Patterns and color 16

[Philippe Hurbain]

Just wanting to add some stuff to this discussion:
I'd like to see clarified the status of color 16 for sticker bounding box. This idea (originally proposed by JC Tchang) has already been discussed here and there, but no clear decision has been made. Point is, using color 16 brings more flexibility for stickers, as they allow user to choose
- color of sticker dominant border color (as suggested today in spec),
- color of the underlying brick (useful to make borders disappear if sticker has several colors on border,
- or even... for purists, white to show the paper white border!

[Orion Pobursky]

I might have missed this discussion but why 16 for transparent part of stickers? Why not 31 (Trans-White)?

[Roland Melkert]

#31 is not really transparent in means it will become light red when the sticker is on a red brick.

edit: i just realize in most renderers it actually would turn up light red because of alpha blending. So maybe this is an option after all.

[Philippe Hurbain]

I guess you mean color 47? (31 is lavender). The problem is - as suggested by Roland - that because of alpha blending the brick color gets washed out, while physical trans stickers doesn't modify brick color.

[Travis Cobbs]

Any LDLite color from 0x7000000 to 0x7FFFFFF is invisible, and would do what you suggest. If someone using a sticker with a color 16 background uses color 0x7000000 as the color for the sticker itself, the background will disappear in renderers that support those color codes (like LDView and probably L3P/L3Lab).

[Steffen]

I support the color 16 suggestion, because it leaves the choice to the user.
----> Flexibility.

[Jean-Philippe Ouellet]

+1

I think it would be better to recommend using 47 when including the sticker in the model, and make the background in the actual part 16.

[Orion Pobursky]

Yes 47. I added 16 not 32. Doh!

[Steffen]

ugh, I am just finding this separate thread

http://forums.ldraw.org/showthread.php?t...23#pid3723

which currently seems to go the opposite direction - which I do not like very much :-|

I would definetely prefer to use color 16, even for non-transparent stickers.

This lets the choice to the user and is the most versatile approach.

I cannot see the benefit of hard-coding all colors.

[Don Heyse]

You could make a subfile with 16 and (maybe hide it in a special stickers subdir) and then import that into the real sticker file with trans-white. Then you've got an option to substitute the 16 file for the real one for better rendering if needed. The actual substitution could possibly be handled by software as an option in the renderer, or by the user if they have old software.

Having a real sticker file with trans-white could be handy for LPUB (or other) inventories.

[Steffen]

no, this would mean 2 files per sticker always.

using simply color 16 for the backgrounds produces just 1.

and the user of that file can decide which color to use
(e.g. black, white, transparent or something else)

[Travis Cobbs]

It may not be obvious, but what the LSC is voting on in the other thread is intended as a clarification of the current rules, not a change to the rules. That doesn't rule out the possibility of changing the rules to allow any sticker background color to be modeled as color 16. I'm not sure where I stand on that right now, to be honest.

[Don Heyse]

I'm not sure what exactly you did vote on. Was it a clarification for the fire engine sticker where the flame color is transparent and meant to be stuck on a red brick?

Anyhow, since it's not ruled out, I think using color 16 for the transparent bits makes sense. Then if you want a "real" sticker for show in an inventory list (or in instructions before it's stuck onto a part) you can use a trans-white sticker part so you can see the whole thing. Then make it invisible by using the color of the part underneath when it's attached. The program (LPUB?) should be able to do this for you, as long as there's something in the name or the header to identify it as a sticker.

Of course this might mean you can't reproduce the real instructions where they cheat and replace transparent flames with red and leave the other transparent bits trans-white, but I'm ok with that.

Keep up the good work.

[Roland Melkert]

I too was under the impression this is only a guideline decission because the part authors couldn't get too a agreement.

I voted yes because this way stickers are self contained, meaning you could place them on any part or assembly using a reference with some transparent color in order to blend the sticker with the whole without messing up a part of the pattern.

[Allen Smith]

I cannot see the benefit of hard-coding all colors.

In DAT code, the only reason I can see to have color 16 is on fully opaque stickers which come in multiple colors. I can't think of all that many (any?) examples offhand, but I don't pay close attention to my stickers either. (They often just go in a bin of mine to be forgotten forever.)

If DAT code is ever finally replaced by texture images—which would bring many, many advantages—every sticker/printed element color would be entirely hard-coded except for transparent bits. If one were to allow dynamic sticker colors, it would impossible to convert such parts into textures without some complicated texture META which I don't really think is worth inventing.

Allen

[Steffen]

no, the usecase here to use color 16 is because different users want the main color differently.
some prefer white, some prefer "the predominant sticker color" etc. pp.

[Steffen]

The voting there also forbids to use color 16 in non-transparent stickers.

But exactly that is what I'm asking for.

We should allow color 16 there.

Reason: currently there is no real consensus about what color the back and side sides of a stickers should get.

By any rule which forces "white" or "black" or anything else here, we will *always* disappoint someone.

Instead, by using color 16, we can make every fraction happy.

That's my reasoning behind that. I really would dislike if stickers would be hardcoded to a color I won't want,
as that would mean that I would have to modify those official parts.
This is not something I would want to do.

That's what basically the color 16 is for at parts,
and I see no reason to not use that mechanism for stickers as well.

[Allen Smith]

no, the usecase here to use color 16 is because different users want the main color differently.
some prefer white, some prefer "the predominant sticker color" etc. pp.

I'm confused. Can you explain what you mean by the "main color"? What parts of the sticker are you wanting to recolor?

