Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[just champi]

Hi, I don't know if this has been asked already, but as I'm curious, I'd like to invite (mostly to people used to work with skins, but also anybody sensible to the issue) to comment about:

what kind of things, have you encountered when working with skins that makes your life harder or easier? can you point to some good and bad examples?
For bad I'm referring here to issues that you think can be improved, fixed or avoided, in the previous step to skinning,  the UVWmapping . And for good, as examples to emulate and from where to learn good practices.

A kind of "more desirable versus less desirable" things that you like to find when having to do a skin from a mapping layout.

 The main constrain you face when doing UVWmapping is size of the texture versus size/complexity of the model, and how to make fit all the parts inside that given space while still retaining enough detail visible in them.

The more complex/bigger the object, the harder it gets, as you have less resolution at disposal. So, to maximize the space as much as possible, sometimes you can find yourself with some strange arrangements for some parts of the skin in the end. I'm lacking the perspective from the practical side of texturing in a decent level, but I believe that maybe this is not so good for the people having to work with this later on.

And seeing the high quality, the amount of work, research and time invested in the actual skins some people are making, I think it would be better, to try to make things a bit easier for them, when/if possible, so that's why I'm asking for opinions, as it would be preferable to take them into account from the start, and include their guidelines (if any) in the process.
So, feel free to post your thougths...

Thanks in advance!

[SAS~Loku]

Few things i hated most when was doing skins:flipped,mirrored and shared areas,best examples are: p-38`s mirrored/shared tail booms,p-39 mirrored rudder,p-40 mirrored underwings,p-47 flipped and mirrored fuselage areas.Having this in mind im always trying to make uv mapping of my 3D models with simple logical layout so skins are easy to paint.Making a "skinnerfriendly" skin is a big + and will pay back with good number of skins made for your aircraft.

[Rock]

Champi, and excellent post and probably a future sticky?

The most important thing about mapping is that there is a little stretching / warping as possible. This deters skinners and makes it bloody hard work for the ones that do choose to tackle these aircraft. The first bad example of stretching that springs to mind is the Spitfire Mk I, II, V and IX fuselage which is mapped in a rectangular shape. As you can imagine, because the fuselage gets narrower, mapping it in a rectangle creates some pretty awful stretching  towards the tail end. (Take a look at any marked skin for this aircraft and you will see just how awful it is).  I know you probably already know this, But it is much better to map the fuselage in 4 sections - Top, Bottom, Left and Right then to try and map in a cylindrical fashion.

Having mirrored parts such as the P-47 fuselage is more of an inconvenience than an issue as it is relatively simple to flip text and images around although if you can avoid this in the first place it would almost definitely save the skinner some time and frustration.  As Loku mentions though, Double mapping (Such as the P-38 tail booms) is a major issue which will put most skinners off.

For me, The next most important thing after avoiding Stretching/Warping is Scale. It's critical that all the parts that you map to the 1024x1024 bitmap (except internals) are all in the same scale. It's pretty much pointless having the wings take up 3/4 of the page with a tiny amount of space for the fuselage as you end up with a very patch work looking skin with the wings looking very sharp/crisp and the fuselage looking very blurry.


I'm sure you know most of this already but as a Skinner this is what I normally look for when I map an aircraft

PS, Something you may want to look at is mapping all internal parts such as the internal cockpit on the exterior model, Gear bays, flap internals and wheels etc... onto a separate TGA file which can be kept in the mods folder. This servers 2 purposes: It allows you to have larger resolution internals so they look much better and also it can leave you with a lot more space on the standard 1024x1024 bitmap for the regular skin. I wouldn't suggest making this internals TGA too large as for older machines, loading more textures may slow them down but when mapping my Hawker Fury project I used a 512x512  size and found this to be suitable.

[Rudi_Jaeger]

Some very good examples have been mentioned so far.

To emphasize what Rock said about warping.. I've always built my own templates, despite it being a time consuming process; and the first thing I do when considering a candidate for a template build, is to throw a 'grid skin' over the model to see how much warping is present. If the warping is mostly 'uniform', like a symetrical bubble for example, then it's workable; I don't exactly look forward to it, but it's workable. But if it's more like a crumpled piece of paper, then forget it; it's just not worth investing the time when there are other subjects available. 

