TSB
screen printing => Separations => Topic started by: DDSol on November 25, 2013, 12:37:09 AM
-
Hello everyone. I've been so busy I don't even know what people are talking about these days.
In any case, all this business has led to me finally producing Seps! Okay, the package you get is a little meager. Also, not having an account is a little annoying as well. You can never reuse settings you used yesterday (unless you left your browser open).
Browser Color Seps here (http://68.80.79.166/seps.html)
Okay. Just to make sure we all understand that this is an Alpha version. Still. Yes, it's true, it would be nice if you yourself could decide on what the ink opacity is and/or how much dot gain you're expecting. And you can't. Also, these values are a bit fidgeted. You can't adjust them and I didn't spend much time on them, so this means they're bound to not be so great. Now, you could mess with curves yourself if you like, or you could see what these default settings look like on a shirt.
There's a lot more to be said, but I added a nifty feedback thingy so you can spill the beans on what you love and what you hate. This in turn gives me the opportunity to fix what is hated first and to enhance what is loved afterwards. That in turn gives you the tools you want, the way you want them.
Please do not get too attached to things being in certain locations. Everything is still subject to change! That may not seem so bad now, with the UI being very ... uh... lame. But when I go to move things to a new home, you may miss it being where I dumped it for lack of a better temporary place.
That said, in recap, art->seps->shirts! Now in a snap.
Awesome.
-
Just tried one, check my feedback...
Steve
-
Yes, I saw you left me feedback as well. That was bad, indeed. Here's the thing, though: It can use any inks you toss in there. And to make that work better, I added an "ink optimizer", which pre-selects inks for you. However, this process sometimes chokes, giving you a green when what it really needs is a black. Maybe it combined a black and a dark green and decided it was "close enough" (some error is allowed to save a screen when it's a minor color change).
However this may mess up your seps, you can easily correct them by tossing a black ink in the mix. As a matter of fact, you can use any number of inks of any color you like. The system does not try to decide on how many inks you want or what those inks should be. It only tries to help by pre-selecting some inks.
So this means that that ugly picture you saw was just one way to do it. If you think your black is too green you can always change it.
But there's no manual yet, so you can't be expected to know how to work this system :).
-
I just want to point out that:
The pre-selected colors can be way off. The ink optimizer doesn't work very well yet. This does not mean the seps are bad. It just means the colors it chooses automatically are sub-optimal, especially for images with only a few hues.
Luckily, you are free to change, add and delete any colors you want. You can also reorder them to any order you like to print in.
Note that the button to download your seps is located in the top-left corner.
-
I just want to point out that:
The pre-selected colors can be way off. The ink optimizer doesn't work very well yet. This does not mean the seps are bad. It just means the colors it chooses automatically are sub-optimal, especially for images with only a few hues.
Luckily, you are free to change, add and delete any colors you want. You can also reorder them to any order you like to print in.
Note that the button to download your seps is located in the top-left corner.
I opened my original in PS, sampled the black, and found the RGB values at 15,15,15, instead of 0,0,0, so maybe that had a little something to do with it... I'll download the seps and check them when I get the chance. Pretty buried going into the holiday break...
Steve
-
Han I hate to tell you this but your math is way off, I mean HUGE off and you have strange lines floating around in even simple separations. Why did you release this like this?
-
It says Alpha quite clearly. It's not 'release'. It's just that anyone who feels like giving it a whirl is allowed to. It's free. It's quick. If it works for you then you just saved yourself hours of work or many dollars to outsource it.
But my math isn't off, AFAIK.
I don't know if the lines you speak of are in the final seps you download, or on screen. (I figured it out, see 2 posts ahead)
The lines on screen are actually... well, it's details. It has to do with angles! Not colors. Also you might be using the wrong ink colors (I dunno). (see 2 posts ahead)
The lines on the downloaded seps... I haven't seen them (I have now), but I would not be surprised if there's issues. It's hot off the press and I'm sure there's a bug or 2 to be ironed out.
So, it's not a release. It doesn't run on a webserver. It runs on my home PC. The final version will run in the cloud.
