Self made photoshop plug in halftones. Not what you think.

[Dottonedan]

You know those manual bitmap halftone methods some people use to print halftones to avoid using or buying a RIP for?  Well, don't trust them. LOL.  They are not always accurate. Depending on how high/low your rez is, can affect the outcome.  Check out the differences in these images.

Each square is 1". Each have a 15% fill in it.  When bitmaped into halftone at the same LPI, but at different resolutions, you get different results. Here, I'm using just 5 lpi. We get unexpected results when converting to higher LPI such as 55lpi..but using lower rez files also. The actual number of lines in the result are all different at each resolution.

Just something interesting to know.
Dot-Tone-Dan

[Full-SpectrumSeparator]

It is not a matter of trust or accuracy.   

You make it seem as if this method is always going to produce unpredictable results.

From my tests it appears that you have simply shown the lower-limit when going at 1200dpi it is about 19 LPI,  600 about 10, and 300 DPI can produce a 5 LPI pattern, but I've never gone that low for printing or RIP purposes either with resolution lower than 1200 or with LPI's below about 35 etc, unless for artistic effect.

The 300 dpi result is off as you've shown, for 55 LPI it is 60,  but why would you use 300 dpi halftones for printing?   I have found perfect accuracy in the range that I use at 1200 dpi and the LPIs 20 and up.

It is sort of misleading you're showing a few examples and that there is a lower-limit, and 300 dpi is off, but the rest is fine so have you done any measurements of all the others to know exactly what deviation percentage there is and what resolutions?   

1200 dpi and LPI between 20 and up seem fine.   So what's the problem?   

There are a lot of resolution and LPI errors in other RIPs as well, but there are still similar vector to raster processes at work and the results are almost impossible to tell the difference or are in fact sometimes higher quality when you control it through programs and know their accuracy points.   

The thing with most RIP software is it does not allow viewing and working with the halftone separations in realtime or making changes and edits to the halftones themselves, seeing composites in color, etc... its just a matter of preference to working style, not quality in terms of RIP capability.    There are some times where output settings can be controlled by a RIP software that are harder to access manually but that has to do with film dot density and other control in printing, not the halftone screening techniques.    It's a deep and complex science, but yes in photoshop there is a lower limit to getting accuracy from resolutions below 1200 and LPI below 20 etc.    Good find and nice to point out, but keep in mind that true science demands as much data as possible before making sweeping generalizations about the overall conclusion or results...  you've only plotted a few points, maybe we should find out the rest of the percentages and DPI/LPI deviations in order to determine the accuracy or error % range, or at least a few more points at 1200 dpi and above 20 LPI to determine if its off even at all?   

While we're at it, we should do the same for the channels dot-gain % previews and blending math, to see if perhaps there are inaccuracies in how inks are shown in compared to what you'd really get, as I've seen in various tests having to do with light and dark inks, various backgrounds, various dot gain preview % curves, and the solidity, also testing any mode combination of gamma and color profiling, there are just certain inaccuracies to be aware of when it comes to any program we use.   This is always what i've advocated is to know your ruler is straight before you build with it.  So having accurate working methods in any programs is key to getting a reliable and consistent, repeatable result or having a WYSIWYG - What you see is what you get - model, and especially when giving specific instructions to a program like halftone conversion you want to know its doing what you asked.   This is a good example that you don't throw the baby out with the bath water,  just because there are some errors in the bitmap halftone conversion at low LPI and resolution and vastly inaccurate ink preview in multi-channel doesn't make it totally useless, many still get amazing results from the combination of both, and as long as you know what you're doing and the way your tools and methods work then it is all the same.