Another aspect, is the 'mechanical' quality of the 3D model itself. Are there any gaps/holes in the mesh, like the triangular section on the inboard portside engine nacelle of the A20G? Are there parts suspended in mid-air, like the venturi tube on the I-16? Are there 'non-model' objects protruding through the mesh, like the pilot's knuckles on the portside fuselage of the B-534? ..of course these examples are minor, and not necessarily 'show-stoppers', but I have seen a few models that were. So not only should the model be 'skinner friendly', but it's quality of construction should also warrant the effort of laying out all those panel lines and rivet sets, not to mention all the research involved.. and the model might even be easy to skin, but if it has obvious mechanical flaws, then it's not likely to see much use. No sense skinning something if no one is going to use it.

Lastly, I'll mention 'light mapping'. Something that's always been a real pain for me, is working around that blasted 'nose cowl ring' on the SBD.. the lighting just doesn't match the rest of the aircraft. No matter how I compensate for 'brightness/contrast', it never seems to be quite right when viewed in-sim; it's like a dog chasing it's tail.. one correction only leads to another. I've seen this on a few other models as well, but having done quite a few Texan/SNJ hacks, this is the one responsible for most of my hair loss.

Champi, it was very considerate of you, to ask for our input. I don't know the first thing about model building, but if this helps to make a better product for the 'brush slingers' in our community, then you certainly have my thanks.

[Knochenlutscher]

I'm no Template creator, but Template-user, like mentioned before, stretched areas are not very helpfull.
Mirrored Parts, hm, not the big deal.
Best to work with are scaled like, rolled up, but logically.

The Spitfire, tried my hands on it, honestly, I don't have the time tinkering, even with the recently published UV-Maps, Help-Lines, guides and Examples added.

Another not so heavy, but still frustrating Template, Me 262, everythings fine and sorted, but the cut into pieces Nose, then the complete fuselage upper and undersides seaming is differently stretched, not just one value, but actually like a rattle-papercut, it has wavy, unliniar trace. This means you need to tinker when applying large/historical Balkenkreuze adding markings before the pit, cut in shape, stretch, urrghh... you end up using smaller sized, cuz you can't exactly cut to this papercut shapes on undersides-rear fuselage. I spent 2 entire weeks on just this prob, still I'm not satisfied.
Due to this stretching you can't add proper Stencilling, i.e. oxygen panel on bottom of fuselage too, it would look like a blotch of colour, not something fine.
What I feel it leaves too much unused space too.
Another thing, but is 1C's fault, the gun-shell ejector ports are lil way to the nose, misplaced. I took pics of the Munich Mutke Me 262 and it's panels/sheet covering on the Nose. In real relation, compared to panelling and these ports, there's a lag. Turned out one segment's Panel is faulty as well, placed around the faulty ports. MK 108 ingame on the 262 are very short-barrelled, probably not in scale. The stock 3D for the 262 could have lil overhaul too, it's such a waste, such a tremendous and outstanding craft, but such a bricky Nose-cone. 
As the Me 262 turned out to be my preffered ground, it's frustrating and sad to work with it, simply because it slows the creative process, two sides of a coin. But humans accomodate with such stuff, just see the Spit-Skinners.

A well known IL-2 example which is just a complete horror for beginners and we don't have smileys that show what I think or feel.
Me 163 Komet mapping. Try adding a rear fuselage Balken in it's exact predefined scale... Respect to the author, but I ask myself what stuff he was on, when arranging it this way, the most neat area?! This Mapping is only good for unmarked one-colour Schemes, historical use?, expect some probs. 

Real good Templates, as for research done, 3D model and Mapping in comparison, to work with in my preferred manner, are the Do 335, Ta 152 H-1, Bf 109 and Fw 190 Templates, simply perfect, and easy to understand. Have some slight mirroring, but nothing crazy.

+1000000000³ for your Post, thank you very much on behalf of us skinners.

[Jarink]

In addition to the above suggestions about avoiding stretching, mirrored or re-using parts, let me add that the mapping should be consistent and in a fashion that's relatively easy to follow.

