Replacing secondary skill icons

Outsane, first a big thanks for actually having made something, I think it’s a decent start and find it comforting to know there’s a modeler out there with an interest in creating new assets.
Having said that, if you like, I can offer some assistance with the last step of your assets. If you’re willing to share the 3d model (the softimage scene), I can take it from there and improve textures, normal and/or displacement maps, spec maps, adjust the lighting, add final gathering and rerender it with mental ray. That is, if you find the actual modeling of the assets the most fun part. If you like the texturing, lighting and rendering just as much or more then you may not want any help from me. :wink: Anyway, just thought I’d throw it out there, since it would be a good way of keeping this ball rolling.

As for assets, do you know that you can use online subversion services for ascii scene files? Check this comparison site for all the alternatives: svnhostingcomparison.com/
That way, you can post your softimage scenes with the models and every time you make a revision, it’ll commit to the online repository by just sending the actual changes to the file. Also, all previous versions are accessible, in case you wish to go back a few revisions. SVN is great.

Cheers!

Well haven’t decided yet, If you are willing to work on archery then sure I won’t touch it.
Other one you could try is estates, not sure how easy it is in AutoCad but from what I see its modeling a bag, a coin and then copy pasting the coin around 50 times :P.

Sorry I don’t know how easy will be to swap models from one program to another either, never worked in autocad only know of it.

Yes plz. Not that much that I hate lightning and the other stuff but don’t really understand em yet in a way to get what I want. If its an easy process for you go ahead.

nope never heard of this wonder full thing Oo. Don’t understand it yet (maybe because I just glanced it:p) but I will upload the scenes when I get it set up.

thank you guys one thing I know is that I can’t do this alone.

And a couple of thoughts on this whole deal.
Main thing that’s an issue is resolution. NWC did it right the first time with the limitations they had. This little project has the same limitations.
On the other side I know devs are playing around with higher resolutions(eyeballs them :P).
Even tho I don’t like to distract them from the actual project it would be nice when all Icons are done to have them in higher resolutions. It would also show one of the strenghts of VCMI over H3 modding. It would be nice to see heroes in a real high res (at least part of it…) instead of stretched up low res images.

Another thing on the actual work of the Icons. Because you need every icon in at least 3 different sizes and resizing small images damages them too much we need a new background texture (a big one). So all the models can be rendered high res on a monocolored background and then slapped over the background.

I suggest rendering them to png or xcf (gimp format), or another loseless compression format supporting transparency (or even alpha channels), and have background already fully transparent. I also suggest xcf or other format supporting layers for combining any 2D images together that is including renders of 3D models.
Note that gimp is free, open source, and availible for windows, linux and other platforms.
Also any other graphics including menus and fancy letters/captions can be done this way.
While talking about letters we can make some bitmap multicolor fonts specially to be used in vcmi (this includes all ascii printable chars and we can make them with pallette to swap or replacing collors internally in vcmi if 256 is too low for pallette).

I also don’t understand claiming skull have to be perfect. He/she died and perhaps been murdered, so cracking or missing teeth are welcome until you find it make them less pretty.

Got them, but don’t use 4shared next time - it requires registration :frowning:

We’re using SVN for code - perhaps it can be used for assets as well. You’d better ask for Tow opinion regarding this though.

In short - this is shared directory with version history. So users and artists will always have latest version of content without waiting for releases or collecting links on forums. It have some more interesting features but they’re mostly used by developers.

This is one of ideas for this project. Don’t expect this feature in VCMI soon but it would be great to have it once main game is finished.

No for xcf. This format can be only used with gimp, have no available tools\libraries, not supported by (most of) other image editors. Png’s are fine, bmp’s are always huge but may have alpha as well and there is targa which is quite popular among artists.

Not sure if this is a good idea. Unless you know a way to create font that will work with all languages. Finding existing non-copyrighted font should be much easier IMO.

For open source fonts, have a look here:
openfontlibrary.org/

I tried setting up SVN project hosting but failed. All the hosting sites I tried are made for programing but not for CG. I am just not techie enough.
Meanwhile here is the whole project
4shared.com/rar/G4r0enX8/H3_online.html
you need softimage to open it and in case any1 is interested there is a free version of it you can download from autodesk 440mb under the agreement you dont use it for profesional work usa.autodesk.com/adsk/servlet/pc … eID=123112

Its on 4shared again… sorry Ivan just mention me a good site I can google and I will use that from now on.
About .PNG I can upload em like that but whats wrong with .BMP (I mean h3 uses bmp wouldn’t that make things a little less complicated)

well that great and anyway its not like the icons will be completed anytime soon as well :slight_smile:

Mirrorcreator. It uploads to a few another mirrors at once.
I appreciate it a much despite using it shortly (warning it has 250MB limit per file so using something as big as CD require multipart archive)

I’m using dropbox - this is not filesharing service per se but it allows direct linking.