Apart from printer-hardware control and certain proprietary halftone or other algorithms, what are you saying is the difference between halftone screening manually through various algorithms?
 Either in photoshop or corel or independent and open source simple math algorithms, threshold arrays work fine in photoshop or other techniques etc,  screening in FM/AM hybrid etc is screening whether done digitally in the computer before sending to a print device or done in the RIP of a print driver at some point, only differences I can see are potential proprietary algo distinctions or the hardware printer controls like droplet sizes etc if not accessible by users in default.   But the screening is basically all the same, conversion of contones to halftones in different shapes, angles, LPI, resolutions, etc...  certain techinques work with halftones together to make things that post-processed RIP of different separations don't allow, however the printing of the screens is different than just the RIP generation of them.   So if someone can attain the same density control and dots printed from a software-generated screen as opposed to a software generated in-print-device or other 3rd party RIP devices screen then there is no difference except when and where you made the halftone dots.   The only thing I see you've pointed out is photoshop has a bitmap conversion lower-LPI limit of 5 for 300dpi, 10 for 600, 20 for 1200,... and that 300 dpi conversion to halftone LPI is unreliable so don't use those low resolutions.... but i've only looked at the 55lpi becoming 60 lpi, what do the others do for 300dpi?  is 600 dpi always fine? 1200?
Let's make accurate judgments before simply saying all bitmap conversion halftone methods are unreliable.    Did you even mention what program you did this in?   It appears to me its the photoshop issue.

Any others have this issue, what about halftone threshold array patterns in photoshop do this also have LPI issues below 5LPI and for 300dpi etc?    This thread brings up great questions, but it should not be looked at as starting with a conclusion that is accurate or complete.

"You know those manual bitmap halftone methods some people use to print halftones to avoid using or buying a RIP for?  Well, don't trust them. LOL.  They are not always accurate. Depending on how high/low your rez is, can affect the outcome.  Check out the differences in these images."

-You can trust them if you know what the input/output variations are and what to avoid... maybe its only a small set?
-They are mostly always accurate in my experience... so how "they are not always accurate" a good generalization?
-Depending on how high/low your rez --  right... sooo...
-Check out the differences in the images -- and its a very small set, what does it tell us, maybe this lower-limit of 5 lpi for 300, 10 for 600, 20 lpi for 1200, and 1200 lpi is accurate, but 300 lpi is not?

We need to do more measurements and comparisons and for each program to judge the accuracy and reliability.

[Dottonedan]

Let me guess.  You sell or give out a halftone plug in generator of some type using this method? Then I can see why you might find this post worth your valuable time to pop in and contribute.

Simply put, It's just some information that many people don't look at, know anything about and most of these same people even with this info, will not change anything they do.  I could quote and highlight and make a long list of contradictions countering your accusations but for who's benefit? People see, People comprehend on their own.

I found that resolution impacts the line screen quantity results when using the manual bitmap/halftone method.
So sue me.

Of course 1200 is better. Why then, would I need to mention 300 and 600?  Because there are PLENTY of people out there using that rez right now. Showing this, is another reason not to and to look at what they are doing. With this, it sparks them to dig a little themselves and they learn more as they go. Just because it's something YOU have not ever done, doesn't make the information less valuable. Keep in mind, it's not just YOU we are putting this information up here for. Many who will eventually read this do use all sorts of resolutions. To hard to believe? Well, I would imagine it's also hard to believe that people use 60lpi on a 110 mesh for an underbase and wonder why they lose the small dots. Worse yet, some never give it another thought.

I didn't say the Bitmap halftone method is not good to use at all. Many shops just starting out want to look at how to get halftones and most want it free or very cheap. It can be a viable method used by even the most seasoned artist and separators. It's always better to use the high rez of 600 or 1200 even if the lpi is not 100% accurate. Close is good enough.

People don't use 5, 10 20 lpi?  Sure they do.  Not often, but there are times. For example you may have art that is of a graphic nature and requires some form of a gradation or burst out from the center. It's using large glitter flakes in a clear base. The mesh required might be a 60 mesh so to do a halftone fade (that hold the small dots) in that mesh, you need a 15lpi gradation. Plenty of other examples. HD etc.
For my test. The only reason of using 5lpi was for easy dot visibility and also to test the extremes. 5lpi and a more common 55lpi.

I'm not stating profound scientific ground breaking screen print industry findings. It just some additional information to help some people make good choices. Like everything else on the internet, take it with a grain of salt, use it or don't.

It's not a scientific study. I'm not a scientist, nor do I claim to be. To do so, would be ridiculous, self serving, self proclaimed and narcissistic. If you want to test all of the different variables of 1200lpi @ 1 lpi and again, at 2 lpi, on up to 85lpi. Be my gust to take over. Just know that you need to post that on your own website and link it here as each pic example in a single post would be looooong and lack interest.

There ARE different results (as shown) and that's enough for them to know they may want to check what they are doing. That's all it was intended for. Made you do a few of your own test though, didn't it. Now you know.


This thread will be locked to maintain it's original integrity.