paint mixing, named colors

[Dale]

in painting and maybe many-colorant systems, I think a "mathematical"
software mixing program, in coordination with a spectral/color
measurement device might have a market

paints/painting can be expensive

if you had a software, perhaps with a xy/uv chromaticity diagram
locating all the paints, you could mix on the diagram, then calculate
how much of each paint you want to get the color you want, and mix them
too get an estimate before trial and error mixing

you would need a "measurement tube" to dispense the right amount of
paint, might save expensive colorants like some paints

ICC has named colors, don't know if they are spectrum or color based

I don't know if paint companies have named color profiles for their
pant, they might want you to use more paints in the trial and error process

named colors and xy/uv diagrams are probably how mathematical constructs
are generated for many-colorant ink systems beyond CMYK

[Davoud]

Post by Dale
in painting and maybe many-colorant systems, I think a "mathematical"
software mixing program, in coordination with a spectral/color
measurement device might have a market

I think you're right. My local Sherwin-Williams store has had such a
device for years. Take a color sample to the store and a digital
colorimeter will measure the color and automatically add the correct
pigments to the base white to reproduce your sample color with amazing
precision.

[Dale]

Post by Davoud

Post by Dale
in painting and maybe many-colorant systems, I think a "mathematical"
software mixing program, in coordination with a spectral/color
measurement device might have a market

I think you're right. My local Sherwin-Williams store has had such a
device for years. Take a color sample to the store and a digital
colorimeter will measure the color and automatically add the correct
pigments to the base white to reproduce your sample color with amazing
precision.

forgot about that, they have been doing it for a long time

my thoughts were around expensive fine art paints where you have several
paints that are typical mixed by trial and error or memory of past trial
and error

you could have a software with a chromaticity diagram where you add
paints and mix them oriented to an end user like a fine art painter

[Savageduck]

Post by Dale

Post by Davoud

Post by Dale
in painting and maybe many-colorant systems, I think a "mathematical"
software mixing program, in coordination with a spectral/color
measurement device might have a market

I think you're right. My local Sherwin-Williams store has had such a
device for years. Take a color sample to the store and a digital
colorimeter will measure the color and automatically add the correct
pigments to the base white to reproduce your sample color with amazing
precision.

forgot about that, they have been doing it for a long time
my thoughts were around expensive fine art paints where you have
several paints that are typical mixed by trial and error or memory of
past trial and error
you could have a software with a chromaticity diagram where you add
paints and mix them oriented to an end user like a fine art painter

Check with Pantone. Perhaps they could come up with something for you.
<https://www.pantone.com/pages/pantone/index.aspx>

[Davoud]

Post by Dale

Post by Davoud

Post by Dale
in painting and maybe many-colorant systems, I think a "mathematical"
software mixing program, in coordination with a spectral/color
measurement device might have a market

I think you're right. My local Sherwin-Williams store has had such a
device for years. Take a color sample to the store and a digital
colorimeter will measure the color and automatically add the correct
pigments to the base white to reproduce your sample color with amazing
precision.

forgot about that, they have been doing it for a long time
my thoughts were around expensive fine art paints where you have several
paints that are typical mixed by trial and error or memory of past trial
and error
you could have a software with a chromaticity diagram where you add
paints and mix them oriented to an end user like a fine art painter

I suppose, but Rembrandt and DaVinci got along without digital
colorimetry, and so can you.

[Mayayana]

| > I think you're right. My local Sherwin-Williams store has had such a
| > device for years. Take a color sample to the store and a digital
| > colorimeter will measure the color and automatically add the correct
| > pigments to the base white to reproduce your sample color with amazing
| > precision.
| >
|
| forgot about that, they have been doing it for a long time
| my thoughts were around expensive fine art paints where you have several
| paints that are typical mixed by trial and error or memory of past trial
| and error
|
| you could have a software with a chromaticity diagram where you add
| paints and mix them oriented to an end user like a fine art painter
|

I buy a fair amount of paint for my work and
have had a lot of computer matches in the past.
My experience is that a computer rarely gets it
right, and even then it's only with off-whites.
I don't try to get a match unless I know the clerk
doing the matching and know that he can/will
tinker with it to get it right. Some things just
don't lend themselves to automation.

[isw]

Post by Dale
in painting and maybe many-colorant systems, I think a "mathematical"
software mixing program, in coordination with a spectral/color
measurement device might have a market
paints/painting can be expensive
if you had a software, perhaps with a xy/uv chromaticity diagram
locating all the paints, you could mix on the diagram, then calculate
how much of each paint you want to get the color you want, and mix them
too get an estimate before trial and error mixing

That won't provide enough information. You need to know the *specific*
pigments which were used, and that's nearly (or actually) impossible to
derive from the spectral values. And then, of course, you'd need to have
them on hand ...