-
I just realized that you may have JUST ran stuff... I'm right now working on adding slots for the white base and the shirt (so you can change the shirt color). This messes with the "production" version in that some things are shared between my dev and production system (notably, the engine backend is shared).
So, it's in flux right now. I should have it working correctly (and better) shortly.
I'm sorry for any inconvenience. The final production version will run completely separately, so I can make any modifications on my end without messing with people's seps.
-
Hey Tom, I think the strange lines you're referring to is from the 4 color bar image you uploaded.
Unfortunately, the GPU will only accept images with a width and height that are a power of 2, such as 256 or 2048. If your image is 1700px wide, I have to scale it to something else, say 2048. This introduces a dilemma: I can scale with nearest neighbor, which would prevent any blurring but would introduce aliasing, or I can scale with interpolation/anti aliasing.
For screen printing, it's really best to not do any anti-aliasing (a little known fact). Interpolation introduces intermediate colors that aren't found in the original art. In your bar image, the yellow and magenta border gets interpolated to pale red ((rgb(255,255,0)+rgb(255,0,255))/2=rgb(255,127,127), which is pale red). As such, it uses red ink to represent it.
I decided on the smoother version because the browser does all the seps in real-time. This means that the image is scaled to the final size and then separated. This in turn means that if interpolation were to be totally off, downscaling would mean dropping pixels. You can't draw 500 original pixels into a 10px space, so something needs to go. You can interpolate, by mixing the colors of every 50 pixels together and put them in 1 (x10), OR you can literally throw out pixels that don't fit. Their color is lost forever. If you have a thing horizontal line, for instance, it may completely disappear, depending on the zoom and pan position of the view. This flickering and disappearing of art is not acceptable. Therefore interpolation is used to generate the on-screen image, which is then separated. The original color is therefore lost.
In most cases this is not an issue. Sharp lines are less common in the type of art we separate with gradient/halftone seps. And where they still exist, they will indeed be blurred and may be represented with a different ink (gray on a black/white sharp border or red on a yellow/magenta sharp border, for instance).
This is unfortunate indeed. I also cannot solve it because the original image may not fit in the browser, considering that in screen printing we often work with huge images indeed. Therefore it's impossible to have the browser separate First and Then scale (although I may do this on the down-scaled/browserified image in the future to at least lessen the effect and make zooming more consistent).
HOWEVER: These issues are only cosmetic issues! They only affect the view in the browser, which is limited due to the browser's weaknesses. I have no such weaknesses on the server and the downloaded seps will not include these anomalies.
So, in retrospect, the math isn't off. It's only perfectly separating a slightly blurred image. Note that the smoothing factor is much higher in the small previews than the large center image due it it being down-scaled vastly, which means more pixels get crammed into 1 and therefore there's more mixing.
I have not found any example of in my test art that was as ugly as the image you uploaded. But I have mainly been testing with gradient-type images. Thanks for the pointer. If people have these issues often, I'll have to devise some sort of solution to this issue. I can't have them think their seps will turn out like crap when they in fact totally won't.
-
NEW: Change shirt color and a separate slot for the white base (which you can now turn on and off just like the colors).
ALSO: Improved dragging, and use right-click to select a single sep instantly.
Try it here (http://my.screenseps.com/seps.html).
-
I see people uploading art that has "flat" spot colors a lot. That's fine but there are a few things to consider:
The "ink optimizer", the thing that pre-picks colors, chokes on these. It can't handle the low dimensionality (not enough colors). It can also not handle the high concentration of color (lots of red and white and nothing else, for instance).
Sorry about that. It needs a lot of work to be able to deal with all these issues.
It seemed to me to be a better idea to first make sure you could download some seps.
However, even though it may be confusing to know that the colors it selects may at times be downright redicilous, do not let that fool you into thinking the separation system is malfunctioning. It isn't. It's the ink optimizer only that chokes.
The system is meant for art that would be very hard to separate by hand. Simple spot colors were therefore not tested well as they are not a priority right now. However, in the future, there will be full support for them.
