Colour reversal · Experimental
Cinefotocolor
Spain's two-colour cinema of the early 1950s.
An orange to blue-green palette over a faint yellow tint on a flamenco scene. A model inspired by the process, not a reproduction. Original vs Ashigrain.
At a glance
- Exposure index
- 8
- Daylight
- Introduced
- 1948
- Discontinued
- Process
- Cinefotocolor two-colour print in one emulsion layer (tartrazine yellow, first record iron-toned blue, second record mordanted and dyed with a basic red)
- Prints on
- Reversal, viewed directly
About the film
Cinefotocolor was a two-colour process devised by Daniel Aragonés Puig at the Cinefoto laboratories in Barcelona, and released in France as Realcolor.
Aragonés patented it in Spain between 1946 and 1949, and the print process was also granted US patent 2,597,820 in 1952. The first Cinefotocolor feature, En un rincón de España, was shot in 1948 with a Debrie Super Parvo that ran one panchromatic negative at 48 frames per second behind alternating blue-green and red-orange filters, so fast movement left red and blue fringes. A later version split the light with a semi-reflecting prism onto two 35 mm black-and-white negatives at 24 frames per second, and Cinefoto owned two such cameras. The prints were ordinary black-and-white positive: one record was iron-toned to Prussian blue, the other mordanted and dyed with a basic red dye, and a tartrazine bath added a yellow tint that could be adjusted for each print. Eighteen of the thirty-five Spanish features made wholly or partly in colour between 1948 and 1954 used it, among them folk musicals starring Paquita Rico and Lola Flores, and Doña Francisquita and Duende y misterio del flamenco were its great successes. Its very low sensitivity and unstable baths made prints vary greatly from copy to copy, and the last film released in the process was Tres eran tres in 1954. Filmoteca Española restored Duende y misterio del flamenco in 1989, and new prints made in 1990 had their colour values supervised by Aragonés himself.
Sources for this history
- El Cinefotocolor: técnica, filmografía, análisis (PhD thesis, Universidad Carlos III de Madrid, 2016), Miguel Ángel Herrero Herrezuelo
- Cinefotocolor (Santiago Aguilar), Film Atlas (FIAF / George Eastman Museum)
- Cinefotocolor, Timeline of Historical Film Colors
- US Patent 2,597,820: Process for obtaining colored copies of cinematographic films (1952), Daniel Aragonés Puig
The look
Cinefotocolor was a Spanish two-colour process used for features from 1948 to 1954, printed on black-and-white positive that was toned, dyed and tinted yellow. Ashigrain's version is a labelled surrogate, a model inspired by the process rather than a reproduction. Expect an orange to blue-green palette over a faint yellow tint.






Best for
- Spanish folk musicals and period drama
- Flamenco and stage performance scenes
- Mid-century two-colour pastiche
Compare camera films
Science & sources


A two-colour model inspired by the process, not a reproduction of Cinefotocolor. No curve, sensitivity or dye measurement exists, so every curve is generated from named assumptions guided by the patent, a doctoral thesis and an archive description; its MTF is borrowed from Kodak Vision3 50D and its grain from Kodak Ektachrome 100D.
Response curves and the evidence behind the model
Compare manufacturer data with Ashigrain’s response on the same axes. These curves show how the model is calibrated to its sources; they do not establish a match to photographed film.
Characteristic response
RMS granularity
Density variation (σD). Where the datasheet gives a single grain value, the response away from that point is an approximation.
| Channel | Density | Datasheet-based σD | Model σD |
|---|---|---|---|
| R | 0.3 | 0.0078 | 0.007666 |
| G | 0.3 | 0.005511 | 0.0054 |
| B | 0.3 | 0.009647 | 0.009628 |
| R | 0.6 | 0.014312 | 0.014066 |
| G | 0.6 | 0.011207 | 0.010982 |
| B | 0.6 | 0.022254 | 0.022209 |
| R | 1 | 0.022954 | 0.02256 |
| G | 1 | 0.019265 | 0.018878 |
| B | 1 | 0.038149 | 0.038073 |
| R | 1.5 | 0.029592 | 0.029083 |
| G | 1.5 | 0.028175 | 0.027608 |
| B | 1.5 | 0.059424 | 0.059306 |
| R | 2 | 0.030736 | 0.030207 |
| G | 2 | 0.033158 | 0.032492 |
Model limits
These images show the film model in use. Spectral response, lighting and viewing conditions include approximations. Halation, bloom, grain size, weave and flicker are not fully measured for each film. The look of real film also depends on development and print timing.
- This is a two-colour model inspired by Cinefotocolor, rather than a reproduction of the original process.
- No characteristic curve, spectral sensitivity, dye spectrum, MTF or grain measurement of Cinefotocolor exists in the sources. Every curve is generated from named assumptions, guided by the patent, a doctoral thesis and an archive description.
- The model runs the two records in a three-layer path: the blue layer copies the second record and carries no dye of its own.
- The Prussian blue and tartrazine absorptions are placed at textbook peaks, and the basic red dye is a named proxy chosen for a neutral grey. The tartrazine tint is a uniform density whose strength is assumed; the patent gives only an approximately 3 per cent bath.
- The exposure index of 8, the daylight balance and the single high-contrast tone curve are assumptions.
- Print-to-print variation, the colour fringing of the first camera version and the orthochromatic second negative are not modelled.
- MTF is proxied from Kodak Vision3 50D (5203) and grain from Kodak Ektachrome 100D (5294).
- Custom lighting spectra, camera metering and the spectral reflectance inferred from fixed-white-balance footage are assumptions used by the model, not measurements of this film.
- Push and pull have no published curves for this film and use a generic model.
- Characteristic response outside the published exposure range is an extrapolation, not measured film data.
- The available sources do not measure veiling flare during projection. The model approximates its contribution to lifted blacks on screen, so the result is not a measured projection contrast.
Reciprocity: not modelled. Exposure time has no effect on this film.
Film guide: assumptions and practical choices
Sources
- Reconstructed: M. A. Herrero Herrezuelo, El Cinefotocolor: tecnica, filmografia, analisis (PhD thesis, Universidad Carlos III de Madrid, 2016), e-Archivo handle 10016/23354, bitstream 7161a39b-76d8-4ec4-b9dc-40824ae90860 (404 pp, PDF built 2016-02-29); defended 2016-02-10, p. 98
- Process: US Patent 2,597,820, D. Aragones Puig, Process for obtaining colored copies of cinematographic films (filed 1949-01-25, issued 1952-05-20), Google Patents image of US2597820A (OCR text layer); Spanish priority 1948-01-30, p. 1