3 ms·

Physicists unlock multispectral secrets of earliest color photographs

42 points by samizdis 5y ago

failwhaleshark 5y ago
No. 4 - How hipsters make coffee, and photographs. Okay fine, haters, it's interesting. :) It's depth-dependent wavelength encoding. I never learned about this in a History of Photography course that included a visitor who made a daguerreotype of me in a Civil War uniform. It sounds like a physical, 3D spectrum analysis (2D to physical, 1D to wavelength) The now extinct Kodachrome color K-14 process is also interesting because the colors last longer than early negatives of generic color film E-6 processes, the kind my grandparents took. Does anyone know why K-14 went away? Did the developing chemicals expire rapidly?
- exmadscientist 5y ago
  It seems K14 was very, very different from all other common processes, according to the guy who figured out how to do it himself: https://emulsive.org/articles/darkroom/developing-film/they-took-my-kodachrome-away-so-i-brought-it-back https://emulsive.org/articles/darkroom/developing-film/they-... > Kodachrome’s K-14 process is quite unlike any other. You see, Kodachrome isn’t ACTUALLY a color film, it’s several layers of black and white film sensitized to the different colors and the necessary color dyes are added during the development process. This involves....
  - failwhaleshark 5y ago
    Yep. Similar to the OP article in the sense of different and more complicated than the E-6 process.
- ginko 5y ago
  The Kodachrome process was very complicated. While you can develop E-6 with just 3 baths you need several development and exposure steps to do K-14. Here's an illustration: https://miro.medium.com/proxy/1*w0fPDTcXMkMzSsVHQgGWrw.gif https://miro.medium.com/proxy/1*w0fPDTcXMkMzSsVHQgGWrw.gif The chemistry is apparently quite complicated and unstable as well with film labs requiring a dedicated chemist to keep everything in check. It's not one set-and-mostly-forget developing machine in your back room like with C-41 color negatives.
_Microft 5y ago
It does not say that anywhere on the page of the magazine but the paper seems to be open access: https://www.pnas.org/content/118/17/e2008819118 https://www.pnas.org/content/118/17/e2008819118 (HTML) https://www.pnas.org/content/pnas/118/17/e2008819118.full.pdf https://www.pnas.org/content/pnas/118/17/e2008819118.full.pd... (PDF) There is a photograph of their "digital Lippmann camera" in the appendix but unfortunately no further information on it: https://www.pnas.org/content/pnas/suppl/2021/04/15/2008819118.DCSupplemental/pnas.2008819118.sapp.pdf https://www.pnas.org/content/pnas/suppl/2021/04/15/200881911... (PDF)
- _Microft 5y ago
  Would anyone have a look at the photo of the digital Lippmann camera and let me know what you think? There seems to be a camera on the left-hand side, a beam splitter in the center, a mirror to the top, another mirror on a translation-stage to the right which would leave the bottom for the incoming beam. So this digital Lippmann camera is "just" a Fourier transform spectrometer that has a camera instead of a single detector element as sensor, right? Compare: https://en.wikipedia.org/wiki/Fourier-transform_infrared_spectroscopy#Components https://en.wikipedia.org/wiki/Fourier-transform_infrared_spe...
  - laserphysicist_ 5y ago
    Yes, you can call it a spatially-resolved FTSI, that's correct. There is a whole field of research around FTSI hyperspectral imagers. The big challenge is to get fast enough to get a decent time-resolution.
    - _Microft 5y ago
      To explain why this is „suddenly“ about time-resolution: the mirror on the translation stage is moved back and forth which changes the length of one arm of this interferometer. This movement takes time but is necessary as the interference pattern changes with position of the mirror. These changes are recorded and allow to compute the spectrum. Since the scanning takes time, this limits how fast images/spectra can be recorded.

42 points by samizdis 5y ago

No. 4 - How hipsters make coffee, and photographs. Okay fine, haters, it's interesting. :) It's depth-dependent wavelength encoding. I never learned about this in a History of Photography course that included a visitor who made a daguerreotype of me in a Civil War uniform. It sounds like a physical, 3D spectrum analysis (2D to physical, 1D to wavelength) The now extinct Kodachrome color K-14 process is also interesting because the colors last longer than early negatives of generic color film E-6 processes, the kind my grandparents took. Does anyone know why K-14 went away? Did the developing chemicals expire rapidly?

exmadscientist 5y ago
It seems K14 was very, very different from all other common processes, according to the guy who figured out how to do it himself: https://emulsive.org/articles/darkroom/developing-film/they-took-my-kodachrome-away-so-i-brought-it-back https://emulsive.org/articles/darkroom/developing-film/they-... > Kodachrome’s K-14 process is quite unlike any other. You see, Kodachrome isn’t ACTUALLY a color film, it’s several layers of black and white film sensitized to the different colors and the necessary color dyes are added during the development process. This involves....
- failwhaleshark 5y ago
  Yep. Similar to the OP article in the sense of different and more complicated than the E-6 process.
ginko 5y ago
The Kodachrome process was very complicated. While you can develop E-6 with just 3 baths you need several development and exposure steps to do K-14. Here's an illustration: https://miro.medium.com/proxy/1*w0fPDTcXMkMzSsVHQgGWrw.gif https://miro.medium.com/proxy/1*w0fPDTcXMkMzSsVHQgGWrw.gif The chemistry is apparently quite complicated and unstable as well with film labs requiring a dedicated chemist to keep everything in check. It's not one set-and-mostly-forget developing machine in your back room like with C-41 color negatives.

_Microft 5y ago

It does not say that anywhere on the page of the magazine but the paper seems to be open access: https://www.pnas.org/content/118/17/e2008819118 https://www.pnas.org/content/118/17/e2008819118 (HTML) https://www.pnas.org/content/pnas/118/17/e2008819118.full.pdf https://www.pnas.org/content/pnas/118/17/e2008819118.full.pd... (PDF) There is a photograph of their "digital Lippmann camera" in the appendix but unfortunately no further information on it: https://www.pnas.org/content/pnas/suppl/2021/04/15/2008819118.DCSupplemental/pnas.2008819118.sapp.pdf https://www.pnas.org/content/pnas/suppl/2021/04/15/200881911... (PDF)

_Microft 5y ago
Would anyone have a look at the photo of the digital Lippmann camera and let me know what you think? There seems to be a camera on the left-hand side, a beam splitter in the center, a mirror to the top, another mirror on a translation-stage to the right which would leave the bottom for the incoming beam. So this digital Lippmann camera is "just" a Fourier transform spectrometer that has a camera instead of a single detector element as sensor, right? Compare: https://en.wikipedia.org/wiki/Fourier-transform_infrared_spectroscopy#Components https://en.wikipedia.org/wiki/Fourier-transform_infrared_spe...
- laserphysicist_ 5y ago
  Yes, you can call it a spatially-resolved FTSI, that's correct. There is a whole field of research around FTSI hyperspectral imagers. The big challenge is to get fast enough to get a decent time-resolution.
  - _Microft 5y ago
    To explain why this is „suddenly“ about time-resolution: the mirror on the translation stage is moved back and forth which changes the length of one arm of this interferometer. This movement takes time but is necessary as the interference pattern changes with position of the mirror. These changes are recorded and allow to compute the spectrum. Since the scanning takes time, this limits how fast images/spectra can be recorded.