Nonetheless, if your art is sharp it will do an awesome job even now. It does require you to change the colors around yourself manually. Do not expect the colors it selects to be optimal.
There are certainly a number of issues that require solving:
- If the system accepts png files, I should accept them all. Currently the system is limited to non-interlaced pngs only
- If it rejects png files because they are the wrong type, it should tell you and instead it silently stops working while the "liading..." screen is up. Not very nice.
- It should accept many graphics formats
- It should allow you to choose between flat spot seps, sim process with Rosettes or with Flamanco
- It should allow you to log in and out and store your favorite settings
- It should provide warnings on such things as low art resolution, moire, areas of color shifts, etc
- It should provide a file with the seps that at minimum contains a list of color swatches so you can see what those weird color names really mean
- It should of course suggest the right colors in every situation and even allow you to automatically improve your color selection to get more accurate seps
- Sliders and knobs and buttons for tweaking whatever you want to tweak
- Image adjustment such as making black really black or enhancing contrast, changing gamma, etc
- Many, many more issues
I am aware of these things. Please know that this is an Alpha Test. Some of the items in this list are crucial and others may be optional. They are not available because you are using an unfinished product.
However, this unfinished product can produce some awesome seps. It requires you to:
- Supply high resolution art that is not anti aliased or blurry in any way. Do not start with a jpg.
- Not take the first color suggestion to mean the limit of possibilities. You are free to add, remove or change colors.
- Click that download button
- Optionally tweak the seps a little for dot gain or some such. Can't guarantee the quality of the curves yet. Some testing should be done before I can make guarantees.
Really, that's all there's to it. You can have your seps in 2 minutes every time. They will be of colors you pick.
So, rest assured that I'm working on adding all the relevant features, but also be aware that 95% of the real power is already at your fingertips. It requires a tutorial maybe, sure, but it's a surprisingly short tutorial. Just arrange the colors you want to print with in the order you want to print them. That's it. That's a fairly short tutorial, I think.
Note that the system should never be allowed to have empty seps. If you see them (and you may very well), delete them. It's nonsense. I apologise for the confusing nonsense colors.
A quick note on color editing:
Any sep can be moved to another location or the trash bin, by dragging it. There's no trash bin unless you're dragging. So that may be confusing. You can change a color by double-clicking. You'll get a color selection dialog and in this dialog you can also select from a list of inks. The inks are ordered by how close they match the color selected to the left. This allows for quick searching of the color list by color, but it hinders if you want to pick a specific PMS, for instance. I'll fix that as well. You can also change the shirt color the same way. Lastly, to add a color, click the big + in the ink splash. That's all there's to it. I think that'll do for a quick how-to.
I hope this helps.
-
Han look at your posts... :o Color is not that complicated man. I promise you its just not as complicated as you have made it out to be. I know of open source free access web based apps that can run circles around you right now. Things that have evolved rapidly in only the last several months that can walk all over anything we have worked with relating to color before. My best advise.. do not put all your eggs in on basket. 4 years and 60K lines of code.. I know of open source projects with over a dozen color science PHDs working on them that have compressed the most difficult aspects of color science into simple scripting languages that are processed thru shaders. The color game is over, actually it was a long time ago.
-
Hey Tom, I honestly don't know what to do with your PHD statements. You refer to projects and such with shaders and "simple scripts". These things are very much meaningless to me. You point at vagueness. I've heard you say things like "the math is perfect". Just because 2+2 is 4, that doesn't mean anything, but the math is perfect. The main point therefore is not that the math is perfect, but that it's meaningless. You speak of apps that run in circles, which also is highly meaningless to me. Shaders are meaningless. They are not a secret, they are not special, or complicated, or magic. Simple scripting languages don't get processed through shaders and shaders are not written in scripting languages. They are compiled. Shaders do not process scripts, they process pixels. It is very clear to me that you do not know what you're talking about and that you use lots of big words you don't understand in order to come of as knowledgeable.
