2nd Hex Code Dyeing Experiment

I started these experiments with a theory, that I could use computer hexidecimal codes as a basis for a dye recipe. My first experiment led me on quite an unexpected journey (that series of posts starts here). After some trial and error, I got close enough results to want to try a second experiment.

The second color I chose, was also from my original Cedar Waxwing photo palette ....

... gold. The hex code for this color is #EE9628. From the online Color Converter, I learned that the CMYK (cyan, magenta, yellow, key black) breakdown for this code is:

magenta - 0.37
yellow - 0.83
black - 0.07

I learned a lot from the first experiment, but even so, I decided to follow the same procedure, to see if there is a pattern to the results.

I based my dye recipe on totaling the above and calculating the relative percentages from that total.

0.37 + 0.83 + 0.07 = 1.27
o.37 is 29% of 1.27
0.83 is 65% of 1.27
0.07 is 6% of 1.27

These percentages gave me the starting point for my recipe, substituting fuchsia for magenta, and turquoise for cyan.

From the first experiment I had learned that fuchsia is very strong in it's dyeing power compared to the other colors. I found that to get the yarn color to match the computer color, I had to adjust the percentage of fuchsia (see this post for details on that). Since I wasn't ready to draw a conclusive ratio based on that one experiment, I started with this second recipe "as is," and adjusted the fuchsia for subsequent samples.

I used those percentages to calculate milliliter measurements of individual dye stocks, based on the weight of the yarn I was dyeing for my samples.

The top sample used the percentages as is. In the next sample, I cut the amount of fuchsia by one half, in the bottom sample, by one quarter. Although the last sample is closest to the target color on the left, in actuality, I think it could use just a touch more fuchsia, perhaps up the percentage to 10%(?)

My conclusions? Well, based on only two experiments (which is not a very scientific statistic), I think that hex codes can be used as a basis for developing recipes, though not as recipes. The first step is to adjust any magenta (fuchsia) in the recipe, and that a decrease to 25% of the percentage in the computer hex code is a good starting point for approximate results. There may be other factors involved, but I would need to experiment more to determine that.

I'm not sure yet if I'm going to try any further adjustments on this particular color, or if I'm going to move on to another experiment. Summer seems a good time to dye, and since I have quite a bit of it left, I plan to make good use of it.

Related Posts:
Dye Recipes From Computer Color Codes: A Theory
1st Computer Hex Code Dyeing Experiments, series starts here

©  August 11, 2008 by  Leigh 

at http://leighsfiberjournal.blogspot.com

Computer Hex Code Dyeing 5: 1st Success?

By Leigh

I'm not sure whether or not I can actually call this a success or not, but here it is.

I will say with confidence that this is the closest I've gotten so far to my target blue. Is it exact? Even making allowances for the difference in media and electronic interpretations, I have to say no. Can I get any closer? I don't know.

The recipe used the same color proportions as I used for the sample on the right (turquoise MX-G, 76%; red MX-5B, 10%; and black MX-CWA, 14%.) With that series of samples, I experimented with adjusting the amounts of fuchsia and turquoise. This lighter blue sample was the closest match to my cedar waxwing photo.

For this newest sample, I dyed the yarn to a darker depth of shade (DOS), in this case 8%. Previously, my samples were 4% DOS.

So, is 8% DOS dark enough to match the target color? From what I've read, the 8% DOS is considered "very dark" or "very intense." Can a darker DOS be achieved? I haven't found a reference to anything darker. Some of you with more dyeing experience will be able to answer that.

Other possible experiments could include trying a different red (MX-8B), or a different black.

My conclusions from these exercises so far, is that a color's computer hex code can be used as a starting point for dyeing with Procion MX dyes. However, the density of the dye powders effects their relative strength and has to be accounted for. Is there a specific ratio or some sort of formula for this? I don't know. Can one be developed? I don't know that yet either. Will I continue experimenting? Oh yes.

©  29 July 2008 by Leigh 

at http://leighsfiberjournal.blogspot.com

Related Posts:
Dye Recipes From Computer Color Codes: A Theory
Computer Hex Code Dyeing 1: Palettes From Pictures

Computer Hex Code Dyeing 2: Wrestling With Recipes Computer Hex Code Dyeing 3: 1st Samples - No Joy
Computer Hex Code Dyeing 4: 2nd Samples - Getting Closer?

Computer Hex Code Dyeing 4: 2nd Samples - Getting Closer?

