T-Shirt Forums banner
1 - 9 of 9 Posts

· Registered
Joined
·
97 Posts
Discussion Starter · #1 · (Edited)
Here is a color separation for black shirts. Client has 5 available heads. The first link is the original image. The other link is the separations. White base, cyan/light blue, green, orange and highlight white. Adding a black would have been better but not enough heads. I opted to just do everything in photoshop at high resolution with anti aliasing off. So the file may be a little heavy. What do you guys think? Any critique will help. There is no t-shirt printed yet but hopefully the client can post a picture if it does get printed.


https://drive.google.com/file/d/0B-MT6xPkO5r2SThWMnVPbE1jemM/edit?usp=sharing
https://drive.google.com/file/d/0B-MT6xPkO5r2eU9uNEJuN0FOMVE/edit?usp=sharing
 

· Registered
Joined
·
342 Posts
Hi Denny,

Looking great! One thing to watch out for, it looks like you're enhancing the hues a bit to help the saturation more, the original is very faded tints of the hue.

One thing that can be done in printing is to actually use lighter tints of the hues mixed with white if they are coming out too strong or saturated in the print... but I think you have enhanced them a good amount.

However... it looks like there might be too much enhancement where some of the green was mixing with the yellow/orange of the shades in the people, especially the hair. So I made a quick setup using your file simulating the color seps as inks in photoshop layers, and just bumping up the base a bit which should happen on press, so that without the black you will still be ok... but the green might gain too much and you dont really want or need any green in the people's skin or hair etc, I think it was just part of the blend in the original image very light.

I've shown how I might go in on the seps if I were you and just mask out the people's skin-tones and hair etc from the green separation so that it just won't even be there on press do you see what I mean?


Otherwise it all looks awesome, great job on picking the hues that were part of the original instead of trying to blend with too many colors you don't need. The orange/white/black will blend and capture the flesh-tones and just the light amounts of it in the hair simulates all the blonde and other highlights... I just think the green is showing up from light shades of the yellow/black and you don't want to enhance that, so if anything just cutting it out looks better more chance to capture the orange-tint tones instead of getting them contaminated with the green.

Thanks for sharing, I think it will turn out great, you seem to have already compensated the underbase and highlight white enough so they don't gain too much, although if you were working in photoshop Channels be careful you might have pulled back the highlight white too much (it appears too bright than it really is most times because of certain problems with the channels ink simulation in photoshop).

It's good in just the highlight bright white areas.. but if there is too much gain in the color hues over the white base and not enough white highlight dots to help soften them back... you'll get a more saturated print -- but compared to the original art it may be too much color... and then without enough highlight white dots there to simulate the tint, only way to adjust on-press is to lighten the inks themselves a bit like I mentioned, but this is also a way to print those tones as long as you dont need the really pure bright areas of the color present.

The separation to greyscale gradients here is part 1 of the process... but the conversion to halftone rip is the next step and should usually be done by the separator, although my contention is that the printers should perform the separations however we are not all at that point yet. In doing separations for a client's art and giving to a printer, you'll want to communicate effectively and work closely with the printer in going from the separation to the rip on films, hopefully giving them halftone files yourself so that they do not mess up the separations with bad halftone rip techniques and settings.

However you also want to understand how much control they have over the print, depending on the inks and how they will capture the halftone gradients on screens... the "quality-control" must be carried all the way through from the art input, to separation choices of inks blending, the way you convert and mix the actual halftones in RIP, and then how that translates at the printer to films, screens, and press... the sequence of the inks on press and the ink characteristics, etc.

You are already on the right track, but try to imagine it like a puzzle, and you want the pieces to fit together nice and snug all working to make the image when its done as the inks on the shirt.... but when it is still greyscale instead of halftones - this is the picture on the front of the box and not the puzzle-pieces. You want to control the puzzle-pieces more as the separator yourself -- or as a printer doing the separations, than just giving the pictures to the printer -- they may make the wrong puzzle pieces so to speak.

Great work and like I said, the only thing really is just don't want any of that green enhanced in the hair/flesh-tones.

Take care!

:D
 

Attachments

· Registered
Joined
·
97 Posts
Discussion Starter · #4 ·
Yeah I just noticed some of the colors shouldn't be there when I took a closer look at it. Just when I thought I fine tuned it you find something else. I also just noticed the orange in the other parts of the image instead of the flesh areas. Not sure if that will affect the design too much.

I actually did this separation manually using the color range tool. Maybe thats why Its pulling more colors than I need. Then I tried to fine tune it in channels. Then I gave it to the client ripped already. I used the bitmap conversion and made all channels 22.5 degrees with eliptical dot at 45lpi. I'm not familiar with the clients press so I think this might be ok settings for them. By the way I did it for free.

Thanks for the critique and by the way I'm still testing out the Simple Seps Raster for Photoshop. I need more time to play around with it. It is something I could use for sure.
 