If you want an example of what I mean, let's look at the B-17 engine cowlings.
Look closely and you'll notice the part right above the green area is flipped backwards from the rest. What you can't see from this picture is that it's taken from an area immediately adjacent to and overlapping some of the other parts.

This makes it effectively impossible to properly skin the cowlings.

Also, don't be afraid to map small details on the fuselage or wings to unused areas of the skin! The oil cooler intakes on the leading edges of the B-25 wings are a great example of how doing this allows for more detail! I also feel that pop hubs should always have their own spot on a skin, as they were often painted different colors (much like spinners were).

[Phas3e]

one thing I think is important is to have the main wheel mapped on the skin. often the type of hub and tread changed so its nice to be able to reflect that.

[just champi]

Thanks for the replays, very helpful and interesting (at least for me...)

I wasn't aware of this different kind of issues with stock skins, and I also had some wrong assumptions about some things, like with mirrored fuselage parts. I though that was desirable for alignment of panel lines and easier painting of camouflage areas, in fact, I've seen this in so many skins, that I was using this approach when testing layout for skins as a kind of "mandatory step" thing...

 By the way, I've just remembered another one while reading this: the Ju 87 rudder has one side mapped in two pieces and the other side in three pieces scattered around and rotated, in consequence I remember seeing a bunch of skins that didn't recognized the third small piece as part of the rudder so it wasn't painted as part of it.

I've readen the whole posts so far with great interest, and I've found this kind of info as very valuable for practical purposes, is the kind of things you pass unaware, and maybe won't get by yourself if someone doesn't tell you about. I've copied all of them because I want to have a closer look to the different examples you talk about and read it more slowly.

I'll post further comments later on after some more tinkering and thinking and a couple of screenshots maybe, as your replays trigger more questions. (a bit tired and slow now)
very glad to have this kind of feedback, very appreciatted!

[Pursuivant]

So,

* Don't mirror parts of the skin that are likely to be painted by skinners.

* Don't flip parts of the skin that are likely to be painted by skinners.

* Try not to stretch or warp parts. If you must, do so in a simple, consistent fashion, and reduce stretching and warping as much as possible.

* Try to avoid "cuts" to the skin when mapping it to a particular part of the plane.

* Be consistent in your skinning methods, so that skinners don't have to guess what mapping techniques you used for each part of the skin.

* Use a consistent scale for all parts of the mesh, especially for large parts of the aircraft such as the fuselage and wings.

* Make sure that there are no gaps in the skin mesh.

* Make sure that internal parts of the model don't stick out through the airplane's skin and that all parts are actually attached to the rest of the plane, so that the skin will look good.

* Map small details, such as wheels, spinners/propeller bosses, cowls and cowl flaps, separately.

* Be aware of light mapping when skinning.

[A1_Phoenix]

i must add one thing: most skinner friendly mapping i've found are 109 and hurricane, so are good examples.

the only minor (very minor) points in these are mapping of props (109) and nosecone (hurri), the first being a simple pair of pixel with consequent solid look of the propellor, and the front-viewed nosecone of hurri has an irregular mapping, so is difficult to make a (for example) wooden prop spinner, the separation between nosecone tip and wooden part will never be straight (eg: my early hurri's pack).. maybe a spitfire-type nosecone mapping seems to me the best choice.

useful thread champi! thank you!

[Jarink]

Another excellent skinner-friendly example is the P-51. There's only a couple of odd spots (like the wing root trailing edge and the fairing between the belly scoop and underside of the wing), but they are easy-peasy compared to the mappings on some planes.

[just champi]

Hi, and sorry for the delay. I've been forced recently  to substitute my fried card with a older one and now is a real pain to work with any graphical applications, as I'm having lags, freezes, lock ups and crashes all over the place, so this has cost me more that I though.
Anyway, reading your posts the other day, I though that it could be a good thing to have some images to look at when reading about this issues, and Jarink's example convinced me even more, so I've taken "a couple" of screenshots to help ilustrate some examples.
I hope it's clear enough that I'm not playing the smart ass that points what others made wrong, that's not the spirit of this nor my way of thinking about it.
 I rather think that having a better perspective helps to have a better understanding, so better take this as a a learning approach, in the thinking that, knowing the matter of UVW mapping for skins a bit better, maybe we can try to avoid doing some kind of things or have clear and informed reasons for preferring some aproaches to others, plus, a bit of sharing thoughs and feedback can't do no harm...