Well, Peg was absolutely correct when she said that I was learning a lot. Indeed!

After learning how slow Pro MX turquoise is, I decided that I needed another mini-skein using my original recipe. This time I left it in the dyepot for 2 days, allowing the turquoise to do it's thing. Below is a comparison of those two skeins. They follow exactly the same recipe, with the same proportions of Turquoise MX-G, Red MX-5B (fuchsia), and Black MX-CWA. The only difference is the amount of time the skeins were allowed to soak in the dyebath.

Still too purple. But I had also learned that the Procion fuchsia dyes are strong, so that a little goes a long way. So for my next series of experiments, I began to adjust the amount of fuchsia.

Since I needed 50 milliliters of dye solutions for each skein, I knew that I couldn't simply use less fuchsia. I would have to make up the difference by increasing something else, either the turquoise or the black. Increasing the black didn't seem to be the answer, so I made up the difference with the turquoise.

Here are the results from that (target color is on the left):

Click on above image to enlarge

For each sample, I decreased the fuchsia by half. So the top is 100% of the fuchsia called for, the second is a half, third is a quarter, and bottom is one eighth the amount of fuchsia. Each sample soaked for about 48 hours.

Obviously, none of them are anywhere near the color I was aiming for. Next steps might include finer adjustments of the turquoise and fuchsia. Or experimenting with depth of shade. In fact, Diane has been doing a similar series of hex code dyeing experiments with depth of shade on silk and muslin fabric. Her first set of swatches starts here. They are very interesting so I encourage you to go take a look.

One thing I noticed from my 2nd set of samples, is that the third one looks similar to another color in my original palette, #1470C4 ....

However, part of the problem with these experiments is that there is no way to accurately analyze the results. Why? Because there is no way to establish the "exact" overall hue of the yarn. In order to use my computer to determine the hex code for the yarn, I can use a color picker. However, a color picker will only grab the color from one of millions of pixels which make up a picture. Take a look at a tiny detail from the same yarn photo, enlarged enough to show the pixels.

You can see why it's easy to create an almost infinite palette from just one photo! You can also see why there is no accurate way to choose the exact color which matches the overall appearance of the yarn. This points to another problem, i.e. having to go through multiple layers of color interpretation -- eye > camera > photograph > photo software > computer monitor > eye. In the end, they can only be analyzed visually.

So, to do that, here is the original photo with my 2nd sample skeins.

Technically exact? No. Visually accurate? Close! And you know what? I feel that now at least, I'm somewhere in the ball park instead of wandering around the parking lot.

Conclusions? None yet, but I am still intrigued enough to continue experimenting along this line.

Next ..... Computer Hex Code Dyeing 5: 1st Success?

Related posts:
Dye Recipes From Computer Color Codes: A Theory
Computer Hex Code Dyeing 1: Palettes From Pictures
Computer Hex Code Dyeing 2: Wrestling With Recipes
Computer Hex Code Dyeing 3: 1st Samples - No Joy

Computer Hex Code Dyeing 5: 1st Success?

Originally published July 21, 2008

© by Leigh at https://leighsfiberjournal.blogspot.com/

Computer Hex Code Dyeing 3: 1st Samples - No Joy

Since my last post, I've dyed my first samples. I have to report that while I got some gorgeous colors, they are nowhere near my target samples. They're not even close enough for me to plead computer discrepancies! ;)

Here they are:

Target (see "Palettes From Pictures"):
#0025B9
cyan - 1.0
magenta - 0.8
black - 0.2

Using those numbers to calculate percentages, these are the recipes I used based on a 1% stock solution. As you can see, I experimented a little with different fuchsias and blacks.

Sample #1:
Turquoise MX-G - 24 mls
Fuchsia MX-5B - 19 mls
Black CWA - 7 mls

Sample #2:
Turquoise MX-G - 24 mls
Fuchsia MX-8B - 19 mls
Black CWA - 7 mls

Sample #3:
Turquoise MX-G - 24 mls
Fuchsia MX-5B - 19 mls
Black 602A - 7 mls

Sample #4:
Turquoise MX-G - 24 mls
Fuchsia MX-8B - 19 mls
Black 602A - 7 mls

Nowhere near the target color. In fact, these samples are actually closer to the prediction I got from the Dye Mixer Applet.

Click on image to enlarge.

