For the past few weeks, i've heen refining my screen printing of process color. While overall things were going predictable, i ran across a problem of snow flaking. This is where the process color has a tiny hole which allows the white underbase poke straight though.
To overcome this problem, rather than single stroking the squeegee, i now take 2 back to back strokes without any flash in between. Because 2 strokes transfer more ink, in photoshop i modified the transfer function back 20-30% to compensate for the second squeegee action. (see attached)
I'm now starting to get quite pleased with my results. The images attached are my personal test files and are not anything i ever plan to sell. Each was selected for the color content characteristics along with my personal desire to wear "the coolest" shirts. The photos are taken from actual shirts. Only the level was adjusted.
To better bring out captain Spalding, i printed the magenta screen last. The black background was printed in superblack using 100% of cyan, yellow, magenta and black, rather than just 100% black.
The dali lama is the image i used 2 squeegee strokes and cutback the inks to compensate. If anything, i might have pulled the color back a little too much.
Both images were done at a 55 LPI screen.
The film was output on a laser printer using kimodesk transparencies
Union ink was used for the cyan, magenta, yellow and black
High pigment black was used rather than a process black
High pigment white was used for an underbase.
The white underbase was printed with a 110 line screen
The CMYK was printed with a 305 line mesh.
The exposures were done with a mercury light and vacuum frame
This technique should mean that the person making the shirt does not need to be 100% perfect in order to get a consistent result.
Re: Calibrated double stoke to eliminate snowflakes
Thanks for the positive comments. While i've got a lot of prior experience working with process color, these 2 images are only 5th and 6th process color t-shirt i have ever made. Were still using the back porch to wash out and reclaim screens.
The one thing which is key to getting good halftones which includes process color is a good exposure unit. Mine is an old used Violux which i picked up on ebay which included a vacuum frame which i modified to work with screens. Including the new $150 bulb, i've probably got $500 in the exposure setup.
Other than my kickass exposure setup, everything else i have isn't anything special. We still don't have a washout room, so all of the wash out and reclaiming of screens is being done outdoors on the porch. Process color requires a minimum of 4 units and in my testing, the white underbase is required for all shirts, even white. This means i'm using 5 screens.
This file was opened in photoshop where i performed a couple of sharpen on an unsharp mask filters to sharpen up the detail in anticipation of some loss in detail resulting from halftone printing. The background was already masked, so all i needed to do was select it and delete it. Then i told photoshop to convert it to a CMYK image using a null set of conversion numbers in the "color settings". When i convert the image RGB-> CMYK i see no change in the image on the screen. If you see a change, you don't have neutral settings. Most likely press gain has a value which needs to be zeroed.
Once again i used the magic wand to select the background, but then i hit control-shift-i to reverse the selection, then under channels i made a new alpha channel which i fill the shape with black which will become my white primer channel. To trap the white i once again select the inverse and select->modify->expand the selection a 5-15 pixels then fill the outside with white to clear off the trap area.
From this point i simply select print with preview and select to print separations, set the transfer function, set the screen frequency, set the printer to print high res b&w and click print. The output was printint in kimodesk laster transparencies. For the sake of testing consistency, i continue to use my Phaser 7700 color laser to print the film, but i recently picked up a HP 8150n tabloid size off ebay for $175 (plus 200 new toner) which looks like it will do just as well. I've also made films with inkjets and see no reason that if the proper halftone dot structure is printed, that inkjet would not also work equally as well.
From finding the image till having a full set of films in my hand took less than 10 minutes. The only real bit of magic is the transfer function, which once calculated will remain more or less a constant.
Expose, washout, register, tape and ink just like any other job.
Process color is something to be respected, not feared. Im glad to share whatever information i can. I've still got a long way to go myself.
Re: Calibrated double stoke to eliminate snowflakes
So you are not using any of these expensive RIP programs to get your halftones. You just use photoshop? I have treid to get proper halftones but never tried the instructions you just set out for us. I use an inkjet printer , but it is not postscript 3. So Can I still get proper halftones using a non postscript 3 printer folowing your directions in photoshop? I would like to experiment with process but dont want to waste my time making crap transparencies if you can teach me to do it properly instead.
Re: Calibrated double stoke to eliminate snowflakes
Do you have any larger or closeup pictures? I've yet to dabble in process printing, as I've always thought it best left to an auto press (consistency) and I'd like to see what I should be aiming for.
Re: Calibrated double stoke to eliminate snowflakes
Quote:
Originally Posted by majesticmind
So Can I still get proper halftones using a non postscript 3 printer folowing your directions in photoshop?
Most inkjet printers resist printing halftones dots, but instead us an internal dither pattern which creates a continuous tone. To get around this driver limitation, the image needs to be sent as a b&w bitmap.
For single color it is easy, simply change the image into a bitmap and set a halftone value. 55LPI is as high as reasonable with a 305 mesh.
There is a way to produce all 4 color separations, but it requires a big fast computers. Switch the graphic to a cmyk version of the graphic, then size the graphic to 1/2 the printers resolution. For example,my epson is 1440, so i'd use 720. If the computer is really really fast, you can use the full 1440.
Once you have the high res CMYK image, apply the pixelate filter, "color halftone". Start with a number like 10. From this simply goto the cyan channel, convert to bitmap, print, undo bitmap, goto the magenta channel, convert to bitmap, print, undo convert to bitmap, repeat repeat.
The color halftone filter produces a CMYK separation. When converting to bitmap, use diffusion dither. (i guess). As long as the graphic is an even unit of measure as the printers DPI, it shoud work out close to what a rip would do.
