Arahne's color shading - history
How did it all start?
ITMA 95 in Milan, Italy was our first international show. At that time, our product was rather crude, and so was our marketing. To attract people to our stand, we had a panel reading:
It reads FEAR in vertical, but nobody noticed the pun. A woman pointed at the sign saying: "I like that!" and sat on a chair. She seemed to be experienced in the matter, and went straight for the kill. She asked how can we make a jacquard picture from a full color photo, without any color reduction. I said - no - that is impossible. You have to reduce the number of colors, then choose a weave for each color, and then apply the weaves. As if I was telling it to the regular idiot, who doesn't understand jacquard. She objected, that we are losing the fine details and transitions by doing that. I agreed, and proposed that you should use some dithering while reducing the colors, and with appropriate choice of colors, you still keep the shades. Then, my program crashed for the second time. It was the first day of the show, and our configuration was not entirely stable. She uttered "I don't like that!" and walked away without saying another word, or leaving a business card.
Besides giving us a lesson on importance of the reliability, she has also questioned the traditional way of doing jacquard. It took us couple of years find the solution. We would get there faster, but in a business environment, you can not just concentrate your work on interesting problems. You need to generate income with more mundane tasks, and work on the interesting stuff in during your free time.
All these images were reduced for your viewing pleasure. Click on the image to see the full size version. And do not forget to click back on your browser, once you are done.
We wanted to have this feature ready for ITMA 99 in Paris, to distinguish us from the crowd of the other CAD producers. So we started experimenting with it already in 1998. Students at University of Ljubljana make a couple of projects, and they were woven on the TIS jacquard equipped with 2200 hooks.
Initially, we tried a taste of our own medicine: Reduce the colors in the image to a fixed palette using dithering, and carefully choose a weave for each color. We were working with fixed palettes of 3x3x3=27 colors and 5x5x5=125 colors, trying to cover the primary colors. On the right side of the fabric, you can see a rather pathetic attempt of a 125 color palette.
Results look promising, but leave a lot to be desired on close inspection. There must be a better way. Unfortunately, it involved full dedication of Arahne's staff, and our TO-DO list before ITMA 99 was just too long.
We have solved the problem during the post-ITMA depression on December 1999.
As with the invention of photography, when you implement the right idea, the result is immediately convincing. But we can not say the same about the weaving quality of the cloth. We could never show this to a potential customer. They would say - yeah, but why are you weaving it at University? Don't you have any real customers?
In a world where everyone is busy copying each other, it is very difficult to find someone, who wants to try something really new.
We succeeded in the Summer of 2000, when company G.&M. Gunetti agreed to weave us the sample.
Needles to say, the results was disappointing. It looks vaguely familiar to the sample from our color shading tutorial. But the picture seems Warholized®? Sky is yellow, the ruins are orange, and the olives are inundated by the sea?
The answer to the riddle lies here: You have to put in your weft colors in the correct order.
Fall 2000, another attempt. This time, we have hit the correct sequence of weft colors. Some of you might even recognize the picture. But the colors are all flat, there is no depth to the picture. The crows and the upper part of the sky should be black. Un orage se prepare. Do you feel suicidal already? Don't despair, read on!
The mystery reveals itself on close inspection. Our shading weaves rely both on warp and on weft effects. Warp must be able to cover wefts almost completely. In our sample, weft effects are dominant. Hence - black and white warp do not contribute much to the picture. Conclusion: use thicker yarns for the warp, or increase warp density. Changing the warp density is a definite no-no on jacquard, so we have to fix the yarn count.
Spring 2001. Getting better and better. This time, we do not need to explain the subject of the picture. And yet - something is still wrong.
Yeah, penguins are fat. But this fat? Not really. The image was designed for the density of 75 picks per cm, and only 50 would enter during weaving.
Summer of 2002: Jackpot! Yarn counts, density, warp and weft sequence are all correct.
All that remains is to zoom in a little, so you see that it is not a fake.
We have actually underestimated the power of this technique. The resolution of the starting images was too low, and you can see the pixels on the woven fabric.
You may wonder why did it take us so long to fix those weaving parameters? A loom with 8192 hook jacquard costs mucho dinero, and needs to produce to pay itself back. The black and white warp, which we used, is completely different from the regular production (although the density is the same). The jacquard design, which we prepared, was usually waiting a lot of time for its time slot on a busy loom.
There is a positive aspect of all this trouble: we have simplified the program to the the point, where a complete layman in weaving can prepare the shadings. Load the full color (JPEG) picture, select yarn colors (if you think that you can improve upon the default primary colors), and click OK. It will even calculate the dimensions and scale the image on the fly, to adapt to the weaving density.
Zooming in on the prey...
Here's looking at you, kid!
And last, but not least, a big thanks to everybody, who has helped us in making all this happen:
- The anonymous woman for wanting "the impossible" at ITMA 95 in Milan
- Department of Textiles of University of Ljubljana for weaving the first two samples
- Riccardo Sattier for introducing us to Franco Gunetti, and for Venice & Bambi samples
- Franco Gunetti, the owner of G.&M. Gunetti, Chieri, Italy, for weaving all but the first two samples on this page. He is one of the few weaves by choice and conviction, not just by profession.