4 ms·

Black and blue, or white and gold?

by tomlockwood 5y ago

danbruc 5y ago
How is that relevant?
- pjc50 5y ago
  https://en.wikipedia.org/wiki/The_dress https://en.wikipedia.org/wiki/The_dress : a viral tweet which triggered a huge discussion of colour illusion. Colour really is perceptual. There is, once you get sufficiently pedantic, no such thing as the "actual colour" of the car. You can say things like "spectrally within 1% of Pantone Red 032C under standardised lighting conditions", but the range of responses that will produce in traditional or digital photographs and then on screens and human eyeballs is much wider. I'm not clear whether tomlockwood is trying to define "real" as "consensus (perception of) reality"?
  - tomlockwood 5y ago
    I'm not really trying to define anything - but yeah, unless we concretely set the pantone of the car by cross-referencing it or something, maybe that colour is still up for debate? The pantone thing would be the microscope equivalent in the post.
    - danbruc 5y ago
      No, the color of the car is never up for debate, what is up for debate are your believes about what the color is or how to best describe or quantify what you perceive or belief.
      - tomlockwood 5y ago
        How would you tell?
        danbruc 5y ago
        Tell what?
        tomlockwood 5y ago
        That the colour of the car wasn't up for debate. What would you measure to determine qualities like this aren't up for debate?
        danbruc 5y ago
        You could measure something like the bidirectional scattering distribution function [1] including its wavelength dependence. Or something even fancier that accounts for polarization and displacement due to subsurface scattering and what not. The important thing is that you can objectively quantify the color and appearance of the car in great detail mostly limited by your dedication and budget. And with that information you could recreate the color of the car and how the paint looks under some given lighting conditions which can also be quantified. There is really nothing to be debated here, at best maybe the fidelity and sophistication of an realization of this method and whether it is good enough for a given purpose. And I just say this, there is nothing to be measured in order to determine whether something is up for debate or not. This is a question of logic and philosophy and physics. Being up for debate means that there must be different opinions. This can happen due to a lack of information or knowledge. Or due to unspecified assumptions. And probably more. Whether you should have high or low tax rates is up for debate. This right answer will depend on your goals. On the behavior of people and societies and your knowledge of and assumption about those things. About whom do you care, who is affected how? The better you know how things work and the clearer your goals are, the less room will there be for debate. But what could you debate about the physical properties of the paint of the car? They are facts of the universe - at least unless or our understanding of the universe is really radically wrong. You can measure them. But opinions you can only have about relating things, not what the color is but how it looks to you. What it reminds you of. How you name or describe it. What it looks to a color blind person. What you think it looks to a color blind person. Whether it was worth the money. Whether it should be changed. [1] https://en.wikipedia.org/wiki/Bidirectional_scattering_distribution_function https://en.wikipedia.org/wiki/Bidirectional_scattering_distr...
  - danbruc 5y ago
    This is just somewhat conflating a lot of things. There is a dress with some colors, there is a photo of that dress with different colors because of the specific lighting conditions the photo was taken under and whatever processing the camera performed, and then there are the colors people perceive when looking at the photo, again influenced by all kind of factors. And there is reality to all of them, the actual color of the dress, the pixel values in the photo, and the perceived colors. But whatever color you see looking at the photo does not change the pixel values in the photo, and the pixel values in the photo do not change the spectral properties of the actual dress.

Black and blue, or white and gold?

danbruc 5y ago

How is that relevant?

pjc50 5y ago
https://en.wikipedia.org/wiki/The_dress https://en.wikipedia.org/wiki/The_dress : a viral tweet which triggered a huge discussion of colour illusion. Colour really is perceptual. There is, once you get sufficiently pedantic, no such thing as the "actual colour" of the car. You can say things like "spectrally within 1% of Pantone Red 032C under standardised lighting conditions", but the range of responses that will produce in traditional or digital photographs and then on screens and human eyeballs is much wider. I'm not clear whether tomlockwood is trying to define "real" as "consensus (perception of) reality"?
- tomlockwood 5y ago
  I'm not really trying to define anything - but yeah, unless we concretely set the pantone of the car by cross-referencing it or something, maybe that colour is still up for debate? The pantone thing would be the microscope equivalent in the post.
  - danbruc 5y ago
    No, the color of the car is never up for debate, what is up for debate are your believes about what the color is or how to best describe or quantify what you perceive or belief.
    - tomlockwood 5y ago
      How would you tell?
      - danbruc 5y ago
        Tell what?
        tomlockwood 5y ago
        That the colour of the car wasn't up for debate. What would you measure to determine qualities like this aren't up for debate?
        danbruc 5y ago
        You could measure something like the bidirectional scattering distribution function [1] including its wavelength dependence. Or something even fancier that accounts for polarization and displacement due to subsurface scattering and what not. The important thing is that you can objectively quantify the color and appearance of the car in great detail mostly limited by your dedication and budget. And with that information you could recreate the color of the car and how the paint looks under some given lighting conditions which can also be quantified. There is really nothing to be debated here, at best maybe the fidelity and sophistication of an realization of this method and whether it is good enough for a given purpose. And I just say this, there is nothing to be measured in order to determine whether something is up for debate or not. This is a question of logic and philosophy and physics. Being up for debate means that there must be different opinions. This can happen due to a lack of information or knowledge. Or due to unspecified assumptions. And probably more. Whether you should have high or low tax rates is up for debate. This right answer will depend on your goals. On the behavior of people and societies and your knowledge of and assumption about those things. About whom do you care, who is affected how? The better you know how things work and the clearer your goals are, the less room will there be for debate. But what could you debate about the physical properties of the paint of the car? They are facts of the universe - at least unless or our understanding of the universe is really radically wrong. You can measure them. But opinions you can only have about relating things, not what the color is but how it looks to you. What it reminds you of. How you name or describe it. What it looks to a color blind person. What you think it looks to a color blind person. Whether it was worth the money. Whether it should be changed. [1] https://en.wikipedia.org/wiki/Bidirectional_scattering_distribution_function https://en.wikipedia.org/wiki/Bidirectional_scattering_distr...
- danbruc 5y ago
  This is just somewhat conflating a lot of things. There is a dress with some colors, there is a photo of that dress with different colors because of the specific lighting conditions the photo was taken under and whatever processing the camera performed, and then there are the colors people perceive when looking at the photo, again influenced by all kind of factors. And there is reality to all of them, the actual color of the dress, the pixel values in the photo, and the perceived colors. But whatever color you see looking at the photo does not change the pixel values in the photo, and the pixel values in the photo do not change the spectral properties of the actual dress.