Postscript laser printers make great output devices. They 11x17 ones weight so much, that they are often sold locally "for pickup only" at very low prices. Rather than buying a rip for an inkjet, i'd suggest upgrading to a postscript laser.
Re: Calibrated double stoke to eliminate snowflakes
Quote:
Originally Posted by ImageIt
Process color requires a minimum of 4 units and in my testing, the white underbase is required for all shirts, even white. This means i'm using 5 screens.
How is it possible to get smooth or no hand feel on the shirt when you're using that many layers of paint?
Re: Calibrated double stoke to eliminate snowflakes
Quote:
Originally Posted by Zsoltee
How is it possible to get smooth or no hand feel on the shirt when you're using that many layers of paint?
It is impossible to get a new hand feel when printing process onto black shirt because a solid layer of white must be laid down first. The hand is primarily created on the first white pass. Once the white has been laid down, the hand doesn't noticeably change when the CMYK inks are added.
The CMYK colors are calibrated to do 2 strokes, so the 2 strokes added together does not lay down any more ink than one uncalibrated stroke. It would be like trying to paint a white room with dark paint. You can actually use less paint when using 2 thin coats of the dark to cover the light than 1 heavy coat. If you take a look at the transfer curve attachment, you will see that i've cut back the ink transfer nearly 50%. Also in that transfer curve on the left side, i pumped up the fine highlight screens 5-10%, then at the right side, at 90% i take a quick ramp up to 100% to ensure any shadow area that is near 100% will get 100%.
Process color requires a white underbase. The less hand this can be printed with, the less hand the final result will have. Currently most of our efforts are concentrated on the color itself, but we have planned testing in the future to address the white. For example, we have just purchased a drill attachment to stir ink, so with a mechanical stir we want to see if we an use a finer mesh screen. Yesterday we used the mixer to stir a clear ink which when we opened the bucket look more like jello than an ink. After stirring, it looked smooth and creamy.
My current round of testing is to go though the sample files included with adobe illustrator. Anyone with a current version of illustrator has a vector original of the Bedlam Ballroom illustration which i've attached a photograph of a completed t-shirt. This is the shirt we used the clear coat. WOW!, did a clear coat really make a big difference on a design like this! The glossy did cause more than a few issues trying to take a photo of the shirt.
Today we are testing union triple pigment process ink rather than the union standard process. We also are going to try a finer mesh screen to produce the white underbase. From those tests, we plan to continue on through the Adobe Illustrator sample folder.
Re: Calibrated double stoke to eliminate snowflakes
When you converted the design into CMYK did you have to adjust any of the angles on each color? If so did you use the diamond option on the print option?
Re: Calibrated double stoke to eliminate snowflakes
I am currently using the adobe default process color angles for an "Auto..." 55 line screen. Adobe screens are optimized for output on a square pixel rendering device which prevents moires within the screen. These same screens seem to do well when placed on a screen with a similar configuration of mesh.
Generally i have been using an elliptical dot, but i doubt that switching to a diamond would have a noticeable difference in output. Comparing spot functions would be another interesting test.
The result of todays test on unions triple pigment process didn't do very well when printed on the white primer, but when printed on a white shirt, the triple pigment could print a white shirt without a mask almost identical to the other process ink onto a white mask. If i only want to make one set of screens and print both dark shirts with masks and white shirts without a mask, it looks that switching inks is all it would take.
Mechanical stirring seems to be the answer to preparing inks. Wish just a couple minutes with a drill, 1 gallon of ink is changes from a consistency of vasoline into one more like honey. We picked up a half dozen mixers hand drilled holes in the lids of half dozen inks. No cleanup is required. Just take the lid off and stir. We also picked up a few spares to mix and wash as needed.
Tomorrow we plan to have a shootout between several different whites to try to determine which white gives the best base.
Re: Calibrated double stoke to eliminate snowflakes
Hi, I'm new to the forums but old salt to t-shirt printing, and I've been experimenting with 4color process for a number of years now, and having a bit of success with it as of late. I'd be happy to share any tips I may have.
One trick is printing all line dot at 65lpi on 300 or 310 screens, as well all are at a 61 degree angle, and set it to ellipse. (tricky isn't it) Try it, you'll like it, works great on a wide range of colors. (Maroon at the moment is my bane) keep those secretes handy, see if they work for you, do work for me. I have not tried any other dot settings as far as diamond or line, I don't know what or how it would moray using those with certain screens.
Re: Calibrated double stoke to eliminate snowflakes
Quote:
Originally Posted by ImageIt
This is where the process color has a tiny hole which allows the white underbase poke straight though.
To overcome this problem, rather than single stroking the squeegee, i now take 2 back to back strokes without any flash in between. Because 2 strokes transfer more ink, in photoshop i modified the transfer function back 20-30% to compensate for the second squeegee action. (see attached)
Fred-
I have yet to take the plunge into process printing, but i have been putting my setup to the test with halftone production by printing greyscale. It's been going pretty well.
My problem is printing white onto dark garments. I am getting a fair amount of dot gain with the p/f/p method; obviously one pass doesn't pop enough.
I feel like this transfer function in photoshop's output may help me out, based on your post here, but i am not sure how it works. When i go into the transfer values screen, all the values are blank. I have no clue how you came up with the numbers you did.
But anyway if there is a way to control my dot sizes to accommodate for the dot gain on press i would love to hear about it. I'd rather not burn a second screen and run a black or clear ink.
I am using Photoshop to Corel for postscript settings to GS View/ Script to print on "out of the box" R1800.
Calibrated double stoke to eliminate snowflakes 
The photos are taken from actual shirts. Only the level was adjusted.
