---
title: "Lippmann: colour stored through the depth of a plate"
description: "How can a colourless photographic material return a coloured reflection?"
canonical_url: "https://sandboxphysics.com/stories/probe-evidence/lippmann-recorded-colour/"
language: "en"
page_type: "story"
alternate_languages:
  zh-CN: "https://sandboxphysics.com/zh/stories/probe-evidence/lippmann-recorded-colour/"
last_modified: "2026-10-22"
---

[Interference as a record and a ruler](https://sandboxphysics.com/stories/probe-evidence/#interference)

Apparatus · Observation · Inference

# Lippmann: colour stored through the depth of a plate

Sandbox Physics · 2026-10-22

How can a colourless photographic material return a coloured reflection?

1. Incident image light
2. Reflection behind the emulsion
3. Depth-dependent exposure
4. White-light reconstruction

The layers are a record inside the emulsion. A soap-film colour alone does not reproduce that recording process.

## The reflection existed during exposure

Lippmann placed a reflective backing in contact with a fine-grained photosensitive layer. Incident light and its reflection interfered through the emulsion depth. Exposure therefore varied with depth, not merely across the two-dimensional image. Development retained a fine structure that could later reflect selected colours.

The important invention was the recording process. Many thin films show interference colours without having recorded an image. A modern colour display that computes a hue from wavelength also skips the material step that made the photograph evidence.

Sources: [1]

## The layer spacing uses wavelength in the material

At normal incidence in a uniform recording medium, neighboring standing-wave intensity maxima are separated by half the wavelength in that medium. The vacuum wavelength divided by twice the refractive index gives the ideal spacing. The reflection phase changes where the maxima begin, not this periodic spacing.

A broad spectrum writes overlapping depth modulations. The developed plate has finite thickness and a material response; it does not store a list of independent RGB values. The viewing illumination and angle then determine how that record is read.

$$
\Lambda=\frac{\lambda_0}{2n}\quad\text{(normal incidence)}
$$

Sources: [1]

## Recording and viewing need different light paths

The historical account describes light reaching the reflective backing through the glass and emulsion during exposure. Later white-light viewing selects a reflected spectrum from the developed layered structure. A diagram that sends viewing light through an unchanged mercury bath confuses two stages.

Tilting the plate changes the optical path between layers and can change the observed colour. Development, shrinkage and the backing reflection also affect spectral fidelity. Seeing a convincing colour is not proof that the original spectrum has been recovered exactly.

Sources: [1] [2]

## Modern experiments test the distortions

The EPFL authors’ experimental model shows that plate thickness and backing reflection can distort the reproduced spectrum. This narrows a common claim that interferential photography simply preserves every colour perfectly. Their work concerns forward response and reconstruction, not a replacement account of the 1891 invention.

No spectrum displayed here is a digitized measurement of an original Lippmann plate. The spacing formula explains the monochromatic limit; a faithful broadband reconstruction needs an optical and material-response model.

Sources: [2]

## Use thin-film interference as preparation

The linked model explains how optical thickness selects reflected colours. Use it to predict an angle-dependent colour change, then return to the historical account and identify the additional exposure and development steps. Those steps distinguish a recorded photograph from a pre-existing film.

Sources: [1]

## Explore the related models

Each link states what the existing model can explain and where the historical measurement goes further.

### [Predict an interference colour](https://sandboxphysics.com/labs/thin-film-color-studio/)

Change thickness and viewing angle.

This model has no photosensitive exposure or development, so it is preparation for Lippmann photography, not a reproduction.

## Sources and reading notes

The original lecture and apparatus account are read in a hosted Nobel volume. The EPFL author abstract supplies the stated modern distortion result; its full paper is not claimed as read.

1. [G. Lippmann, Colour photography (1908)](https://www.nobelprize.org/uploads/2018/06/lippmann-lecture.pdf)

   Lecture and preceding apparatus account read, pp.183–188; the recording plate and its reflective backing are described separately from viewing. [Read the hosted original](https://www.nevis.columbia.edu/~haas/documents/physics_nobel_lectures/physics01-21-1.pdf)

2. [Baechler et al., Lippmann Photography: A Signal Processing Perspective (2022), EPFL author abstract](https://graphsearch.epfl.ch/en/publication/2c7d48a5-c66c-456b-bf4b-bd5f49663f70)

   Author abstract read: experimentally tested spectral distortions depend on plate thickness and backing reflection. Full paper not used.

[Probes, light and measurement: reading the evidence](https://sandboxphysics.com/stories/probe-evidence/) [Michelson: a fringe count becomes a length](https://sandboxphysics.com/stories/probe-evidence/michelson-fringes-and-length/) [Gabor: a photograph that preserves a wave reference](https://sandboxphysics.com/stories/probe-evidence/gabor-recording-and-reconstruction/)