screen printing > 4 Color and Simulated Process Printing

halftone measurment results. WOW, dot gain galore!

<< < (2/4) > >>

yorkie:
Let me try to explain the "zen" of dot gain.

Start off by clearing your mind and listen to the press. Imagine the press as a stack of speakers and press calibration is the mixing board. If the speakers are in a small room, the dials need to be tuned down, where in a large room, the dials turned up. If the walls are marble, then some frequencies need to be turned up or down, verses if the walls are carpeted. The speakers (the press) does not get to choose the conditions it will be playing (screen lpi, textile fabric, squeegee pressure, ink viscosity...).

The way to calibrate a press is by starting off by setting all of the knobs the the middle position. In the case of squeegee pressure, find a nice middle setting, that should all else fail, you still can adjust the darkness to be a little darker or a little lighter. If everything else described works perfectly, the pressure setting of the press should never change.

The postscript language used in RIP's (Raster Image Processors) was actually designed by printers!!! The computer screen should display a reasonable representation of the data in the graphic file, but when output to film, the image is transformed to a value which is correct for the press. The simplest example is negative film. This is simply a reversal of the calibration values.

The actual calibration values are very simple. 0.0 represents white, 1.0 represents black, .1 is 10%, .2 is 20%... this gives an array of numbers

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

If the 50% screen is actually 70%, then the value in the array would need to be a .3.

0.0 0.1 0.2 0.3 0.4 0.3 0.6 0.7 0.8 0.9 1.0

Inside photoshop, these values are expressed as a graph.

These basic values are inherent in postscript. There are multiple ways to access these values. The value can be adjusted in photoshop (as attached), these values can be invoked via color correction adjustments in the artwork, the values can be embedded in a .ppd file and accessed through a print driver, the values can be set within the rip... Any and all of these methods can either act cumulatively or absolution, depending on context. In simple words, changing a .5 to a .3, could be multiplied by a .97% somewhere else such as color correction via color space.

Quite often the simplest way to control press gain, is to turn off all other forms of gain control, except the one you control. Don't color manage the file!

For my purposes, photoshop is the easiest place to control the calibration curve. various curves can be stored is setup files to photoshop.

The other easy place to control the calibration curve is via the rip, but typically to control the calibration curve in the rip, is by setting up multiple print queues. Different print queues can define different mesh/lpi/ink combinations. This can mean a LOT of queues, but with good naming is manageable.

Beyond the numbers exists reality. A 5% screen if very difficult to print, so calibration files often bump the low end. T-shirts also look great with "BIG COLOR", the heck with correctness, we want it popping! The translates to a smaller spectrum of tone being used to represent the entire spectrum. So long as in a fountain fill, the dots make a smooth transition, more is often better. :\

The place to begin controlling for gain, is by printing a test pattern, which includes a graduated screen and individual percentage blocks. Turn off all control and see where on the graduated pattern the dots first appear and where the dot pattern disappears. Test different press settings to see what gives the widest band of small dots to no dots. For manual printing, test 2 easy on the arms strokes verses what various levels of muscle will produce. Remember, you want to make it easy on yourself and on the machine!

Test a couple exposure settings and be sure the exposure is where it works best. Again, this should be the natural value of the light curing the screen. Where washout is easy, but crisp. A value which produces consistent results. A value which washes out the smallest of dots...

Once the press, exposure and washout are a constant, then adjust those 11 values to produce a piece of film, which yields the desired printed results (correct or purposely incorrect).




drdot:

--- Quote from: Orion on April 20, 2011, 10:35:32 AM ---Pierre, if you printed a 50% dot that gained to a 72%, that is not a 20% gain it is 44% gain. Yep, you need to go back in the rip and adjust the curves. I don't understand the amount of dot gain that you are experiencing in the lower percentages, the numbers aren't making sense to me. How hard are you flooding your screen?

--- End quote ---

This is incorrect. There are two ways to measure dot gain and by far the most common is using the Murray-Davies calculation where dot gain is reported in absolute percentages, not relative. So in this example, the dot gain is 22%. If it were 44%, the 50% dot would measure 94% on the printed surface. I should also point out that there are two different types of dot gain, mechanical (physical) and optical. My reply here  is based on the mechanical component, although optical gain is also simultaneously occurring.