I've picked up a couple of Loku's examples to start with, first one the P40, so here's the P40E skin with his shared bottom wing:



I think that this layout aims for the greatest resolution possible for the main parts, reducing his number, making big UV shells, even to the point of having to sacrifice some of them, -like the missing bottom wing and the bottom of one of the stabilizers- to allow the necessary space for the smaller ones.
Strangely, also the bottom of the stabilizer is scaled along his lenght to make it fit in the texture.(compare the shape with the upper counterparts)
The bottom of the fuselage is mapped from a side projection wich introduces strecthing in the areas involved.








This issue of having parts lacking a projection from the right direction for all his sides is a recurrent issue encountered in many examples.
It's interesting to compare this with the P40B skin, so similar in shape but with very different layout in skin:



This example has a smaller resolution in general for the bigger parts than the previous, but in this case you can find that every main part has his space in the skin with no shared uv's.



Also, it has bottom side mapped from a bottom projection, making this area as easy to skin as any other. (a small area at the top of the tail is still mapped to the side, though, but maybe is hardly noticeable)





It appears to have more parts mapped and with unique UV space than the anterior example. (cockpit interior as example)
it seems too a good example of what Jarink and Phase commented about dedicated space for both sides of the main wheels.
both examples have mirrorred fuselage sides.

One question that comes to my mind is, do you prefer to work with symetrical layouts (like for the wings arrangement) over a bit bigger ones but no symmetrical? (both cases with no mirrored or shared main parts)
I mean, sometimes is possible to achieve a bit more of resolution if you don't put things symetrically arranged but filling every possible space rotating or moving the parts around.

The other one I've picked from Loku, and also Rock mentioned, is the P38 shared and mirrored tail booms, P38J in this example.



As you can see, there is only 2 sides mapped for both tail booms (keel and rudder included) so we'll face some inverted sides. maybe it would have been preferable to mirror half of the bottom horizontal stabilizers (as they don't have unique features) to try to have space for non mirrored keels and rudders, as usually they carry markings on them. but I don't know, what do yo think about? Liberator double tail comes to my mind, with half inside and bottom parts shared but top and outside parts with dedicate space.

Apart from this, here we encounter again this aproach of mapping the top and bottom parts from a side view with the consequent stretching for top and bottom faces. Orange marked areas in the skin are top and bottom related faces, so you can appreciatte the amount of texture space given when projected from the side. The red faces in the mesh view shows the related top area involved. (Only area on base of turbo is mapped from top projection)
Here you can appreciatte the effect when a texture is placed , regular-sized squares are shown on top of wing:



mirrored fuselage sides is used in this example too.

The next example I've pick up is the spitfire from Rock's post, the model IX. So here's a view of the mapping of fuselage and tail:



I think that the problem comes in fact, from making a cylindrical projection for the entire engine-fuselage-tail.
 The effect of perform a cylindrical projection in a regular cylinder is a  rectangular shape with more or less wide and length, as you can see in the skin. This is useful for meshes with some resemblance with regular cylinders (like engine cowlings, wheel struts, machine guns canons or many others). But in any other cases some amount (being big or small) of distortion is expected.

In the spitfire case, these UV vertices of the tail part, being stretched for the cylindrical projection to conform that rectangular shape, (while the tail, has a truncated-cone-like shape) is what introduces this weird distortion.
Here's a side view with a texture in place to help show the result:



Now some assorted examples about small technical flaws while doing a model, that Rudi jaeger mentioned in his first post:

A20G:


I-16-24, (model 5 doesn't have the issue):


Avia B-534:



And another point from Rudi's post of interest to me, "light mapping issue" with the engine cowling on SBD and alikes:



Rudi, I don't know if I understand well this issue. the problem is only:

- having the front part of the cowling (the front faces of it), using planar mapping? (while the rest of the cowling other than the front, uses cylindrical mapping)
- having them mapped to a separate area of the skin than the rest of cowling?



