Halftone film output Linearization

[Dottonedan]

There is a fair amount of discussion on linearization here often, but I don't see too many people compensating for (print gain) after or on top of the linearization file[/size].To me, the Linearization process is all good once you get that far, BUT, there is still gain on the press. Sure, it's going to be in various degrees based on setup, (but you should be able to get closer) than just linearization.  I mean, the idea is to actually see some dots printing in the shadow tone.With just linearization, when when your films print out a correct 95 as 95 and 99 as 99%, you still don't get that (on the shirt). So why not bump that linearization up a hair again, ...till you actually hold a neg dot...on press...in that shadow tone. After all, it's intended to show in the print. Make it show. No, of course it won't "be accurate", but who rally cares? You're getting freak'n dots where you were not.... The 99 won't be 99...but it will show some info there, where your standard linearization won't... and that's the important part.Nobody's going to get their densitometer out at the store or when they unpack their box of shirts, but they will see info in the shadows...but won't know why they see it, nor care. All they care about is that it looks good and that's one more step towards a perfect shirt.

Now if you really want to get nuts on it, use a reflective densitometer on the printed shirt (as best you can).[/font]

That's my thoughts.[/font]

[Colin]

In all my seps, I compensate for dot gain on press. I also started out as a printer, so I do look at art differently than the majority of shop operators.

Imo:  Learn to compensate in your seps.  Make sure you are printing (and printing) clean precise dots.

Now..... For a shop that has artist turnover or whose skill level may not be up to par:  I suggest adjust some in their linearization.  But once you get an artist who can adjust their seps, or a shop who gets seps in from an artist who knows what they are doing - they will have to adjust their dot gain curve at the printer.... so now you have to change your curve depending on your artist/separator.

My 2 cents

[Sbrem]

I agree, I see the linearization as the starting point, a fixed baseline or whatever you might call it, then adjust in the seps...

Steve

[Dottonedan]

In all my seps, I compensate for dot gain on press. I also started out as a printer, so I do look at art differently than the majority of shop operators.

Imo:  Learn to compensate in your seps.  Make sure you are printing (and printing) clean precise dots.

Now..... For a shop that has artist turnover or whose skill level may not be up to par:  I suggest adjust some in their linearization.  But once you get an artist who can adjust their seps, or a shop who gets seps in from an artist who knows what they are doing - they will have to adjust their dot gain curve at the printer.... so now you have to change your curve depending on your artist/separator.

My 2 cents

Good call.  There IS a huge difference in output sometimes when (I provided art and seps) and from what they have been doing in house.  If you have all of your your gain cut back for your inhouse output, and you then use an outside source every now and then, theirs may be much lighter or much heavier. On several occasions and it's been extreme.

[Dottonedan]

I agree, I see the linearization as the starting point, a fixed baseline or whatever you might call it, then adjust in the seps...

Steve

Yep. Baseline for sure. I guess I was trying to remind people that (linearisation is not the end of it).  Adjusting manually (per job) does rely on the skills of that one individual rather than across the board for all output tho. In a perfect scenario, one should be able to just paint and draw and not be concerned about gain like we do.

[ABuffington]

Yes there are two linearizations needed, film and print.   You need a densitometer that measures transmissive light for film, and one for reflective readings for print.  The later is totally crazy from my tests.  A 4.8 oz low end shirt will have different dot gain than a nice 24-32 singles ring spun with a fine flat surface.  Mesh counts also can have unique curves due to amount of ink deposited. In my work with a few companies you have a subjective adjustment you can do.  Once the film reads correctly with a transmissive densitometer we would adjust the 'press curve' in wasatch by lowering tonal values i problem areas incrementally and make several different adjustments, or apply an S curve, or pull down the right handle to dramatically reduce 60-100 vales save them with a recognizable name. Then output film, make a screen, optimize press setup, squeegee, durometer, edge, pressure, speed, well you can see the problem here.  So many press adjustments that the second curve is ball park only, unless you calibrate every facet of your press, and with textile even this is only ball park to improve the subjective viewing of your print.  So with about 4-6 tonal ramps output with the different press curves made you print as best you can and compare it to your first linearization visually, subjectively.  The quarter tones above 75 and below 25 pretty much tell you if you are right or wrong.  For 4/color process the files may need global adjustment on the image and different curves since you are printing direct to shirt, black sim process, or black 4/C process could be different press curves.