When mechanical dot gain is described, it is always measured at the 50% value using Murray-Davies equation. In order to get an accurate value you must know the starting and ending density values of your ink. This means zero to the value of the print surface and measure the value of the solid ink. You are calibrating based on what a 50% DOT AREA actually ends up. This means you need to measure with a spectro-densitometer if doing sim process or a color reflection densitometer if doing 4 CP.  There are only a few of us crazy enough to invest in these measuring devices as they are all over $1000 and some as much as $4500. 

Dot gain occurs as the perimeter (border) of the dot grows. This is based on Zone Theory (also called Border Theory) based on the mechanical growth of the dot due to ink spread. For any given linecount, the increase in the dot gain will be constant to the perimeter. So, if the gain is .001" (made up example) increase around the perimeter, it will be the same for every percentage of dot. Since the maximum perimeter occurs on the 50% dot, it will have the lowest relative gain. The very small highlight dots can easily double or even triple in value  (eg 5% dot becomes 10% -15%) As the halftone dot gets larger, the amount of gain will decrease as both an absolute and relative percentage.

We can cheat the systems somewhat by changing the dot shape and how the dots connect across the tone range. This is why we use elliptical dots to give us the most control over the tone transitions and to minimize the effects of mechanical dot gain.

Like most things in this business, the vast majority of printers "eyeball" their results by comparing their printed values to some type visual reference. This is fine as a general rule, but the problem comes when different colors (pigments) have different dot gain profiles.  This means the printer must know how each color will behave and adjust accordingly.  This is also one of the main reasons it's  almost impossible to get a dead-on color match when you need to hit a measured color like a Pantone color. It can definitely be done, but it's outside the realm for 99.9% of t shirt printers. Thankfully there are not that many jobs that require this level of accuracy.

Clark:

--- Quote from: drdot on May 05, 2011, 11:50:14 AM ---
--- Quote from: Orion on April 20, 2011, 10:35:32 AM ---Pierre, if you printed a 50% dot that gained to a 72%, that is not a 20% gain it is 44% gain. Yep, you need to go back in the rip and adjust the curves. I don't understand the amount of dot gain that you are experiencing in the lower percentages, the numbers aren't making sense to me. How hard are you flooding your screen?

--- End quote ---

This is incorrect. There are two ways to measure dot gain and by far the most common is using the Murray-Davies calculation where dot gain is reported in absolute percentages, not relative. So in this example, the dot gain is 22%. If it were 44%, the 50% dot would measure 94% on the printed surface. I should also point out that there are two different types of dot gain, mechanical (physical) and optical. My reply here  is based on the mechanical component, although optical gain is also simultaneously occurring.

When mechanical dot gain is described, it is always measured at the 50% value using Murray-Davies equation. In order to get an accurate value you must know the starting and ending density values of your ink. This means zero to the value of the print surface and measure the value of the solid ink. You are calibrating based on what a 50% DOT AREA actually ends up. This means you need to measure with a spectro-densitometer if doing sim process or a color reflection densitometer if doing 4 CP.  There are only a few of us crazy enough to invest in these measuring devices as they are all over $1000 and some as much as $4500. 

Dot gain occurs as the perimeter (border) of the dot grows. This is based on Zone Theory (also called Border Theory) based on the mechanical growth of the dot due to ink spread. For any given linecount, the increase in the dot gain will be constant to the perimeter. So, if the gain is .001" (made up example) increase around the perimeter, it will be the same for every percentage of dot. Since the maximum perimeter occurs on the 50% dot, it will have the lowest relative gain. The very small highlight dots can easily double or even triple in value  (eg 5% dot becomes 10% -15%) As the halftone dot gets larger, the amount of gain will decrease as both an absolute and relative percentage.

We can cheat the systems somewhat by changing the dot shape and how the dots connect across the tone range. This is why we use elliptical dots to give us the most control over the tone transitions and to minimize the effects of mechanical dot gain.

Like most things in this business, the vast majority of printers "eyeball" their results by comparing their printed values to some type visual reference. This is fine as a general rule, but the problem comes when different colors (pigments) have different dot gain profiles.  This means the printer must know how each color will behave and adjust accordingly.  This is also one of the main reasons it's  almost impossible to get a dead-on color match when you need to hit a measured color like a Pantone color. It can definitely be done, but it's outside the realm for 99.9% of t shirt printers. Thankfully there are not that many jobs that require this level of accuracy.

