Thread replies: 12
Thread images: 2
Thread images: 2
File: 1470452053752.jpg (146KB, 1000x733px) Image search:
[Google]
146KB, 1000x733px
/g, does there exist some usb type device with which one could "sample" the color of say, a piece of paper, or some other material, and output the color code in whatever hex, or rgb format?
This would be used to prepare paints/dyes at another location of samples taken with the device. So the idea you could just give the person that makes the dyes the code and get the exact match that the device output.
Does anyone know what this might be called, if it exists?
Picture unrelated.
>>
>>56363349
Yeah, it's called a camera.
This is a stupid idea, OP.
The color detected by the photosensor will change based on ambient light.
As /diy/ will tell you, it is essentially impossible to reproduce an exact color. Pigments decay over time due to UV radiation.
>>
>>56363870
actually, depending on the use, I think there does exist such a device. It's not stupid.
>>
>>56363349
I think there's a few laser color sensors I can't recall where I saw them.
>>
>>56363349
>>56363942
Found it
>>
>>56363904
Attempting to get an exact color and texture match for paints and dyes is stupid, given that the idea is to patch a damaged section of an object, i.e. painting over a section of drywall patch on a wall.
If the idea is to copy the color and then do a complete recoloring of the object, i.e. painting an entire wall, then the idea isn't quite as stupid.
It then becomes a regular kind of stupid, because you're expending excessive amounts of effort to match a color to a certain precision standard, when it would otherwise look untouched.
>>
>>56363952
I'm retarded not to post the link
https://www.sparkfun.com/products/retired/11195
http://www.keyence.com/products/sensor/photoelectric/index.jsp
>>
It wouldn't be exact anyway, OP. Light emitted from your computer and all these sensors has different properties than paints and dyes. You'll only ever get approximations, so you might as well just eyeball it instead of spending money on snake oil
>>
File: microspectrometer.jpg (111KB, 804x245px) Image search:
[Google]
111KB, 804x245px
>>56363349
>>56363870
This use a camera that you have calibrated with a calibration pattern, if you need some SERIOUS COLOR SENSING ACTION then buy a serious color sensor https://www.sick.com/us/en/product-portfolio/registration-sensors/color-sensors/c/g113666
Fuck, Hamamatsu sells a goddamn visible light spectrometer(C12666MA) so you can get the color down with 15 nanometer resolution.
https://www.hamamatsu.com/jp/en/C12666MA.html
Pic related. Honestly I would really like to know why the Japanese think a dimmer needs 15 fucking nanometer wavelength resolution.
You can also get some shitty color sensors from adafruit/the chink copy of adafruit for much cheaper, but 15 NANOMETER WAVELENGTH RESOLUTION!
>>
>>56364109
This is the adafruit, I think?
https://www.adafruit.com/products/1334?gclid=CKX4grf27M4CFYs9gQodvrMAFA
And a calibration pattern would be? http://www.curtpalme.com/images/SMPTE.gif Something like that?
>>
>>56363959
no, sweet child.
this shit's been standardized since 1937.
the key nobody so far has mentioned is that you need a full-spectrum reference light to compliment the sensor. then you can account for metameric error and translate it accurately to human color vision.
>>56364109
15nm is pretty junk, actually. we only see a range of roughly 320nm. the difference between 500nm and 530 is massive.
>>56364559
buy (not print) a macbeth chart.
>>
>>56366229
>full-spectrum reference
could you just buy a 65w 120v bulb to do that or is that something more specialized?
Thread posts: 12
Thread images: 2
Thread images: 2