Look up "metamerism".

Isaac

[Dale]

Post by isw
Look up "metamerism".

yes ... you could do mixing from spectral data to avoid metamerism

[isw]

Post by Dale

Post by isw
Look up "metamerism".

yes ... you could do mixing from spectral data to avoid metamerism

But it would be *really difficult* if you could not identify the
specific colorants from their spectra (impossible?). And, of course,
you'd be unlikely to have every possible item on hand anyhow.

Possible in theory, not very likely in practice.

Isaac

[Dale]

Post by isw

Post by Dale

Post by isw
Look up "metamerism".

yes ... you could do mixing from spectral data to avoid metamerism

But it would be *really difficult* if you could not identify the
specific colorants from their spectra (impossible?). And, of course,
you'd be unlikely to have every possible item on hand anyhow.
Possible in theory, not very likely in practice.
Isaac

colorimetry alone would get you to "good enough color", accurate enough
too edit from

[isw]

Post by Dale

Post by isw

Post by Dale

Post by isw
Look up "metamerism".

yes ... you could do mixing from spectral data to avoid metamerism

But it would be *really difficult* if you could not identify the
specific colorants from their spectra (impossible?). And, of course,
you'd be unlikely to have every possible item on hand anyhow.
Possible in theory, not very likely in practice.
Isaac

colorimetry alone would get you to "good enough color", accurate enough
too edit from

Under very controlled lighting, true.

Isaac

[Martin Brown]

Post by isw

Post by Dale

Post by isw

Post by Dale

Post by isw
Look up "metamerism".

yes ... you could do mixing from spectral data to avoid metamerism

But it would be *really difficult* if you could not identify the
specific colorants from their spectra (impossible?). And, of course,
you'd be unlikely to have every possible item on hand anyhow.
Possible in theory, not very likely in practice.
Isaac

colorimetry alone would get you to "good enough color", accurate enough
too edit from

Under very controlled lighting, true.
Isaac

As a part of the UK National Gallery's "Making Colour" exhibition last
year they had an old master and a fake exactly matched for 2700K
incandescent room lighting in a lecture room where the lighting
conditions could be varied to simulate everything from natural full
sunlight to fluorescent lamps. The experiment which involved public
participation isn't shown on summary which concentrates only on the
exhibits in the exhibition rather than the modern demo forgery.

http://www.nationalgallery.org.uk/whats-on/exhibitions/making-colour

Part of the interactive test was to determine proportion of people who
were colour blind and to what extent and then who could see what. There
were startling differences between the two under the right lighting.

[Martin Brown]

Post by Dale
in painting and maybe many-colorant systems, I think a "mathematical"
software mixing program, in coordination with a spectral/color
measurement device might have a market

It does and it was pioneered in the 1960's by ICI Dyestuffs Division the
original being the Imperial Match Predictor and it was as good as the
best human colour matcher soon after its invention. Calibrated against a
Hardy's (sp?) spectrophotometer and human expert colourists.

http://textilefashionguide.blogspot.co.uk/2014/09/computer-color-matching-system-ccms.html

and

http://industrial.datacolor.com/wp-content/uploads/datacolorproductliterature/Instruments%20for%20Color%20Measurement_72.pdf

The original was for internal use only and gave them a significant edge
for about a decade before ICS began manufacturing similar instruments.

It is actually a lot harder to do with fabrics because the composition,
surface texture, process temperature and impurities in the mordants can
all affect the final colour outcome.

Post by Dale
paints/painting can be expensive

These days there are apps for your mobile phone that will take a picture
of a material and a set of reference colours on a test card and work out
the right mix for a dedicated custom colour paint system.

Post by Dale
if you had a software, perhaps with a xy/uv chromaticity diagram
locating all the paints, you could mix on the diagram, then calculate
how much of each paint you want to get the color you want, and mix them
too get an estimate before trial and error mixing
you would need a "measurement tube" to dispense the right amount of
paint, might save expensive colorants like some paints

They are in almost every DIY store. Great fun unless or when the OPO
fails to secure the lid properly before starting the shaking up!
(depending on your point of view)

Post by Dale
ICC has named colors, don't know if they are spectrum or color based
I don't know if paint companies have named color profiles for their
pant, they might want you to use more paints in the trial and error process
named colors and xy/uv diagrams are probably how mathematical constructs
are generated for many-colorant ink systems beyond CMYK

You are so out of date it is funny. No wonder Kodak went to the wall.

[Dale]

Post by Martin Brown
You are so out of date it is funny. No wonder Kodak went to the wall.

I have been out of the industry for going on 18 years...

I see Kodak has a spot color mixing solution, I don't know when this
came out ...

I think mixing might be a better algorithm than matching in some
many-colorant systems, just my opinion