Sticker Standard Questions

[Greg Teft]

I just made the detail geometry for my first sticker, and was considering uploading once I wrap up the loose ends.
Before I do, I have some questions that the standard didn't answer, and after looking at some certified stickers was uncertain if these considerations would cause rework, or if it didn't matter. I searched for "sticker" on the forum, and didn't see these asked by a quick scan at the topics, so I apologize if this is old news.

I've attached my almost-done sticker in context with a brick. Yes the origin and header are wrong. As I said, it's almost done.


file name - set number or sheet number? The standard says sheet's SKU number, but I see certified files having used set numbers. Is the set's number for where the sheet's SKU is unknown? And which is the SKU number? Looking at 10221's sheet, there's 22637 in one corner and 10010521/6005049 with the copyright statement in another. I'm leaning toward 6005049 based on 4622303a.dat.

file name - If I intend to create more from the same sheet, can I name them out of order so that the final sequence makes sense (as called for in the instructions)? Or should the first uploaded be 'a' and so on. Alternately, should I upload 'needs work' placeholders for the ones in the sequence I want to use.

round corners - yes or no? Most stickers have a 2 LDU radius rounded corner, didn't see any on the files I looked at. Extra geometry to render, but I wanted to consider close-up rendering.

edge lines - yes or no? The files I looked at didn't seem to have them. I used edged primitives for the sides,
as I think edges add detail if when ambient lighting or flat shading is used.

inset - yes or no? I've noticed that most of the stickers I've looked at are inset at about 2 LDU from the edge of the element face. Is the sticker to be modeled to the edge of it's target element, or to the edge of the sticker considering the inset?

geometry and obvious color offset problems - looking at repeats of the same sticker for 10221 (and the shield for 10214), should the model correct for blatant printing issues where the ink is not where it should be, or leave out where it bled into a neighboring space that might not have been masked well?

sub-parts - I'm still not sure where the line between file size and number of files is. 10221 has three stickers that vary only by the color of some of the details, and two that have slightly different detail. I'd prefer to sub-part everything but the varied details, but not sure if the CPU cost of loading another file and applying the transformation is worth the smaller memory footprint. Additionally a couple can be mirrored (the doors), or used some mirrored geometry (the windows).


I was also considering a primitive for certain common sticker sizes (1 x 2 tile face), that is if rounded corners can be used. For 10221, I created a couple of seed files for the back and sides, but would prefer to use a sub-file rather than repeat the geometry. If rounded corners are prohibited, then the 'box' will suffice.

[Travis Cobbs]

The sticker spec is very clear on the file name prefix. So the certified stickers that don't match were either certified before the spec existed, or were certified without meeting the spec. I'm not sure about skipping suffixes.

I personally feel that there shouldn't be round corners, but I understand why you might want them. Also, since color 24 is disallowed in stickers (see next point), they wouldn't be allowed to use cylinder primitives.

I believe that edge lines are disallowed by the sticker spec. Color number 24 is never allowed in stickers, and it states that only real colors are allowed. A hard-coded color for an edge line around the sticker doesn't meet the "real" color requirement. (Note: I believe bullet point 3 in the sticker spec was intended to replace bullet point 4, but the editing was done incorrectly.)

I'm not sure I understand your inset question, but I feel the stickers should be modeled to as closely as possible match the size of the actual physical sticker.

I believe that correcting blatant printing issues is done for patterned parts, and if this is correct, I think it should also be done for stickers.

Sub-part preferences can start holy wars, so I'm staying out of that, other than to say, "Use your best judgement."

[Greg Teft]

Thanks for the answers Travis.
I interpreted the use of color 24 to apply to the sticker's file, rather than primitive sub-files. Your explanation makes sense to clarify the rule for me.
I'll work on some other things before I rework, and see if there are any other opinions on the rounded corners before I swap out the primitives.
Your 'close as possible' remark answers the question about the sticker size with regard to the inset from the edge of the element.
In retrospect, I wouldn't want to skip suffixes, but I started with what appeared to be the easiest sticker on the sheet. I can wait until I've got the others in a usable state before I upload any.