Therefore, I suggest you stick to your pack of PHDs and haggle with them over how long it would take for you to copy my system. It is certainly telling that you have never uttered a word about WebGL until you found it used in my system. I want to cautiously point out that it is not WebGL that provides the magic for my system, it is instead the math itself, the pure math, the kind that means the same thing on paper as it does written in a script or a shader or on subway walls. And you couldn't write a batch file if your life depended on it. I'd be surprised if you even know what a batch file is.
I'm saddened and disappointed in myself that I allow myself to be dragged into this tit-for-tat conversation. I think it would be best to leave your claim with you, so that you may make the whole world believe that you are the God of the Color Game. Just leave me out of it. You should really stop with the flame wars and instead focus on trying to reverse engineer what I did, because I'm moving fairly quick with it. That 'color game' that you say is over may very well be, and this means you'll have some catching up to do. I saw you had some "color rendering issues", but the math is perfect. It's the browser that "can't render it". Well, it does a bang-up job for me, so I doubt it's the browser.
Lastly, may I warn you that my 4 years of development weren't me sitting around flailing in ignorance, they were me building a sophisticated piece of software. I am very capable in the software world. I am more capable than the people you hire even if they have PHDs, I'm quite sure of that. Not that I'm so great, but I have a passion for this stuff. Nobody is paying me anything. I do it because I really, really want to. And there's a difference between doing what you want and doing what you're paid to do. This difference is passion. I never set out to copy anyone. I set out to fix Wilflex EasyArt, because it hurt when I saw it. It made my eyes bleed when I saw the results. And I had seps to produce. I could scarcely believe that such a lousy job was done on these sep systems, so I built my own. I am hereby sharing. It's not second rate. It's not built to make a quick buck. I built it for me.
May I note that you have my Skype? Instead of telling me this stuff in a public forum and putting your dirty laundry out there, you can simply talk to me directly. This public smear campaign, frankly, is beneath us. I also believe it to be tiresome to others to have to read these mud slinging contests. Then again, I believe it is your intention to make me look bad, which means you're not communicating with me at all, but instead with all the others on this forum. So it would do you little good to say these things to me directly. You have nothing to say to me to begin with. You just want to tell everyone that you know someone that has a PHD and that they therefore will produce better seps and that I am 'old hat'.
I'm certainly not the first you've dragged into one of your little wars. Still, I am ashamed that I am not mature enough to let it be. Strong feelings of being wronged well up in me when you use your slick (but meaningless) name-dropping and underhanded stabs. These feelings are powerful, but I should know better anyway. I wish to apologize to everyone whose stomach turns having to read through this crap.
Tom, please, I wish you all the luck with your endeavors, honest or otherwise. Go and build your empire, but leave me out of it. This is such a waste of time and energy.
-
Wow you two seem rather passionate about this. Granted it is pretty sweet, I thought I would maybe break the tension a bit...
While it is clear you both know quite a bit about art and colors, I hate to break it to you but a customer of mine turned me on to a great graphics program you both may like-
https://www.youtube.com/watch?v=SPP6lFHg_1E&feature=youtube_gdata_player (https://www.youtube.com/watch?v=SPP6lFHg_1E&feature=youtube_gdata_player)
(Someone please insert the video for me, and then tell me how to do it!)
-
Granted, that is a totally awesome video of a totally awesome program. But the MS Machine is a little big and you can't expect me to compete with them... Word is probably the most awesome software ever written. I don't think it's fair to throw that in my face like that.
I hear Word also does cool things like dynamic arrows and even thought bubbles! Really, trying to compete with such a powerhouse is a David and Goliath type deal.
-
You know I always get blamed for fighting when I am telling the truth. When we first rolled out Corel Seps.. Oh my God we were blasphemers.
Disagreeing with you based facts and truth is not fighting with you Han. And of course now I am the evil Darth Vader building an EMPIRE! ROFLOL! to conquer HAN SOLO!
I am amazed at how you and this industry have turned this color stuff into some mystery of the AGES with the Wizard Programmers, the Mighty Pro Separators and Lords of High End Printing!
Have you ever just looked at simple rain bow.. Oh that is so deep. Ever applied white or black to a color oh oh oh that just so magical! Digital color 15, 15, 15, or 220, 45, 60 Oh my that is so difficult.