In comparing my four samples, the color differences are subtle, in fact barely discernible on my computer monitor. Most notably, the fuchsia 5B samples (1 & 3) appear a little cooler than the fuchsia 8Bs (2 & 4). Not much difference between the blacks however. I also note that I didn't stir well enough, as the color is spotty. About the only thing I am satisfied with is the depth of strength (DOS).

Comparing them to the target color, they are obviously purple. Too purple, which means the fuchsia was too strong.

The second color I tried to duplicate was the gold: 

Target:
#EE9628
magenta - 0.37
yellow - 0.83
black - 0.07

Sample #1
fuchsia 8B - 19.5 mls
yellow 8G - 32.5 mls
black CWA - 3 mls

Sample #2
fuchsia 8B - 19.5 mls
yellow 3RA - 32.5 mls
black CWA - 3 mls

For these, the only color I experimented with was the yellow. #1 is the cool yellow sample, and #2 used a warm yellow. But! Neither one is anywhere near my target color. And again, the fuchsia dominated, creating orange. (I did do a better job stirring during the leveling phase though.)

These experiments affirm what I wondered about, and what many of you confirmed in the comments; that equal weights of various Procion MX colors are not equal in dyeing strength.

So it would seem that the next logical step would be to experiment with the proportions. It would be wonderful if I could come up with some concrete numbers for these differences, but that remains to be seen.

Next ....... Computer Hex Code Dyeing 4: 2nd Samples - Getting Closer

Related Posts:
Dye Recipes From Computer Color Codes: A Theory
Computer Hex Code Dyeing 1: Palettes From Pictures
Computer Hex Code Dyeing 2: Wrestling With Recipes
Computer Hex Code Dyeing 4: 2nd Samples - Getting Closer
Computer Hex Code Dyeing 5: 1st Success?

Originally published July 12, 2008

© by Leigh at https://leighsfiberjournal.blogspot.com/

Computer Hex Code Dyeing 2: Wrestling With Recipes

Questions! (And my proposed answers.)

The first question is what to do with those numbers:

#0025B9
cyan - 1.0
magenta - 0.8
black - 0.28


Do they need to be percentages? Proportions? Weights?

To test the numbers on my various colors, I have been plugging them into the Dye Mixer Applet. Sometimes the applet results matched the html sample, sometimes they didn't. After a little research, I have learned some interesting things about the applet in this comment on Paula Burch's Dye Forum. Namely, that the "mix amount" numbers only give approximate results. Considering the samples that didn't match, that's a relief. I also learned that those numbers don't indicate any specific unit of measurement. That's not a relief, that's a "now what."

The CMYK (cyan, magenta, yellow, & key black) color model that I'm trying to use, is for printing inks on paper. The numbers stand for percentages of each color. So for computer color #0025B9 (above), the printer would reproduce it on paper by using 100% cyan, 80% magenta, and 28% black inks. Obviously this cannot be translated directly into proportions for a dye recipe, because then I would have to be working with 208%.

For my first experiment I've decided to calculate these as percentages of their total 208. This gives me:

Cyan (for which I will substitute turquoise) - 48%
Magenta (for which I'll substitute fuchsia) - 38%
Black - 14%

Which lead me to another question, what unit of measurement? 48 teaspoons (or even 48 eighth teaspoons), is out of the question here in terms of dye amounts. And from what I have learned from the DyersLIST archives, Procion MX dyes do not all weigh the same anyway. This means that a teaspoon of one color doesn't necessarily weigh the same as a teaspoon of another color. So, kitchen measurements are definitely out and metric measurements are definitely in.

The next question is on mixing. Do I mix the individual powders to make one dye stock, or do I make up individual dye stocks and then mix them? Considering that I will likely need to do quite a bit of sampling, mixing individual dye stocks makes more sense. Plus, it is safer to work with the dyes in solution rather than in powdered form. True, solutions won't keep all that long, but without the activator (soda ash), I should be able to use them up quickly enough.

And then there's the question of how well fuchsia will substitute for magenta and and turquoise for cyan. Also, which yellow is best? And what about black? There is no true Procion MX black as all the various blacks are mixtures of other colors. I think only experimentation can answer this for me.

Here's my plan:

1. Measure mini-skeins of 12.5 grams each as samples. Based on the yards per pound, I have calculated that my yarn has 5.55 yards per gram, so I will need to measure out 69 and 1/3 yards per mini-skein.
2. Mix up 1% stock solutions for the individual colors and mix the colors based on the percentages I calculated. I'm doing 1% instead of the usual 2% as it will be easier to work with the hex code numbers this way.
3. Start with a dark depth of shade (dos), since the color #0025B9 (above) is a fairly deep blue. I will make adjustments from there.