-  added to this, the fact that the cowling is mapped in a cylindrical fashion, and being in a separate area than the rest of fuselage (plannar mapping here), with the same end result than before?



all of this making you having problems when painting light effects on it, consistent with the other separated parts?

or all the above approaches added one to another?

for me is interesting to know, because having the cowling separated from fuselage, may help in having a bit more resolution for the entire model, as the lenght/pixel ratio increases a bit with this approach, and this is proportional for all the parts. I'm used to make a kind of puzzle game throwing pieces cutted from 3views in photoshop and playing with them trying some arrangements to see the possibilities.
Here you can see an example from Hs 126 mapping test, from starting point to a couple of iterations:






As you can see the last example has a bit bigger parts than the previous ones.
Hs 126 is just one example. but there are many others planes that face the same situation.
So I'd like to have a better understanding about this light mapping issue to help me decide the most preferable/less drawbacks for anyone approach.

I've taken the next couple of examples from Knochenlutscher post, but I think both deal, more or less, with one important point that Jarink makes in his first post: inconsistent mapping
First one is Me262. As you know, me 262 has a triangular shape when viewed from the front, so maybe that's why the autor (respect to him) has chosen to map the Me only from the sides and from the bottom. The only part mapped from the top is the area around the holes of the cannons, presumibly because, if mapped from the side they will be very stretched.



Like before, the orange marked areas corresponds with top and bottom faces but mapped from a side projection, so having some amount of stretching on them.
An talking about inconsitent mapping, in the case of the bottom rear that Knochenlutscher commented, you can see, that while the front bottom part is mapped from a bottom projection and everything is right there, the rear part that faces the same direction is not, because is has his faces mapped from the side, so this tiny area is badly stretched. (my condolences...):




The other infamous example is Me-163. (your post made me laugh about it) In reality is not that bad. I mean, I think the me-262 is a worst case than this, due to his inconsistency, and both are caused, probably having in mind to save space while making the mapping.



The shape of me 163 in his rear part is like made of three shapes: one rectangular stretched box for the bottom, one truncated cone for the middle and one half of a stretched cone with different diameter for the top part.



The problem maybe comes from simplify this complexity having for all of them a single planar mapping viewed from the side.



This introduces some stretching mainly in the rounded areas (yes, this includes the area where the historical markings should go... my condolences again...)

The B-17 problem posted by Jarink speaks by itself, so I guess no further comments on this is needed.

Another interesting point is given by A1_Phoenix about the mapping of prop spinners, so I've taken a couple of screenshots from both examples to show these two different approaches and the final result (I left out the bf 109 example). First one is the mappping layout for the prop spinner of a finnish Hurricane MkI:



and second the mapping of a prop spinner for a british Spitfire Vb:



and the result (hurricane MKI):



this shows that it's a bit difficult to draw something in the base of spinner without distorsion, maybe better for one colour spinner without details and probably for painted spirals (someone can tell about this?)

(spitfire Vb):



you can see that there is some small stretching going on, that runs toward the tip of the nose cone, but shared in an even fashion.
The main advantage looks that the rear area, where most detail will show, doesn't show noticeable distortion and allows details to be painted there, and you can draw lines going around the spinner too.

I would also add to the sumary made by Pursuivant (thanks):
 
-uncheck "normalize texture" when doing UVW mapping for consistent scale. Wait till have everything UVmapped before scale everything down.
-always maximize space in the texture. unused or misused space will be loaded anyway... so better make use of the whole of it as much as you can (without become obsesed) and make it worth for this full 1024x1024 pixels to be loaded in memory

and a couple of question to finish this looong post:

-do you usually work in the same scale and resolution of the skin or do you work at a bigger resolutions?
-As you know many single engined planes have some way of deal with torque, one of this being, having an offset to some parts, as engines (many italian planes and bloch 152 as extreme example) or tail surfaces (most of them) So, is right to assume that is better to have offset parts mapped in the skin without the offset?

And that's all for now, thanks for the comments so far, and I hope this can help to have a better understanding about the examples posted by skinners and some of the issues (and the challenges) encountered by them and by anyone approaching this task of making UVW mapping.