Or as mentioned work backwards to the file and reduce tonal values there, or even curve there.  Some of our master printers have massaged the files for color, sharpness, curves etc that all lend a hand in obtaining good film, and then add curves controlling output and press dot gain to an exact science and is part of the reason their work is so crisp with beautiful tonal transitions, or sim process combinations. This is one reason I prefer imagesetters over ink jet output.  Tonals are more accurate, halftones are precise with sharp edges to shoot and produce crisper edges.  But even this needs the press curve to avoid loss of quarter tones above 75 when printed.   

[ZooCity]

I'm glad to see this being discussed.  Half the time I feel like I'm either talking to brick walls or a brick wall myself on this subject. 

We linearize to film, at the print head of whatever the output device is- epson inkjet head, wax cts -doesn't matter, it all gets truly linearized with a densitometer using the RIPs native test files.

After that it's all a game of averages.  I apply a tone curve to the true linearization to:

• Compress the tonal range.  Which technically speaking sucks but for textile printers it's analagous to insisting on certain tonal range being present in your music recordings when your speakers can't reproduce it. We don't print through 400 tpi + mesh to hold true 3% dots and, if we did, half of them would disappear into the void that is jersey weave fabric.  So we tune up the mid range and compress in the 1-10 and 90-99.  The data from the channel prints and is part of the art on the shirt.
• Account for the ink/screen/blade/press/fabric and average it all out to be "good" across the widest range of situations.

[Maff]

Yeah, and then throw manual printing into this mess and the variables get even worse... In my shop, sometimes I print and sometimes my partner prints and we both push/pull the squeegee differently 

[ScreenFoo]

I think Zoo's point 2 is a good one to look at here, at least from what I've seen 'in the trenches'. 

If you linearize your output device very well, that's an excellent first step that enables you to do accurate films.

If you take that screen to press and print it, as mentioned, you will see the gain/loss effects of exposure on the dots of the stencil, as well as those of the printing process--these can be measured and compensated for as well.

You can take that info and make a gain curve for press--but it will only work properly with those variables in place.

What do you need when you're running ring spun high fiber mass T's and the cheapest hoodies you will print on the same run?
Two sets of screens?  Press adjustments?  More rhetorical questions?

Yeah, and then throw manual printing into this mess and the variables get even worse... In my shop, sometimes I print and sometimes my partner prints and we both push/pull the squeegee differently 

That's a tough one.
I've spent years trying to linearize people, have not had any luck yet.

[Colin]

One design on good t's and crap hoods?

The trenches have voted for decades to make the seps/halftones/print parameters for the hoodies.  The t's will end up not being as amazing as they can be, but the hoodies will look really nice.  Easier to compensate for the fleece than to try and make the hoodies work with screens and halftones that are not ideal.

Otherwise - 2 set ups.

[Dottonedan]

One design on good t's and crap hoods?

The trenches have voted for decades to make the seps/halftones/print parameters for the hoodies.  The t's will end up not being as amazing as they can be, but the hoodies will look really nice.  Easier to compensate for the fleece than to try and make the hoodies work with screens and halftones that are not ideal.

Otherwise - 2 set ups.

I've done this a ton while at Disney. Almost every sim process job had to go on tees (and hoods).  We ended up shooting for the best print on hoods first, then printed that on tees and sent that off to get buy off on the prints together. This way, the expectations for the tees, were a closer match to the hoods.  All in all, still a very good looking design/print yet not a museum piece for sure.