--- End quote ---

Huh?


:)


No, really this is awesome stuff Mark..thanks for contributing.

blue moon:

--- Quote from: Clark on May 05, 2011, 11:56:23 AM ---
--- Quote from: drdot on May 05, 2011, 11:50:14 AM ---
--- Quote from: Orion on April 20, 2011, 10:35:32 AM ---Pierre, if you printed a 50% dot that gained to a 72%, that is not a 20% gain it is 44% gain. Yep, you need to go back in the rip and adjust the curves. I don't understand the amount of dot gain that you are experiencing in the lower percentages, the numbers aren't making sense to me. How hard are you flooding your screen?

--- End quote ---

This is incorrect. There are two ways to measure dot gain and by far the most common is using the Murray-Davies calculation where dot gain is reported in absolute percentages, not relative. So in this example, the dot gain is 22%. If it were 44%, the 50% dot would measure 94% on the printed surface. I should also point out that there are two different types of dot gain, mechanical (physical) and optical. My reply here  is based on the mechanical component, although optical gain is also simultaneously occurring.

When mechanical dot gain is described, it is always measured at the 50% value using Murray-Davies equation. In order to get an accurate value you must know the starting and ending density values of your ink. This means zero to the value of the print surface and measure the value of the solid ink. You are calibrating based on what a 50% DOT AREA actually ends up. This means you need to measure with a spectro-densitometer if doing sim process or a color reflection densitometer if doing 4 CP.  There are only a few of us crazy enough to invest in these measuring devices as they are all over $1000 and some as much as $4500. 

Dot gain occurs as the perimeter (border) of the dot grows. This is based on Zone Theory (also called Border Theory) based on the mechanical growth of the dot due to ink spread. For any given linecount, the increase in the dot gain will be constant to the perimeter. So, if the gain is .001" (made up example) increase around the perimeter, it will be the same for every percentage of dot. Since the maximum perimeter occurs on the 50% dot, it will have the lowest relative gain. The very small highlight dots can easily double or even triple in value  (eg 5% dot becomes 10% -15%) As the halftone dot gets larger, the amount of gain will decrease as both an absolute and relative percentage.

We can cheat the systems somewhat by changing the dot shape and how the dots connect across the tone range. This is why we use elliptical dots to give us the most control over the tone transitions and to minimize the effects of mechanical dot gain.

Like most things in this business, the vast majority of printers "eyeball" their results by comparing their printed values to some type visual reference. This is fine as a general rule, but the problem comes when different colors (pigments) have different dot gain profiles.  This means the printer must know how each color will behave and adjust accordingly.  This is also one of the main reasons it's  almost impossible to get a dead-on color match when you need to hit a measured color like a Pantone color. It can definitely be done, but it's outside the realm for 99.9% of t shirt printers. Thankfully there are not that many jobs that require this level of accuracy.

--- End quote ---

Huh?


:)


No, really this is awesome stuff Mark..thanks for contributing.

--- End quote ---


aaaah, few more things make sense now (the high gain in lower percentages)! FWIW, I did go ahead and purchase a refl/trans densitometer and it was used to take the readings straight of the shirt. The readings were very erratic, due to the weave interference and I am certain that the measurements were off a little.

All of the RIPs and even Photoshop have the adjustment curves to compensate for the dot gain. Those are easy to use and set up if a densitometer is available.

I guess I answered my  next question in my head as I type. Should I correct and suffer low ink coverage? I am thinking, yes, and then adjust up as needed. Better to know and adjust than eyeball . . .

Thanx for sharing Mark!

pierre

Dottonedan:
@DrDot.


--- Quote ---This is also one of the main reasons it's  almost impossible to get a dead-on color match when you need to hit a measured color like a Pantone color. It can definitely be done, but it's outside the realm for 99.9% of t shirt printers. Thankfully there are not that many jobs that require this level of accuracy.
--- End quote ---

HA!  You haven't watched those Mickey Mouse artist creating that vector art file with 60pms colors and them wanting the printer to match each pms color exactly. HAHA. Just like Santa Claus said about those little M&M characters. "They do exist". ::)

Excellent food to chew on!  Good food. "Food for thought" as they say.

Thanks
Dan

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version