INDEX Vrs Stochastic halftones.

[Dottonedan]

I posted so much about this subject in another group, that I had to copy it and post it here for my own keepsake.

A bit about a couple more unique methods used in screen printing pertaining to (dots) and the various types of output.

This is a subject about SQUARE dots. There are two main methods to achieve and use square dots.  Why use square you  might ask? The goal in these cases where you would use this method, is not really to intended to use SQUARE dots, but for the benefit of using the SAME SIZE dot. This method puts out the same size of the square dot across the board in the separation for the fact that you are litteraly using the file and it’s resolution...to output the shapes. Traditional dots vary in size and represent tone. Smaller dots = lighter percentages of color. Larger dots represent a more consistent range of the color until it becomes a solid. In these two methods we are avoiding the transitional phase of going from a 1% dot on up into the 49% dot and the cross over between the 49% positive and the 51% negative shapes that the traditional dot makes up. With the square dot method, you reach tone by spreading the square dots further and further apart.

What two types of square dot methods are there?

Stochastic and Index.
Both of these are resolution dependant. Meaning, you get different quality results based on the files own resolution...and not the resolution of the imager (film printer or CTS printer).

Stochastic generally comes from a RIP conversion, but the shape of dots and the need for considering the files resolution is the same.  The end result of the dot is exactly the same square dots (all square dots end up coming out at the same size) and how they represent TONE is by the spacing. Similar to hand stippling. (old school art and seps). You can think of Stochastic as more of a true blending of colors like in pointillism. In a pointillism painting, the dots or dabs of paint are often physically bended over and under each other to help form the color blending....while INDEX is visually more like computer generated plotting of colors placing the squares of color side by side. Not over or under.
Therefore, much of the blending is due to the files resolution (Higher resolution means smaller dots...and the need for higher mesh to hold those smaller dots. As a result, there is a thinner ink lay down but at the sacrifice of color PUNCH. Meaning, it’s thinner, less opaque. Can seem more washed out over a base white...and can disappear when printed with no base.  With higher resolution comes smaller dots. (Smaller is better for image resolution). Think of it like a 72ppi web images...and a 300ppi image for magazine printing.

Stochastic, is identical to sim process (blending colors together, running one under and another over colors to create other colors (like 4 color process does, but with more specific colors that might not othersie be acheived with CMYK. In this case, we can use what we loosely call "spot color" inks).  That term SPOT COLOR stems from old school terminology for printing a solid area of custom color ink. You put it in this spot, or that spot. The only difference between the stochastic method and sim process method is the shape and size of the dot. The stochastic dot is also file resolution dependent (all come out at the same size, just spaced differently). The benefit of stochastic is that since it blends, you do not need (as many) colors as with indexing...and your eye is less needed for the blending purposes. You physically blend on press with stochastic while index requires both more colors...and the human eye to be tricked into thinking that a red dot place next to a black, and blue and yellow dot, create brown.

A benefit of index or stochastic is that if you can hold one of the small dots in the screen, you can hold them all.

When doing this (attempting to simulate Stochastic) via indexing or diffusion dither, this process is...of a lesser randomly generated quality...than if you were to use a RIP that offers the option for Stochastic. If you attempt to do the Stochasic via a lower costing RIP such as AccuRip.  It does it, but it’s not that great. The results from Photoshop are better.

Index...Is a method using multiple colors of squares placed side by side. It’s the lesser of the two square dot methods (IMO). Index, often requires adding in more spot colors to assist with building the full shades of the image. A Dark blue, a medium blue and a light blue for example rather than with Stochastic. With the Stochastic method, you can mix a single blue with white under base and top white..to create a medium and lighter blue. Via indexing...the colors can be hand assigned, or randomly assigned based on the art file color content and the methods used during the indexing conversion. INDEX was and is, intended for reducing color or breaking an image down so that it is easier to transfer from computer to computer or web image to web image. Some of the files you might download that won't do the same things your jpg does...is INDEX. The great thing about the indexing method is that you don’t have to be a color separation guru to achieve separating a very difficult job. It’s very plug and play.

To test this out, open a full color image that you want to reduce to 6-8 colors..and go to IMAGE, MODE, INDEX COLOR.
In the Palette, choose CUSTOM, and pick only as many colors as you want to print with (referencing the MAIN colors in the art. Under OPTIONS, choose DIFFUSION and AMOUNT start off with  something near 80 (this depends on the art). Play around with it.
I don’t use this method or haven’t for over a decade, so you will need to try the various options....and then make selection using the color picker...and then while still selected, bring those over into a new spot channel in the channel window. It can be very handy in a pinch for some of the newer people trying out separations.

There is diffusion dither, (random) placement meaning no pattern, and there is indexing by way of pattern. The pattern does NOT work as well. Patterns within patterns (mesh) cause further issues such as mesh interference and moire..

As a side note, do not confuse this RESOLUTION with LPI or (Lines of dots Per Inch). LPI is not in this picture. It has NOTHING to do with indexing or stochastic since there is no line count of dots at all.