[Willy Tschager]

The standard says sheet's SKU number, but I see certified files having used set numbers. Is the set's number for where the sheet's SKU is unknown? And which is the SKU number? Looking at 10221's sheet, there's 22637 in one corner and 10010521/6005049 with the copyright statement in another. I'm leaning toward 6005049 based on 4622303a.dat.

I used the second number for the Fort LEGOREDO stickers:

file name - If I intend to create more from the same sheet, can I name them out of order so that the final sequence makes sense (as called for in the instructions)? Or should the first uploaded be 'a' and so on. Alternately, should I upload 'needs work' placeholders for the ones in the sequence I want to use.

No, go with "a" for the first sticker you submit to the PT.

round corners - yes or no? Most stickers have a 2 LDU radius rounded corner, didn't see any on the files I looked at. Extra geometry to render, but I wanted to consider close-up rendering.

Hmm ... a matter of taste. IMHO too much geometry.

edge lines - yes or no? The files I looked at didn't seem to have them. I used edged primitives for the sides,
as I think edges add detail if when ambient lighting or flat shading is used.

No edges and no edged primitives - "but color 24 is not" in the specs stands for this.

inset - yes or no? I've noticed that most of the stickers I've looked at are inset at about 2 LDU from the edge of the element face. Is the sticker to be modeled to the edge of it's target element, or to the edge of the sticker considering the inset?

Go for the real figures.

http://news.lugnet.com/cad/dat/parts/?n=5692&t=i&v=a

geometry and obvious color offset problems - looking at repeats of the same sticker for 10221 (and the shield for 10214), should the model correct for blatant printing issues where the ink is not where it should be, or leave out where it bled into a neighboring space that might not have been masked well?

Looks like that most people prefer the ideal.

http://news.lugnet.com/cad/dat/parts/?n=6047

w.

[Greg Teft]

Thanks for the answers, Willy.
I forgot all about searching for these answers in lugnet since I started back up with LDraw this winter!!!

That last post is quite the eye-opener.
I have to say, a strict, or nit-picking review is discouraging. Especially in the grey areas where the standard can not be used to mediate (sub-parts). The community should be happy people are still out here modelling, and not make this hobby painful. It's not a job!!!

I'll try just using a box for the back and see. I do think a 2 LDU radius isn't worth the overhead of pulling in the 1/4 cylinders, chords and discs, but I didn't want to be wrong.

With all the drive toward accuracy, I would think rounded corners, and white sides/back for white sticker sheets would be the preference!!! And looking at parts like the 1 x 1 plate w/ tooth, there's an edge line right down the middle of it, and it's not even a sharp edge!!!

[Travis Cobbs]

With all the drive toward accuracy, I would think rounded corners, and white sides/back for white sticker sheets would be the preference!!! And looking at parts like the 1 x 1 plate w/ tooth, there's an edge line right down the middle of it, and it's not even a sharp edge!!!

You have to understand that stickers in LDraw are a compromise. Real stickers aren't 2LDU thick, but LDraw stickers are, and the reason that they are is due to rendering considerations. (Making them much thinner would likely result in bad rendering.) So having white sides and back would result in a visible white stripe that's much thicker than the actual sticker. Similarly, edge lines are disallowed to prevent the eye from being drawn to the edge of the sticker.

[Philippe Hurbain]

Real stickers aren't 2LDU thick, but LDraw stickers are

Ahem... LDraw stickers are 0.25ldu thick!

[Greg Teft]

after reading some of the old lugnet posts, I get the reasoning behind these restrictions.
I'll clean up what I have and see what the reviewers say when I get the first one in.

[Travis Cobbs]

Whoops. My bad. I suspect real stickers are less than 0.1mm, but it's not nearly so clear-cut.