Are you referring to a completely opaque sticker, or a sticker with transparency? (I understand the use of 16 for the transparent background.)

Thanks,
Allen

[Don Heyse]

The wording voted on mentions the sticker pattern and printed colour. I'd interpret that as the top of the sticker. The wording doesn't really say anything to me about the sides or the back of the printed part. Does it?

[Steffen]

I am referring to both transparent and opaque stickers.
I am referring to their rear and side color, which sometimes extends to the front.
For transparent stickers, color 16 is needed there and currently officially allowed / required.
Using color 16 for that allows the user to set that color to many things he/she likes. That can be:
the color of the part behind the sticker, or some transparent special material or color or whatever else.

My request is to allow that same color 16 use for non-transparent stickers as well.
There are some on the Parts Tracker where the parts authors can not find a common consensus
to use either e.g. white or black for that background. It is a question of personal taste.
Some people want to mimic the white plastic sticker material, and also want to see that at the sticker sides,
some people want the sticker to better blend with the background and choose a completely custom color,
and some people want to use yet another color, because it predominantly occurs at the sticker pattern
(e.g. especially at its border).
This non-consensus for me is a clear indicator that using color 16 there would make all people happy.
Everybody can use the color he/she wants. And we already are using the same technique for parts.
I don't really understand why extending that technique to stickers should be a problem.

I hope I made myself clearer now -
best,
Steffen

[Willy Tschager]

The base for this entire fuss is a simple misunderstanding. While we fixed the gap with the fire logo pattern Steffen was talking about a change in the first paragraph of the sticker specs. Since our editing hasn't change a single comma in the current color restrictions for the surrounding sticker box trying to interfer with the ongoing voting was IMHO completely out of place. What Steffen is suggesting is changing:

The sticker is a 0.25 LDU thin box (or whatever shape sticker is in), oriented such that the top face is at -0.25 Y and parallel with the X-Z plane. The pattern is on the top face; the other (five) sides of the box correspond to main background colour of the sticker pattern (i.e. the predominant colour of the pattern at its edges). If there isn't a predominant colour at the edges of the sticker, the actual colour of the sticker material should be used instead (usually white).

to

The sticker is a 0.25 LDU thin box (or whatever shape sticker is in), without any linetypes 2 (edgelines) and coloured with main color 16. The sticker is oriented such that the top face is at -0.25 Y and parallel with the X-Z plane. The pattern with the hardcoded colours is on the top face.

or something like this down the line. Right?

w.

[Philippe Hurbain]

Right?

Right for me!

[Allen Smith]

Ahh, I get it now. You want to be able to recolor the sticker substrate for stickers printed on an opaque material. That sounds reasonable.

Allen

[Travis Cobbs]

I think he also wants the "background" color for the sticker pattern itself to be color 16:

I am referring to both transparent and opaque stickers.
I am referring to their rear and side color, which sometimes extends to the front.

The "which sometimes extends to the front" above means to me that he's referring to that in addition to the back and sides.

[Willy Tschager]

That's what I'm talking about. To make it simple he wants to see color 16 used for all sticker boxes like I used it her:

http://www.ldraw.org/library/updates/vie...16816a.dat

0 // Thin Box
1 16 -16.5 0 14.5 -1 0 0 0 -1 0 0 0 1 1-4disc.dat
1 16 16.5 0 14.5 0 0 1 0 -1 0 1 0 0 1-4disc.dat
...

(where it has been used 'cos the actual box was transparent).

w.

[Steffen]

I just wanted to answer at
http://forums.ldraw.org/showthread.php?tid=3857,3902
but couldn't due to forum restrictions :-((((((

In your 3rd example there, the background substrate is visible at the front,
i.e., at the sticker with the crossed swords and the wavy black outside border.
Outside those waves, I'd like to be able to use color 16, even for the front surface.
I think the current spec text has a gap there, it doesn't treat that case.

[Philippe Hurbain]

I don't quite get it - if we do see the white paper outside wavy black line, then shouldn't it be modelled white?

Otherwise, here is another interesting way of doing things... http://www.ldraw.org/cgi-bin/ptdetail.cg...13857b.dat

Philo

[Willy Tschager]

Steffen,

I perfectly know what you're talking about - problem is how do you define what the substrate and its color is? Taking these:




and following your request there would be no RED or GRAY used but color 16. It would leave the decision to the part author and no rule is required. How do we define what is substrate and what is part of the pattern? What if the author has no substrate at hand but just the sticker attached to his space ship from '79?

w.

[Travis Cobbs]

If color 16 is to be used for the sides and back, what about requiring (not allowing, but requiring) color 16 on the front for any portion that matches the color of the sides and back (which also means there's presumably no printed ink there)? There would then be no room for questions about what color the substrate is. It is the color of the sides and back.

[Steffen]

Yes, that's what I had meant.

[Don Heyse]

I hate to poke this thread, but I'm curious about Willy's pictures. What exactly are we looking at? Are those transparent stickers on red and grey backing sheets? If so, color 16 still makes the most sense for the sticker because the backing sheet is a separate bit, just like the brick the sticker eventually gets stuck on. If you want to model the backing sheet you have 2 easy choices: A very flat box, or just import the sticker with a real color like red or grey on the type 1 line.

[Don Heyse]

And this reminds me of a similar issue. Do we model the disposable sprues attached to the minifig tools, flippers, coins, and flowers? That might be handy for inventory purposes, just like the disposable sticker backing paper.

[Steffen]

We currently do not model backing paper and sprues.
If you'd like to discuss that, could you please start a separate thread?
These things would bring this discussion here completely off topic.