Fact is we not to far away from Gimp eating PhotoShop for breakfast as far as screen printing separations are concerned. Of course you would need to spend allot of time on research to understand that statement but do not worry the days of PS are coming to an end. Sort of like Kodak, who would have imagined. Wanna see the future just test drive Canva.
Oppps did I blaspheme again :o
-
Tom, I am okay with you telling the truth. I just want you to not tell it to me. Your 'truth' is irrelevant. You love open source, which is super. Enjoy. I have no problem with open source. Why would I? I just don't see how it has anything to do with "Drag and Drop browser seps", which is the topic of this thread.
As for "me and this industry"... I am now in 'cahoots' with an industry? A 'conspiracy', maybe? Thanks for letting me know, because it means have some secret meetings to arrange. This changes everything! Anyway, it's hard for us 'cahooters' to turn color into a mystery. Yeah, turning anything into a mystery is hard. It wasn't a mystery yesterday, so we have to mystify it somehow. You understand how this task is hard, since everyone already accepts it as demystified and all... It's like trying to make everyone believe the earth is flat again!
Gimp may or may not eat PS, for breakfast, lunch, dinner or whatever. How on earth is that relevant? Open source isn't special. The software is pretty well the same. Using Gimp for color separations is something I would personally discourage, but not because Gimp is crappy or not as good as PS or whatever, but because Gimp was not designed to do seps. Neither was PS. They are both the wrong tool. It's like using a screwdriver to drill a hole. It works, somewhat, but it takes forever and the result is not quite optimal. That said, whoever feels strongly about it should not take offense by my saying this. Instead they should do whatever they want and use the Gimp or their beloved PS. To each their own.
I myself only use OpenOffice instead of Microsoft's suites. It's a matter of principal, although I personally don't like the ribbon interface.
As far as "seeing the future"... that turns out to be completely impossible. It violates causality. Maybe you've heard of time paradoxes? Well, maybe that's a little over your head... but whatever this "Canva" is, I am quite certain I cannot use it to see any 'future'. It may have impressive graphics abilities but seeing the future isn't one of them.
What exactly, btw, did you disagree with me on? I can't for the life of me figure it out.
I have an idea: Could you please vehemently disagree with me elsewhere? Maybe you can paint me black in a different thread at least? I think it's fine you blaspheme or whatever you do. Just don't force it on me please.
-
Your right Han, I am a delusion conspiracy believer building my empire. But OS is the future Han...
How are you going to complete with this Han?
https://gmicol.greyc.fr/
I suppose, in this industry of screen printing there just might be 2 to 5 people that just might be able to comprehend that page perhaps you are one of them.
I am a wee bit occupied at the moment but the fact is with that OS script your 4 years of labor and 60K lines of code could be dominated and disposed of in less than a month or 2 or 3 months max.
It is 1000 times more powerful than anything in PhotoShop if you have the color understanding!
It is a free open source solution.. The color freak dream team cranking out open source.
-
Okay, that link you posted points to a silly thing that can run filters on images.
On the server.
Why is that special? I really don't see the benefit of this system. Plus it's really ugly as well. And I have to scroll by selecting a subsection of the image... and not a random one, but one of 9 predefined sections.
All-in-all, it isn't sexy, or special, and we've seen this stuff for decades. It's not new, or ingenious.
2 to 5 people that just might be able to comprehend that page perhaps you are one of them
Well, I understand that page, yes. As I said, it renders on the server.
And lol on this: It is 1000 times more powerful than anything in PhotoShop if you have the color understanding!
What this "it" is that you're referring to is unclear. If the "it" is open source, then you're comparing apples and camels. It's not PS vs OS. I mean it's like saying that the hybrid car industry is 1000 times more powerful than a Prius. Not sure what to make of this. And I don't see why it is required that one has "color understanding"... or what that even means. I mean, any 3 yr old that's not color blind will have color understanding.
But your most irksome statement is of course that my project could be 'dominated' in 2 or 3 months "max". I don't see how that would happen because first, there would need to be a whole pile of someones interested in building the thing.
