Jump to content


Photo

Artwork Submission Guidelines & Faq


  • Please log in to reply
4 replies to this topic

#1 Deimos

Deimos

    Colonel

  • Forum Members
  • PipPipPipPipPip
  • 1,591 posts

Posted 03 June 2003 - 11:04 PM

1. If a thread does not exist in the Images forum, create one called 'ItemName concept discussion' or something similar. Talk about the object you are interested in seeing. Give references you find to the design if it helps get your point across. Get feedback on points from other member comments, and try to incorporate other ideas if it improves the overall design..
2. Draw up concept sketches. It's way easier to redesign a pencil sketch or hand drawn concept than to endlessly tweak a model, but use whatever tool you are most comfortable with.
3. Discuss points of the sketch, fix and tweak it until everyone is generally happy. Minor items like one color versus another don't count, minor details like that can be handled later. If you make changes to a concept, always save the previous designs! You might find out the first concept was the best, you don't want to redo it later.
4. The concept might sit for a while at this point, especially if other concepts are in the works. We don't necessarily take the first concept we get, we like to see several ideas to compare them. Once a concept is liked by most of the team and is approved by the senior members as appropriate for the style guidelines (see below), it is placed on "the short list" of completed concepts.
5. Build a 3D model to the concept standard to see the design in 3d. The model must be to scale with this block, which is 1 meter square and 2.5 meters tall. (It is sideways if you don't use Milkshape 3D, which has the Z and Y axes switched compared to 3DMax) Make sure you model is properly aligned-sitting on the ground if it will be on the ground in the scene for example. You must use triangles for the mesh, not NURBS or 4-sided faces. Post the model in a .zip file that includes the exported model in 3ds format (and any textures you have mapped in .png format) for everyone to look at in their modeling packages. Also post a screen shot of the model in flat-shaded view, 50% grey background, with no special lighting as well as a textured view if you have a texture assigned. Try to post the screenshot as a .jpg and try to keep the picture size small for those who have 56k modems.
6. Tweak the model if necessary after more discussion.
7. Finalize the design, and submit for skinning and any animation if not already done.
8. Once the high poly, X-Net model is complete, create the low poly, battlescape model. The texture will be scaled down as necessary as well. Submit both as a second .zip file in the same thread as the concept. Make sure your model is aligned and scaled properly, it will not be done for you.
9. Currently our designated 3D engine is Ogre. There are export plug-ins for several 3D apps on their site, take a look for your version. Once you have installed the export plug-in, you should export the model to .mesh format. Here are some additional comments from Drewid regarding Ogre: The main restriction is to do with the number of materials per object. Ogre does not support per-face material assignments. The mapping from Maya to Ogre is like this:

Maya scene --> Ogre mesh
Maya mesh --> Ogre submesh

Each Ogre submesh can have one material, so therefore each Maya mesh can have one material. A material is different than a texture though -- each material can have up to 8 texture layers applied.

Remember that to render efficiently you want to minimize material changes. So making a model with different materials on every face, or lots of submeshes, will slow things down horribly. Better to have a larger texture with smaller sections, and get creative with the UV mapping.

Multiple objects can share a single texture, but keep the material name unique to each object. Best Practice is to do the naming as you create the materials.

If the objects share a single texture then try to make that obvious in the texture name. For instance all the baseview objects will share a single texture map, with a name like "allbasestructures.png".

Animation: at the moment all animation is done through bones. So if you want to rotate a whole object it has to have a vertex assignment to a single bone. It seems possible that it might work to have the object as a child of the bone instead, which would be a lot more efficient.

Only Max and Milkshape have animation exporters as of now. The Maya animation exporters are being worked on. There is a work-round if you want to animate in Maya, but it adds extra steps to the whole export process, so really should be avoided. Maya is easy to work in so I think the animation exporter will be running soon.


A note regarding polycount - the first model should be the high quality version that we'll see in the X-Net (Ufopedia equivalent). For the X-Net, keep models under 6000 polys and use a 512x512 texture map or less. More details can be found in the asset list.
For all other areas of the game polycount on character models should be 1500 or less, weapons should be 500 or less, and small environmental objects should be as small as possible without looking bad. Objects you won't see in the X-Net should be modeled as the low poly versions from the start.

A note regarding textures - High detail X-Net models should start with a 512x512 map (if modeling heads separately, use another map for the skin detail). Texture maps must be sized in pixels by a power of two: 64x64, 128x128, 256x256, 64x128, etc. This map can also be used for the lower detail battlescape model, but will certainly be scaled down to save video memory. You can scale the texture down prior to posting, just make it as small as you can without losing too much detail, and maintain the power of two sizes above. You can use 32x32 or 32x64 as well, and other variations up to 512x512. The texture should be named to match the model name, for example aliengrenade.3ds would have aliengrenade.png as the texture. The X-Net model should be named highaliengrenade.3ds. Don't put spaces or underscores into the name, and preferably make the name all lowercase. For generic models that we might have more than 1 of (for eample a street light), include a unique designation in the name, preferably your initials followed by a number. This will make sure all model names are unique between members. An example from John A Doe would be streetlightjad1.3ds and streetlightjad1.png for the texture.