Here are my calculations (from this pdf help sheet from the University of Hertfordshire):

% shade ÷ % dye stock x WOF (dry weight of fabric) = total mls of dye stock 4% ÷ 1% x 12.5 gms = 50 mls total dye stock to add to dye bath

So for each color I need:

• Turquoise: 48% of 25 ml = 24 ml
• Fuchsia: 38% of 25 ml = 19 ml
• Black: 14% of 25 ml - 7 ml

One last question I have is based on some of the research I've been doing. That the depth of hue can vary amongst the various Procion MX colors, even if the same weight /amount of dye is being used. I hope I'm saying that right. I found this as a question in the DyersLIST archives, but with no clear answer. I've seen this somewhat with my own experiments: I got a bright yellow but a softer red, even though I used the same amount of dye powder for each color. I got a red more equal in intensity to the yellow by doubling the amount of dye powder. Am I making sense? I'm probably not using all these terms accurately, but I hope you get my drift.

Anyway, I don't have any concrete facts to work with here, so I'm not going to worry about for now. I figure that I can't correct my course if I don't begin the journey.

I have absolutely no idea where I will end up colorwise, but at least I'm making a start.

Next .... Computer Hex Code Dyeing 3: 1st Samples - No Joy

Related Posts:
Dye Recipes From Computer Color Codes: A Theory
Computer Hex Code Dyeing 1: Palettes From Pictures
Computer Hex Code Dyeing 3: 1st Samples - No Joy
Computer Hex Code Dyeing 4: 2nd Samples - Getting Closer
Computer Hex Code Dyeing 5: 1st Success?

Originally published July 9, 2008

© by Leigh at https://leighsfiberjournal.blogspot.com/

Computer Hex Code Dyeing 1: Palettes From Pictures

The first step in testing my theory, was to choose some colors to work with. Rather than use a computer image, I decided to start with hard copy picture. There are a couple of reasons for this. One is what Jane mentioned in the comments to that post; that computer monitors do not display colors accurately. Even though computer technology has advanced beyond the traditional "web safe" colors, computer color vision just isn't that accurate. (For an interesting article on this, "Is Your Computer Color Blind?," click here for )

The other reason is that it will be easier to compare and share my results.

The photo I chose was from an old Audubon Society calendar. My first problem was getting it into the computer without a whole lot of color distortion. I quickly discovered that my scanner couldn't do it. [Note: I have an HP OfficeJet 5600 series All-in-One. The scanner stinks. The best scanner I ever had was an Epson Perfection.]

I discovered that photographing the picture in natural light rendered the best color reproduction after inputting to the computer.

There are a number of ways to create a color palette from a photo. I could have done it one color at a time at a site like LunaPic, or with my photo editing software's color picker tool. However, I wanted to show you two websites which automatically create palettes for free.

First at Color Hunter. All I had to do was upload the photo and click on "upload image."

You can click on these screenshots to enlarge them. Color Hunter offers two options for their color palettes, vibrant (above), and dull (below.)

Both seem to coordinate with the photo quite well.

Another option is Big Huge Labs Palette Generator:

The Big Huge Labs palette is closer to Color Hunter's dull palette, though the hex code numbers are not the same. And considering the number of colored pixels there are to choose from in any photo, this isn't surprising.

I arbitrarily chose the Color Hunter vibrant palette to work with first. I picked three colors and wrote down their hex codes. Then I went to the Color Converter and plugged in those codes to get the CMYK proportions. (Like I did in this photo from my previous post.)

I rounded the proportions to hundredths, because that's what my own photo editing software does.

So, here are the colors I've chosen to start experimenting with:

#0025B9
cyan - 1.0
magenta - 0.8
black - 0.28


#1470C4 cyan - 0.91
magenta - 0.5
black - 0.11


#EE9628 magenta - 0.37
yellow - 0.83
black - 0.07


The next step will be to try to figure out how to translate these numbers into recipes. Since the test run with the Dye Mixer Applet wasn't all that accurate .....

...... I have several doubts and a lot of questions.

Next ..... Computer Hex Code Dyeing 2: Wrestling With Recipes

Related post: Dye Recipes From Computer Color Codes: A Theory
Computer Hex Code Dyeing 2: Wrestling With Recipes
Computer Hex Code Dyeing 3: 1st Samples - No Joy
Computer Hex Code Dyeing 4: 2nd Samples - Getting Closer
Computer Hex Code Dyeing 5: 1st Success?

