Choosing LPI For Output

[drdot]

This is a very, very common question and there's usually some "machoism" associated with it. You know, I can print 85 lpi all day, and so on. Here are some things to consider. Just because you can print a finer dot, doesn't mean you should. This will be the first of several posts on choosing lpi based on production considerations and related factors.

The choice of LPI for your halftone is determined by several factors. Chief among these are:

Substrate Limitations.
Viewing Distance.
Screen stencil resolving capabilities.
Production variables (mainly moire and dot gain.)

The coarser the dot you choose, the easier it's going to be to make an accurate stencil, control moire, dot gain, and tonal range reproduction.

On the flip side, the finer the dot, the more detail you can hold and the less noise the image will have (surface texture that interferes with the image.)

For textiles, the most common ranges fall between 45 lpi and 85 lpi. Some very experienced printers have gone as high as 110. Here are my recommendations to help you choose the best resolution for your purpose.

First and foremost, consider the "dot density". This is how many dots are in a given area.  This is really enlightening and will have the single biggest impact on your choice.

45 lpi      2025 dots/sq inch
55 lpi      3025 dots/sq inch      49.4% increase over 45 lpi
65 lpi      4225 dots/sq inch      39.7% increase over 55 lpi
75 lpi      5625 dots/sq inch      33.1% increase over 65 lpi
85 lpi      7225 dots/sq inch      28.4% increase over 75 lpi
110 lpi    12,100 dots/sq inch   67.5%  increase over 85 lpi

The dot density is the lpi squared.  Since dot gain is uniform around the perimeter of  ANY size dot, having the dots smaller and closer together radically magnifies the gain you will experience on press. For an ideally controlled halftone, normal dot gain(defined and explained in another post) will typically be:

45 lpi       22% - 28%
55 lpi       32% - 45%
65 lpi       38% - 50%
85 lpi       50%+

High dot gain is not necessarily a problem. It only has to be managed effectively. When you consider the relationship between lpi  and dot gain, the most control occurs in the 50 lpi - 60 lpi range. This is why it is so often recommended. 

Add to this the fact that there will be increased tone clipping on both the highlight and shadow end of the tone range as the smaller highlight and shadow dots will be clipped by the thread diameter of the mesh.  Based on 34 micron thread diameter the smallest consistently (meaning all dots print with no eclipsing) printed dot will be:

45 lpi     2%
55 lpi     4%
65 lpi     6%
75 lpi     7%
85 lpi     8%
110 lpi   14%

With todays high resolution dual cure emulsions, you can resolve 1% dots for ALL of these linecounts, you just cannot print them due to the threads of the mesh being thicker than the dot itself thereby blocking the mesh opening.

I'm sure this is going to raise some questions, so let the discussion begin.

[Donnie]

Good stuff.

[Orion]

On the flip side, the finer the dot, the more detail you can hold and the less noise the image will have (surface texture that interferes with the image.)

Great post Mark. If I am reading this right higher line counts help when printing on t-shirt fabric? I was thinking that if the line count is too high that some dots hang in the screen because they don't contact the substrate during the squeegee pass which in turn may cause more dot gain.

Also thanks for providing the data on the dot gain percentages and smallest printable dot.

[Northland]

I'm begining to see the value of larger dots as part of the design element.
For example... I really like the way the dots add to the design Artelf2xs posted last week. (see link... REDLINE ROCKETS text)
http://www.norwb.com/index.php?topic=301.0

[alan802]

I think I'm going to have to read that a few more times.

[Evo]

I think I'm going to have to read that a few more times.

I know, right?


DING....school's in.

[blue moon]

I think I'm going to have to read that a few more times.