When a model is posted as a .zip or .rar attachment, it is important to realize that the original artist is the one working on the model unless they pass it off to another member. No modifying of the model and reposting should be done, unless you can't explain a point of the concept any other way. The original artist is responsible for integrating ideas into their model, but if they want help with changes they can ask others of course. If a member leaves or is on extended absence and changes are needed, they can be done by another member with a senior member's permission.

We won't be using any art that is a derivative or modified version of someone else's work in Xenocide, which includes anything which doesn't have the open source license disclaimer on it or we have the creator's explicit permission to use it. It applies to all art, not just models.

We're not saying that just because a model makes it into the repository that it will be guaranteed to make it into the game, but the high quality models in there will be considered the 'short list'. Anything that doesn't make V1.0 that is still in there could be offered as an optional download pack. Also realise that submitting concepts and models for an item is not "first come-first served", you can post models for items even if someone has already posted one and both will be considered.

We can see that some people might well get upset by the fact their design didn't make it, which we are going to have to face. There's no way around it so we should be aware of it and accept that people might get upset. We do have a style guide and if we don't stick to it the game will look patched together very quickly. So if we just accept any ideas thrown at us we're going to end up with a mish-mash game. We don't want to see that happen.

Edited by Breunor, 09 January 2004 - 09:13 AM.


#2 Breunor

Breunor

    Creative Text Department

  • Xenocide Inactive
  • 3,234 posts

Posted 21 December 2003 - 11:29 AM

To help facilitate importing all the models into the game engine, we need to organize our files to match what will be used in the final system. For those with ftp access, you'll notice on the ftp site a data folder, and I've moved all the work posted there into what I think will be used for the directory layout. Moving textures will cause the 3ds file to lose the mapping, but that can be fixed. When you create a new model to put up there, make sure to follow the new system and put it all in the data folder subdirectories. Here are the instructions posted up there in the readme file:

These directories are for Xenocide-related material only. Non-project use is not allowed. This server is moderated on a daily basis, so no sneaky business.

You can link to these files in the forums, everything is case sensitive so please use lower case for the entire name. File names should not have spaces or underscores in them either, just run it all together. Make sure you test your links before posting.

For 3D models, the 3DS file goes into the /data/models folder under the appropriate subfolder. All textures need to be in PNG format and go into /data/textures, using the matching subfolder. The texture name should match the model name, for example, plasmarifle.3ds would have a matching plasmarifle.png. For common items like benches or tables, include the artist's initials and a number on the end of both files, like benchabc1.3ds/benchabc1.png.

Edited 01/09/04 by Breunor, and comments or corrections please PM me.

Edited by Breunor, 09 January 2004 - 09:17 AM.


#3 Breunor

Breunor

    Creative Text Department

  • Xenocide Inactive
  • 3,234 posts

Posted 25 December 2003 - 09:46 AM

Note: significant updates to the first post were made 12-25-03. Please read the first post again and note these changes, they are significant items that affect most of the work being done here. Thanks!

#4 Breunor

Breunor

    Creative Text Department

  • Xenocide Inactive
  • 3,234 posts

Posted 09 January 2004 - 09:27 AM

A few more tweaks the first posts, and here are some reminders:

Make sure your model is scaled and positioned correctly. Compare it to the 2.5m tall block model and make sure it is centered on the origin, sitting on the "ground" plane. In 3DMax this is the X-Y plane, in Milkshape it's the X-Z plane.

Make sure your model and matching texture is named correctly. Do not use any caps in the name, including the extensions (it should be 3ds, not 3DS). Make sure to include your initials and a number at the end of the name so we don't have duplicate names. Not only do we have to rename, we have to reassign the texture with the new name as well.

For the final model submission, include the exported .mesh (and .skeleton if the model has animation) file for the Ogre 3D engine. As long as the model is properly built, we will export it to .mesh anyway, so this will save us that step (and earn you brownie points for saving us time!).

#5 Breunor

Breunor

    Creative Text Department

  • Xenocide Inactive
  • 3,234 posts

Posted 10 January 2004 - 11:03 AM

One more point regarding the scale of items relative to the block. For X-Net models, the model must fit inside this block. For battlescape models and items you don't see in the X-Net, the block is a scale reference to a 1m-1m-2.5m volume.

The difference for the X-Net version is so that when the X-Net loads a model, the camera zoom can be at a set distance and focal length, and every model regardless of "real" size can be seen without adjustments. Otherwise you could switch between the avenger and a pistol, and you'd either see a tiny pistol that requires zooming in, or you'd see a tiny patch of the avenger hull, if you weren't actually zoomed inside. Note that because the block is taller than wide, there will typically be extra space above the model. Please place the model "in the middle" of the block. This is likely where the original will be for the camera, allowing better rotation of the model.