Regarding format - bmp, png and targa are all lossless formats, all of them support transparency\alpha channel so they can be used in VCMI as it or converted into each other without any issues.
Problems with BMP are: 1) alpha channel - some editors completely ignore it (for example MS Paint) so saved image won’t have alpha anymore. 2) No compression - bmps are HUGE.
For example your “ExpertScholarshipHighRes.bmp”:
16.8 MB in bmp format
6.4 MB in targa (2.5 times smaller)
3.3 MB in png (5 times smaller)
This may be small difference right now but when you’ll get 1 000 of such images size will become a notable issue :slight_smile:

Does that mean that vcmi will work with .PNG or you just want to make the transfer easier (sorry for a million questions)
BTW I kinda revisited the scholarship (again) starting to use HDRI lightning and things are looking good. After that I will try to make a decent replacement for the background and after that I pick balistic’s as next Icon.

http://i1152.photobucket.com/albums/p495/dimovgoce/ExpertHDRI82x93.jpg http://i1152.photobucket.com/albums/p495/dimovgoce/ExpertScholarshipIcon82x93.jpg http://i1152.photobucket.com/albums/p495/dimovgoce/skl8219c.jpg

Great work! BTW I am currently working on archery a should “finish” it sometimes soon (just have to do lights and textures).

Will you share the byckground texture when finished? It would save some time… and be more consistent.

wow that was fast I am working pretty slow and seldom.

of course. Just its not gonna be done today. And btw if autocad can have HDRI generated lightning I definitely recomend you try it. I know its a software oriented for engineers and architects but if it has support for that and final gather it makes anything look good.
Here is a link where you can find images like that hdrmill.com/Freebies.htm

Thanks, that would be great!
I hoped to finish it today, however doing textures takes quite a long time, so when finished, I will post it here.

I just googled for HDRI, however, I think that AutoCad does not allow this, but I am not sure -> if there is anyone experienced with this, it would be great. As far as I now, there is only metal ray.

I’m sure you all know this already but remember that exchanging 3d models between applications is very easy, by exporting them as .obj. I’m presuming AutoCAD has support for exporting as .obj, if not you can surely get a plugin for that. All 3d applications I know of (Maya, Softimage, Max, Lightwave, Houdini and probably also Blender) support .obj. It preserves all info including uv’s and texture info (at least I believe it does).

When it comes to 3d file interchange, rigging is the really complicated part. Animation can be difficult if it’s anything other than pure FK (forward kinematics) rotation (like what motion capture data is comprised of). But pure rigid 3d models have always been the easiest part of 3d to share between applications. Hope some of this was helpful. :slight_smile:

VCMI will work with any of them. So if bmp\tga are more suitable for you - use them. But for distribution\releases I’d rather use png.

And regarding background: I think there is no need to use it at all - images with transparent background should work as well.

So true did the first model for like 3 hours ( ~2hours spent on skull) and said to myself I am done. After that I got feedback and spent 10-12 hours struggling back and forth between photoshop and softimage.

All right PNG it is :).
And just to get the ‘official’ format for uploading images:

  1. Large image .png image with a transparent background for resizing
  2. a 44x44 Icon with transparency .png
  3. a 58x64 Icon with transparency .png
  4. 82x93 Icon with transparency .png
  5. extras like models, scenes and any bonus images
    Correct ? You can change the format for the images from there on

There is a tiny problem with that, and its shadows. The first one I uploaded with background didn’t have any so they kinda looked like they floated, after that I added shadows which grounded the icons. But on the other side I tried making a replacement background texture and managed to make two which were close but not close enough in my opinion. Anyway I’d rather just make the renders with transparency right now and deal with that further on plus I could add the shadow faster if I do them all at once (plus it would look more consistent)

If anyone is interested, I tried using sourceforge.net as a repository for scene files and it was not a problem at all. I registered a test project called VCMI_newIcons, uploaded a sample .obj scene (I just downloaded one online) with its .mtl (material info) file as well. Here’s the link to the files in this project: sourceforge.net/projects/vcminewicons/files/

Obviously, this is just one way of sharing the files we’ll work on, but since I believe .obj is an easy file format since it’s interchangeable AND in ascii, I just presume using subversioning for it would be ideal. Why? It keeps track of all previous versions (they’re all available) and there will be only one file per scene, as opposed to one for each version. In addition, since it supports SVN, Mercurial, Git etc., setting it up locally is easy. I believe you Ivan and Tow would agree with me on this?

So a long story short, using subversioning doesn’t require actual code, only a text-based (ascii) file format. Yay. :slight_smile:

Nice but I won’t be home today so I will check it out tomorrow I guess it works with softimage scenes as well. Hopefully I will understand how to get this thing working without problems :slight_smile:

Softimage scenes work for me, as long as you make sure to save them as ascii, not binary format. I have Maya, Softimage and Max at home, which are all Autodesk software. They pretty much work (or are supposed to) together well.

Looks to be it.

What’s wrong with creating shadows without any background? Background-independent icons should be more easy to use. Replacement texture may be useful for GUI replacement later on so it won’t be wasted.

All you need is to install TortoiseSVN (in case of Windows) and copy-paste URL. I think that can be qualified as “easy” :slight_smile:

SVN works with binaries as well (for example images). Only difference is that you won’t have any diffs for them (list of changed lines) which aren’t used by artists anyway AFAIK.

Having ability to use a Diff is always good.
I suggest using ASCII scene files, because then SVN would work better, and since SVN handler have to upload only changes and not whole file, it should be faster too. Also if it would happen two artist would do changes on same file, there exist a chance that the changes could be automatically merged.

On another side binary files may have lesser size themselves. Still I think using SVN without versioning stuff is something like a waste.