And then there's the minor problem of doing that math stuff that's so hard. I mean, hard as in almost like gold digging: You may find gold just strolling along, but it's unlikely, and it's even hard to locate it with power tools, explosives, prospecting and all kinds of advanced techniques. Gold digging isn't necessarily a good way to get rich, so to say. It's far from easy. It's even a chance game.
It took me months of laying awake at night and building and tweaking dozens of prototypes to come to a formula that works, and works well.
Even though it is possible that someone else comes across this same thing, I consider it highly unlikely. So all the effort in the world with large OS armies, isn't going to come across a discovery that requires both dedication and luck. It's a very thankless job all the way up to the point where you hit pay dirt. And I'm talking months of nothing except huge piles of effort... no results, no reason to believe you'll get any results, and a lot of reason to give up.
So, open source is powerful, sure, and nice, yes. But it doesn't work well in situations where there's not a lot of people doing work on it. And as you said "there just might be 2 to 5 people that just might..." etc, that's not a lot of people. What's the percentage of those people that a) Have time b) Have programming skills and c) Have interest? Maybe 10% at best... so 5*10%=0.5 people are willing and able to work on it.
Not sure that this 0.5 person can pull it off in 2 or 3 months.
Or, to put it in different words: That link you put in here may be OS (I didn't look, actually), but it has 1 teensie issue:
It doesn't do simulated process seps.
And this is where my project is different. It does do seps.
And to use it requires zero "color understanding" beyond what anybody already can handle when they're 8 yrs old.
OS is by no means "the future". Things that are ubiquitous will likely run a lot of open source (Android), but things that are specialized require specialized knowledge and a lot of effort for a (relatively) small user base.
And I would like to note that your own open source solution is designed to make you money. You will let them have it for free, with the understanding that they will have trouble getting it up and running and you offer to do it for them for a sizable fee. You'll also offer proprietary plugins, I'm sure. Art packs maybe. Who knows what all you're planning, but the truth is you're planning to make money from it. It's a vehicle to get a lot of users. Paying users.
There's nothing wrong with that, of course. You have a business to run, ends to meet, etc, just like everyone else. Moreover, a lot of open source projects are designed to produce income in this very same way. There's a myriad of products that are free and open source up to a certain extent, but for more potent features, you need to pay. Take Riak (http://basho.com/riak/) for instance. If you want to run this in multiple data centers, you need to pay. A lot (http://basho.com/riak-enterprise/). A power outage at a data center will bring down your entire infrastructure, unless you pay the big bucks. The open source thing is a way to get massive adoption. Everybody knows about it. Everybody uses it. References galore. And it works, billions of dollars worth (http://www.wired.com/wiredenterprise/2012/03/red-hat/).
So much for "free". There ain't no such thing as a free lunch. (http://en.wikipedia.org/wiki/There_ain't_no_such_thing_as_a_free_lunch)
And to close, if someone can "dominate" my project in 2 or 3 months "max", then let them do it. The talk gets weary. I have yet to see a shred of it.
-
Han the scripting language in that tool is magnificent it short cuts every aspect of pixel manipulation. It opens up any and every possible math you might want to run on digital color. Just hook it up to a GPU server with one of these http://en.wikipedia.org/wiki/Application_streaming (http://en.wikipedia.org/wiki/Application_streaming) or something like this http://aws.amazon.com/appstream/ (http://aws.amazon.com/appstream/)
It can be made to process seps very very easily it is a pixel math scripting language. Because the color engine is wide open that makes it, among many other young emerging open source solutions so powerful.
I really find it hard to believe that you cannot fathom what a tool like that scripting language really amounts too. I am glad we have had this discussion because we can revisit this in the not so distant future and time will tell if I was correct or incorrect. But the posts will be here and we can rewind in this very thread.
-
Scripting languages don't do anything. It's the programs you write with them do do stuff. Any language that is Turing-complete can be used to make a program to run anything.
Moreover, by far most high level languages are extremely similar, choosing a paradigm such as procedural or functional and then implementing the same stuff as all the others in its class. For loops. While loops. If. Else. Function. Some syntactic sugar gets tossed over top and it may or may not have macro support. Awesome.