I have and am still trying to wrap my head around it  . . .

pierre

[squeegee]

Mark,

I'm wondering something about the thread diameter and if those percentage limits could be be pushed.  When we do simulated process, we work with 55 lpi output with 300/34 mesh, and I am happy to stay at 55lpi as customers are happy, so I'm happy.  Considering your statement that on a 55lpi output that a 4% dot is the smallest consistently printable size on 34 micron thread, what do you think of using a 330/30 S mesh in terms of printabilty of smaller percentages?  It appears to me from Murakami's mesh chart that a 300/34 and a 330/30 should behave similarly in terms of ink depost.

Pierre (bluemoon) mentioned this somewhere else and I've been intrigued by the idea ever since but haven't tried it yet.  If it is the case that 330/30 might allow a printable dot a couple percentage points lower (or higher) how much differnce could it make visably on a print?

I'd love to hear Pierre's take on this as well.

TIA

[Dottonedan]

I work closely with Pierre and know he's getting a better halftone because of the new mesh. His coverage is better and at the same time he's able to hold at least an additional 1% lower. We seem to get a good consistent 3% dot out of a 55 line screen.

I have what I think is another interesting topic that is somewhat related to this post. I would not want to begin to stir this one in another direction so I will create another post.

[drdot]

I'm wondering something about the thread diameter and if those percentage limits could be be pushed.  When we do simulated process, we work with 55 lpi output with 300/34 mesh, and I am happy to stay at 55lpi as customers are happy, so I'm happy.  Considering your statement that on a 55lpi output that a 4% dot is the smallest consistently printable size on 34 micron thread, what do you think of using a 330/30 S mesh in terms of printabilty of smaller percentages?  It appears to me from Murakami's mesh chart that a 300/34 and a 330/30 should behave similarly in terms of ink depost.

It is possible to pick up that extra 1% by dropping to a 30 micron thread. It gets more complicated, but it is doable. I need to put up several more posts about the relationship of actual thread count to dot area. It gets pretty deep, but it will explain a lot of the behavior that takes place in the very low percentage areas.

There are so many related behaviors, it's hard to make a general post like this and get the full picture of what's actually going on.

[drdot]

@Orion,

It's a tradeoff between detail and dot gain. I've got all sorts of graphs that show the optimium possible. It comes down to how calibrated (level and parallel in the same plane) your press is, how sharp your squeegees are and how careful you are. The single biggest player is the ink. With the right thixotropic profile, you can do amazing things. Here is a really cool trick.   Add 2% - 5% high density base to you colors and you will dramatically improve (reduce) the dot gain in the midtones and significantly open up the shadow detail.  This allows you to use a finer line count to capture more detail and still maintain full tone range.

[bimmridder]

High density base? Not high density clear? Same thing or different? 

[tonypep]

Mark...... I'm pleased you mentioned blade sharpness. Quite often one of the most overlooked interdependant variables. I know of many flatbed printers that dispose of the squeegee material when finished with the run and simply replace it just for the secure knowledge that it is a stable variable. Correct me if I'm wrong; I know of no tool to measure this.

On a quick sidebar often when printing straight process on white we often create a "wet white" (non flashed)underbase that helps to control staurated hues as well as soft pastel shades. Also improves fibrillation. Sort of goes against the rules and some may call it a bandaid but I have to say it works quite well.
Best tp

[Orion]

On the high density base, I believe Mark is talking about an additive. I have some made by Wilflex.

http://www.polyone.com/en-us/about/businessgroups/Inks/Literature/Wilflex%20High%20Density%20Additive.pdf

[Colin]

Mark is making an excellent point about how the "body" (the rheological properties, how much it flows/pours vs. keeping it's physical shape) of your ink has a large impact on how the dots hold their shape when deposited on the substrate.  Be it the garment itself or a white base plate.

It is VERY important that the ink be just ridgid enough to maintain it's dot shape both during the initial ink deposit (will not flow out under the edge of the stencil) and when it is "steped on" during the wet-on-wet process.

Some ink lines will need an additive to acheive this "controlled ridgity" while others already exhibit these properties.

Loving the discussions here by the way!  I've really missed this level of interaction...