· Registered
Joined
·
342 Posts
"All channels 22.5 degrees with eliptical dot" = Color/Ink Blending Death-Sentence

You are directly overprinting your inks which depending on the opacity and response curve and especially the sequence of the inks will completely change whatever you thought it looked like in channels preview (which was incorrect in preview to begin with as it can never be shown correctly with both light and dark inks in the mix)... and in halftones preview this can really be seen..

Attached is an image showing a flesh-tone color blend of orange, white, and black, in halftones, over a blue shirt color background.... at same-angle, rosettes, and interlocked halftones.

All of the color curves represented by the gradients in greyscale completely change in the "puzzle" of the halftones... and the same-angle dots ruin your color ink blends and the attempted mix you're going for in the separation... especially this is seen in the black and base mixture... which if it is not inverted then shirt-color comes through all over the print -- changing the entire shift of colors towards the shirt.

Much more to get into with these things in the future... but please be aware that color range tool, multichannel mode, and same-angle dots are like the 3 worst things you can do in color separation for screenprint. I wish someone had told me these things when I started in this industry 10 years ago, it would have saved me a priceless amount of time, energy and frustration over trying to make an incorrect workflow produce correct results.

The time has come to be brutally honest about the lies and false teachings in this industry that are spread around as if they are some sort of holy scripture or almighty standard... when the entire time it has actually been the last approach one would want to use in any sort of attempt to do professional color separation work for screenprint.


Let's say you have a 2-color gradient...

Color range tool will not pull enough of the blend to have each color reach the correct curve in the gradient... you cannot adjust to push levels up to grab more color where zero color was extracted from the original. This can be shown and proven compared to other techniques which are mathematically logical and accurate to how they extract color gradients from any image, and the color range tool is definitely not the way to do it.... you'll end up with an arbitrary banded and more contrasted blend where it drops off too soon and the darker regions are too dark... not enough smooth transitions in the blend.

Multichannel mode will especially never be capable of showing correct previews of the gradients of the two colors pulls - even if you pull them perfectly, if one is a lighter ink and the other a darker ink over certain color backgrounds it will always throw opposition-errors.. it is most extreme with black and white - which we use all the time as screenprinters, so you can never preview a base and hl white and black with your other colors and have it look correct... this leads to making incorrect decisions to adjust curves/levels etc while viewing in this mode and what one does the art no longer has any real-world bearing and could be potentially making the opposite decisions of what you intend.

Finally, even if you pull the perfect gradient blend, then ignore any preview issues and convert to halftones -- if you convert to the same-angle in the halftones then whatever gradient of color you were attempting to make is now a completely different color-mix -- including space between all the dots where shirt-color or underbase come through, or other colors, and the inks now sitting directly on top of each other in the blend are fighting each other for positions in the "puzzle" of the "average" result you are trying to get from the blend of the inks.

As mentioned in the posts above.. the separation to gradients is only the first step, and making your intended color-reproduction turn out the same requires unique halftone methods of placing the inks in as many places as they can be to maximize coverage without overprinting , consider it like a halftone-index is the best approach, rosette halftones are the next best, and flamenco or same-angle is the worst approach to blending screenprinting inks. The final "puzzle" mixture of the halftones themselves can make or break a separation just like the screenmaking which needs to capture those halftones etc etc.

I will keep you posted on future developments, going along with all I've mentioned is why I have the renovation of the SSR photoshop beta going on to put in the rip capability in layers mode so I can actually show exactly the issues faced in the workflow whether one uses channels or layers, and how to actually work with the right goggles on as professional color separators instead of in the dark with blinders on. Thanks for sharing your work, good luck and hopefully I can help you to learn the right way of approaching this kind of work if you want it to make sense and have accurate, predictable results.
 

Attachments

· Registered
Joined
·
1,558 Posts
I looked at your file,

I put back all file into a multichannel in photoshop. all color set to 20% opacity expect underbase.

First impression is " I have a good output file" , dot gain is good...
It will need minor adjustment but not big deal.

Your orange is fine, it will not affect the other part of the design, as it is only 2% - 4%, not easy to hold this on the screen.
 

· Registered
Joined
·
1,558 Posts
@ FullSpectrumSeps

Could you please share the same file output with Simple Seps Raster for Photoshop ?

The file as it is with no tweak, because that would be a great opportunity for me to compare both techniques.

I am not use to work with layers, because 10 years with channels, I have my own routines and habits.
And changing habits is not easy :)

Cheers !
 

· Registered
Joined
·
97 Posts
Discussion Starter · #9 ·
Hope you have time to take a look at this separation? This was done with SSR in layers. After making adjustments in layers visually and was satisfied I converted the layers to channels. This was the result.
Underbase White - cyan/light blue - green - redish orange - highlight white.https://drive.google.com/file/d/0B-MT6xPkO5r2dm9EeTZRRGE0RDA/edit?usp=sharing
https://drive.google.com/file/d/0B-MT6xPkO5r2SThWMnVPbE1jemM/edit?usp=sharing

As far as the halftones should I make the underbase and highlight white 67.5 and all the colors 22.5? I previewed it in bitmap halftones and some of the colored halftones fall into the black of the shirt. Is that ok?
 
1 - 9 of 9 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top