I did this Cobra using Stochastic (similar to index) but again, stochastic is like sim process and blends. I am using 5 colors for a 6 color press allowing for the flash so that I can have it printed in most any shop. I’ve had this customer for over 13 years repeating the same art. I did this via Photoshop due to not having a higher end RIP at the time... that will produce better stochastic random placement. The algorithms (math to do so) via Photoshop are not as high as what you can get from a good RIP. I did this cobra in stochastic for the benefits of capturing the more subtle details and color changes within the original painting on poster board.  You will see in the next picture, an example of the square dots close up and the files resolution and in the picture below, is the actual printed result. The snakes head is about 5” wide. We run it on either a 280 in a pinch (since this resolution works well on the 280 also) or we run it on the intended 305 mesh. Works well every time.

[Lizard]

Good stuff Dan

[Sbrem]

We used index a number of times in the '90's, but got away from it. The last time I went with it, it wasn't very good, so we did the job over with traditional halftones, but I did generally like the "look" it gave some designs. A press needs to have pretty tight register to get the best results.

Steve

[Dottonedan]

With indexing, yes.  The press needs dead on reg for it to work well (and more of the colors) along with a higher resolution of small dots to help form a seamless photo print.  Similar to the old method of knocking out halftones from each other.  One guy today, has brought that method back, but I still see the same issue with it. I don’t really have a problem with anyone doing the method of knocking out the dots from each sep, for one of the benefits claimed to be the result of having only one layer of ink and a softer hand.


I do see some flaws in the method similar to what we see with indexing. On press reg.  Knocking out the dots from each other also takes away from the blending that happens on press. He pushed this by claiming that traditional overlapping halftones don’t make correct separations. (That, is obviously based on how the seps are done). A correct method of separations are done by pushing just enough of one color ...under or over another in the right areas to cause your intended end result.


In his example of a poor outcome, he doesn’t do that (like as if all of us who separate using traditional methods) do our separations in his poor quality method and do not push ink further under or over another. This then aids in supporting his claim a bit further toward his promotion of this old technique that to some, (is a new invented process). For him, it’s just another “product) he can offer. Like some of his other separation “products”. It can be done, but it’s not like the invention of sliced bread. It’s an old technique for a reason IMO. Like indexing, it needs to be exact or the intention in areas are off in color. Some gaps of line work etc. then due to the fact that it’s not actually bending well if you do have it in perfect reg, leans more towards requiring to use more colors. There is less true blending...and requires the eye to do more of the blending.  But what do I know. He’s closer to being a millionaire now. I’m not.

[Dottonedan]

Some my best printed results were at Disney, where we printed very small dots in stochastic form with 10-14 colors. This type of printing combined with no real short limits on the number of colors, and wet on wet, allowed for those dots to sort of mash together on press, forming more of a photo real print result. Less dots can be seen and more of the true blending is seen as a result. Like continuous tone.  Loved those days.

[Sbrem]

Some my best printed results were at Disney, where we printed very small dots in stochastic form with 10-14 colors. This type of printing combined with no real short limits on the number of colors, and wet on wet, allowed for those dots to sort of mash together on press, forming more of a photo real print result. Less dots can be seen and more of the true blending is seen as a result. Like continuous tone.  Loved those days.

I have a Grumpy sweatshirt that is exactly this, including some gel lettering. It is without a doubt, the most commented on shirt I own, complete strangers comment on it. It's 10 years old at least, and been washed a good hundred times. I'll have to take a shot of it later and post it, to see if it's one of your's Dan...

Steve

[Orion]

Played with four color process random dot back in the '90's. Easy to control but images just were not as sharp as traditional halftones.  We did use it one more time for a discharge underbase simproc job. We were having issues with our Union white plasticharge base, drying up in the lower percentages. We changed the base plate to a 156t and employed random dots in the 170-180 resolution range. The images are from a vintage shirt site. This was from my time at Cotton Cargo Tour Merchandise.

[Dottonedan]

Orion,  that's awesomeness right there!.  Great work!




SBrem,  It would have to had been at least 13 years old for it to have been one of mine.  I got kicked to the curb in 2009. lol.

[Sbrem]

Could be, 10 - 20 years is all the same now...

Steve

Here it is

[Dottonedan]

Could be, 10 - 20 years is all the same now...

Steve

Here it is

HAHA!  Yep.  thats one of them.  One of the last groupings in a program I worked on. We did Grumpy & Sleepy then Tinkerbell and then the fab 5. Being Mickey, Donald, Goofy, and of course Minnie & Daisy.  Might have been Fortune Fashions that had printed it at the time.

[tonypep]

Yes we printed that at FF in Cali. Seem to remember an HD gel plate on that one

[Sbrem]

Yes we printed that at FF in Cali. Seem to remember an HD gel plate on that one

Yes, 2 of them, a light gray, and a dark gray. Every time I wear it I get a comment in a store, just a great print and concept. Great work you two!

Steve