Every time some grifting corporate shill talks about “AI”.
Every.
Time.
Remote
troy_s 
@troy_s@mastodon.art
mastodon 4.7.3Digital colour gadfly. Antihumanist. Antipositivist. Acolyte of moving / static pixel hermeneutics and thaumaturgy. Make creative things matter.
817 Followers
367 Following
29 Posts
Joined November 05, 2022
Website:
GitHub:
This is absolutely the finest work ever produced on LinkedIn.
Marvel in all of its #pissoncarpet glory.
“Why I Built My Own Camera App (And You Should Too)”
“Why I Built My Own Camera App: RAW Data”
https://blog.thepixelary.com/post/810503615233048576/raw-data-my-first-taste-of-unprocessed-beauty
Replying to
When we form a pictorial depiction of these colourimetrically formed relative wattages? Our basic spatial invariance is broken. Our “energy upon wall” gradients will tragically yield busted pictures that have *no know solutions*.
All efforts of “gamut mapping” turn into charlatan games of moving coordinates around without a consideration of the goal. Yet more ill defined problems.
Under a spatial invariance lensing, a reasonable goal is far more clear, and changes the framing of the problem.
Replying to
@philippe@tchafia.be It seems like Pan might or might not release it.
In the short term, you can always give @petrikas@mastodon.art active web app a try.
https://www.saulala.com
Replying to
Davidoff, Jules B. Cognition through Color. Issues in the Biology of Language and Cognition. Cambridge, Mass: MIT Press, 1991.
Replying to
Ironically, anyone who examines a greyscale photograph can rule out ranges of colours within a segmented field.
Just as musical notes reside within a given octave, it seems colour has a similar correlate with luminance positions; the segmented form becomes the octave, and the notes reside in a hierarchy of increment to decrement relationships within the segmented regions.
“Exposing 'faceless' YouTubers pushing Canada to join the U.S.”
I've been finding a few interesting books of late...
Cole M, Scribner S. Culture and Thought; a Psychological Introduction. Wiley; 1974.
“Image registration based on the Fourier-Mellin transform
Log-polar transform”
https://sthoduka.github.io/imreg_fmt/docs/log-polar-transform/
Replying to
More folks should invest the time to think about what a pictorial depiction should encompass in terms of *invariants* with respect to what they cognized in front of the camera, or their PBR models.
We can clearly show that the idea of colourimetry reproduction is a fatally flawed construct, supported by absolute reams of evidence. Further, it leads to hilariously bad takes of requiring specific colourimetric hardware for a depiction.
Absurdity at its finest.
Replying to
For folks who are technically savvy, I would encourage someone to consider a flat plane and some camera. Put a lamp source close to the plane and make a record of the quantal interaction. Filter the lamp energy with a “blue” gel.
From the camera vantage, there are quantal records that pattern after the energy gradient that falls across the plane. In a pictorial depiction derived from these records, we would have a “shape invariance” emerge if the pictorial depiction is formed reasonably.
Replying to
However, if we apply a transform to form colourimetry? The most basic principle of shape invariance is violated.
Camera colourimetry stretches and warps the quantal records to “match” know coordinates. We can end up with virtual record primaries that are negative relative wattages, all because of this ridiculous obsession with colourimetry.
Those negative relative wattages, in terms of the *integrations* of the geometric point samples of R+G+B records, will “remove energy” from the sums.
Replying to
Perhaps related, it is intriguing that only a “shape” can be “black”. It seems challenging to imagine a “light” as “black” outside of the convention of using the term “black light” to describe a wholly different phenomena.
At which point can a surface be distinguished from a light? This would seem to critically hinge upon the idea of a local increment in the field.
Extending to colours, an energy *increment* would appear to be a prerequisite for something to be bound as a “light”.
Hats off to jpsalas… you have made some stunning virtual pinball tables!
Replying to
@pq
Apples to oranges.
But yes, it’s irrelevant.
Otherwise CMYK printers would not likely exist, and after all, that’s “painting”.
Having pre-mixed paints is a short cut that comes with advantages.
Imagine trying to suggest “Oh you can’t depict that rainbow without this paint.”
Absurd delusions.
Replying to
@pq I just don’t think the paints matter at all in the broader sense.
The most striking part of Impression Sunrise is the equiluminant condition of the sun, not the particular paints.
Painters know how to mix paints.
Replying to
@TonyVladusich Things being a “light bulb” for example.
It seems that it is at least challenging to suggest a “blackish” light bulb exists. Hence the idea that it has to at least start with a local segmented increment?
Replying to
@TonyVladusich That is the observation. Some things seem oddly non-existent in a duality mode.
Replying to
@TonyVladusich Except the cognition thereof we agree is not physicalist in the sense of the transport.
It is interesting that we have this particular structure, and that there appear to be some demarcations of phenomena that are less ambiguous.
Replying to
@TonyVladusich I am specifically speaking of the colour categorization.
If we accept broadly that there is some segmented energy field that increments and decrements sit within, it just seems odd that in no case can a “light” ever be of the decrement variety.
Plenty of colours etc. exist in this peculiar dual universe, but *some* things seem to be strictly increment only within the locally segmented form.
Replying to
@petrikas @TonyVladusich True.
It’s a constrained emission, but as the Staircase Gelb experiment shows, we need *some other* complexity in the articulation to identify it as such.
Tony suggested that Adelson (?) tried the test with surface plaster, and it did not manifest as “white”.