This means I can manipulate pixels in any language. As for GMIC, I took a minute or 2 to peek at its language, and it's very similar to what I've seen with ImageMagick (http://www.imagemagick.org/)/GraphicsMagick (http://www.graphicsmagick.org/). Nothing interesting at all there.
And then there's this huge pile of bullcrap you dump by referencing "app streaming". The wiki article talks about only sending the necessary parts of an app to the client device, as in it runs on the client like any normal app. This has nothing at all to do with GMIC. The second link, to Amazon AppStream, is not the same thing at all. It involves doing all the rendering on a server and streaming video of the screen to the device. This is useful if the device can't render complex 3D stuff, such as games. OnLive (http://www.onlive.com/) started (http://www.ign.com/articles/2013/11/19/amazon-appstream-offers-cloud-based-game-streaming) this, btw, not Amazon. It's a useful technology. It has nothing at all to do with GMIC. For one thing, GMIC cannot and will not run on the GPU. Therefore it seems mentioning the GPU is... uh... weird?
So, you mentioned that GMIC can be made to process seps "very easily" since it's a "pixel math scripting language". There's no such thing as pixel math (http://en.wikipedia.org/wiki/Pixel_math). There's pixels and there's math. There's also no color engine in GMIC, although it is capable of transforming between various color spaces. It cannot do anything with color profiles, though. So the color engine isn't wide open. There is no color engine.
But it's not even relevant whether or not there is or isn't a color engine. As a matter of fact, I could probably make some script in GMIC to do color seps. This script could be extremely accurate, possible even fast. I could probably make it as a good as what I currently use.
This isn't, however, because GMIC is so miraculously incredible (or "Magic", as found in it's name, just like ImageMagick and GraphicsMagick), but because I have developed the algorithms for color seps. It's not GMIC that provides the solution, it is me... It's similar to when you write a program in VB for applications in Corel (such as SimpleSeps), it's not VB that provides the solution, it's you.
I can however guarantee that you cannot make GMIC sim-process-separate art "very very" easily. It's no easier than doing it in any other language. It can probably convert to HSB easily and inversion is also easy, so maybe your HSB/HSW/Whatever seps can be done quite simply in GMIC. But it's again not GMIC that makes it simple, it's the HSB seps you do that are simple to begin with.
So, it all comes down to that I can fathom what a tool like GMIC amounts to. I can also fathom that you have no idea what you're talking about. It's becoming very clear indeed.
Therefore time cannot tell whether or not you are correct. The things you talk about are convoluted, vague, and meaningless. You're not saying anything. I can't figure out where you're going with this.
But if I apply my best read-between-the-lines skills, you seem to simply attempt to discredit me (and in the process yammering such nonsense that I couldn't take you serious) and at the same time uplift the idea of Open Source so that you may sell a lot of OpenTShirt sites. I also think it's a prelude to you adding your web seps to OpenTShirts. You'll tout the openness of it all as the ultimate benefit. That it barely works is irrelevant. But even if it worked well, why does this mean you have to engage me with this nonsense?
I'm spending so much time trying to talk reasonably about the nonsense you're spewing. And the truth is, it's really not possible to do that. As the verse says: Do not answer a fool according to his folly, lest you become like him.
And boy don't I feel stupid now. Allowing you to drag me into your whirlpool of nonsense yet another day.
So, it is indeed good that there is a record of this conversation. It will remind me to not be dragged into a conversation of this nature. I'm just hoping that it works.
I note that only you reply in this thread. Those with more wisdom have long left this thread. Only the sound of crickets and your nonsense remain. This is what I get for my folly.
I again request that you leave me out of it. If you feel threatened in any way, take it to your therapist. Please don't talk to me about pixel math. I know math, I know pixels, you need not explain to me. I'm good. I do not require your teachings. If I were to require teachings, I would not request them from you. I'm sure that you have better things to do. I know I do. Go do the better things. Tell your pack of PhDs to use pixel math engines. Build Open Source. Teach the industry about how wrong they are. Go ahead, but leave me out of it.