Originally published July 7, 2008

© by Leigh at https://leighsfiberjournal.blogspot.com/

Dye Recipes From Computer Color Codes: A Theory

Creating color schemes from photos is one way in which fiber and textile artists find inspiration for their work. Kris's Color Stripes blog is a wonderful example of color palettes that can be obtained from photos.

What I've been wondering however, is how to duplicate the colors from my own palettes onto fibers, fabrics, and yarns. I've been thinking that there must be a way to create dye formulae to match colors from a photo. After a lot of poking around on the internet, I've come up with an idea which I plan to try soon. Here it is in a nutshell.

Computers use what is known as hexadecimal codes to interpret color. These are based on the RGB (red, green, blue) color model which is the same model used to mix colored lights. Why RGB instead of the traditional artists' RYB (red, yellow, blue)? Because devices like digital cameras, scanners, and computers interpret colors like lights rather than paints.

Hexidecimal (or hex) codes use values from 0 to 9 or A to F and look like this: #40E0D0 or #EE82EE. The first two places refer to red, the middle two to green, and the last two to blue values. A hash mark always precedes them. By changing the codes, a web designer can determine the colors for background, text, links, etc on a web page. Several websites have charts of these codes, such as this one at december.com, or this one at web.forret.com.

Obviously dyes don't work like lights, so potential recipes would have to be interpreted in a different model. For my experiments, I'm going to try the CMYK (cyan, magenta, yellow, key or black) model. Why? Well for one thing, according to my experiments, dyes don't mix like paints do. For another, it's easier to convert hex codes to CMYK codes rather than RYB codes.

Most photo editing software can pick out individual colors and give you the hex code. Many of them (like The GIMP) can give you the CMYK code too. But you don't need your own editing software to pick colors from your own photos. There are many online editors that can do it for you.

I found the one below at LunaPic.

[NOTE: You can click on any of these photos to enlarge.]

I had the choice of either using their free photo, or uploading my own. Wherever I click on the photo, the color picker isolates that color and displays it at the top, along with it's hex code.

Next, I need to convert the hex code to a CMYK code. This can be done Peter Forret's online Color Converter.

I type in the hex code at the bottom and click the RGB -> CM.... button. At the top, the color displays as several swatches. Under that, the CMYK and RGB codes are given.

My theory is that by substituting fuchsia for magenta, and turquoise for cyan, those CMYK values should give me proportions for my dye recipe. But how to test? With Olli Niemitalo's free, online Dye Mixer Applet (which is also available for downloading).

There is a column on the left side of the dye mixer applet entitled "Dye." If I click on any "blank," I get a drop down menu with a number of different types of dyes and colors to choose from, including several Procion MX dyes. The choices are somewhat limited, but hopefully I can get an idea of whether or not I am on track. A color swatch for each individual color is displayed at the right.

The "Mix Amount" column in the center lets me adjust dye amounts with either a slider, or by plugging in actual values. I input the CMYK values from the Color Converter. At the bottom of the applet, I can see the mixed swatch.

How well does the mixed swatch match the original chip from LunaPic? It's not exact. There might be several reasons for this.

One thing I notice is that the color swatches on the right for the individual dyes display lighter hues than the CMYK key on the color converter. I can note too, that my mixed swatch on the applet, appears to be a lighter hue than the original color I'm trying to match.

Another problem is the dye color choices available in the dye mixer applet. There are two reds and four yellows to choose from, and none of them matches the CYMK key. However! The mixed swatch is enough in the ball park that I am encouraged to take this a step further and try some dye experiments. Looks like I've got my work cut out for me, doesn't it?

Next ..... Computer Hex Code Dyeing 1: Palettes From Pictures

Related Posts: Summary of Procion MX Dye Experiments - Learning how dye colors work together Computer Hex Code Dyeing 1: Palettes From Pictures - using free online tools
Computer Hex Code Dyeing 2: Wrestling With Recipes
Computer Hex Code Dyeing 3: 1st Samples - No Joy - but a starting point
Computer Hex Code Dyeing 4: 2nd Samples - Getting Closer
Computer Hex Code Dyeing 5: 1st Success? - achieving the closest match so far

Originally published July 1, 2008

© by Leigh at https://leighsfiberjournal.blogspot.com/