 Nobel Prize in Physics. Unlike standard trichromacy, Lippmann photography relies on an interferential process that records the entire colour spectrum. Producing and viewing Lippmann plates is complex and challenging, therefore Lippmann photography remains a rare curiosity.

We first considered producing faithful digital copies of Lippmann photographs, requiring high-dimensional multispectral acquisition. This led to a precise mathematical modelling of the two key aspects in Lippmann photography: the way in which standing waves are recorded in the light-sensitive medium, and the way in which these waves are reproduced when white light excites the recorded interference patterns. A physical model combined with mathematical analysis provided us with a new and much more complete understanding of the Lippmann process, which was verified by experimental acquisitions and by analysing the resulting plates via X-ray imaging and tomography. Some phenomena, absent from previous analyses, can be explained and verified with this new model. In particular, the folk theorem of perfect colour reproduction by Lippmann photography was shown to be just that, a folk theorem.

We also designed a “digital” Lippmann camera that mimics the acquisition process carried out by the original analogue Lippmann method. This is an alternative multispectral camera based on interferometry, and illustrates interesting trade-offs between number of channels, number of pixels, and exposure time.

We then considered the idea of reproducing the Lippmann recording process using femto-second lasers on an appropriate substrate. Printing in three-dimensions and with a full spectrum could lead to new high density permanent storage methods, a technology of great interest given the explosion of digital data.