Beast is based off of a very powerful duplication script created by Johan Badenhorst (ezual). The script was initially designed to generate cities but thanks to ezual, it will soon power the generation of fur as well. The idea behind the concept is simple. Take a 4 vertice plane and apply an alpha mapped image of a few hairs. This plane is then duplicated (up until now manually) and applied to a mesh. This process is also known as the “card system”.
How much does it cost?
If you were using some fangdangled commercial 3d application, it would probably cost you quite a bit, especiallly considering the fact that this technology has already proven itself at the box office. However, you are a part of an Open Source community and that means competitive tools at no cost! All we ask is that you contribute to the community in some way shape or form. Blend on baby!
How does beast work?
Beast will be offered in the form of a python script to be used inside the Open Source 3D Animation/ Game Creation suite titled “Blender” exclusively. A detialed description and tutorials on using beast will be available soon.
What exactly will beast do?
As I said before, up until now the “cards” were created one by one and carefully placed onto the surface of a mesh in a way so that the end result would look as natural as possible. Beast will allow the user to select from a list of prefabricated patches of fur applicable to his or her project or even create their own. Once beast knows WHAT to grow all it then needs is to know WHERE to grow it! In the case of hair and fur, beast simply duplicates cards along the verts of a mesh. In the case of a city, beast works in the same manner discussed with the fur only with building duplication instead of cards.
When will beast be available for download?
Dont rush the man! No seriously, ezual is busting his keester on more than one thing right now (as with all of us and our toooooo many projects) so we need to give him some space. As soon as there is an update I will le you all know so keep checking in.
Can anyone contribute to this effort?
Absolutley! Anyone who has anything to contribute is welcome. Input for features would be nice but ultimately ideas on how to achieve basic features (length, density, color) should be our focus.
This post is subject to change like you wouldnt believe.
More details coming soon!
Thanks Landis, I’m glad people are working on hair alternatives, because I can’t get the dang fibre script to work (grumble). I may have a couple of things to contribute at some point. Thanks for your efforts.
<edit> You should try controlling those puppies with my PET method. Somewhere buried in General.
Yep, you have it right. Just select the faces, specify a couple of parameters, and it does the rest. Good for fur, scales, grass, weeds, rocks, buildings, etc. They can sit straight up and down, or set according to the face normal. setting parameters for randomization of size and attitude gives a natural looking result. Build forests, cities, or put hair on your bear. Just depends on what you want to use it for.
This is one of the lowest vert count methods available.This thread describes the card method. There is only 1 quad for every few hairs, so the poly count comes out much lower than other methods. I don’t know how efficiently the script will work, but when doing this by hand, as few as 200 quads can produce a good effect (more is better). This method is also relatively easy to animate.
to everyone involved, keep up the good work, I can’t wait to use the script.
Landis, you’re one of the true masters of blender. Keep the ideas coming!
There are a couple way to handle this, and they’re still being sorted out. For armature control of fur fully covering a creature, the process is pretty straight-forward. You can join all the cards to a single mesh and create your vertex groups (manually or weight-painting… autoskinning is still slow). Once you create your vertex groups and parent it to the armature, you have a choice. You can:
Unjoin the mesh back to the individual cards. Don’t worry, Blender is cool and smart enough to remember the weight groups in the cards. Unjoining the mesh allows you to have an alternative method of individually animating the cards (animating rotation or having groups of cards track to an empty). The unfortunate thing, though, is that it’s ungodly slow. On a high-detail model, you might have as many as 1200 individual cards. Having each card parented to the armature, requiring individual calculation per card gets processor-intensive.
Leave the mesh as-is. This limits (though it doesn’t eradicate) your ability to animate the individual cards, but it’s heaps faster to animate, and the results aren’t all that bad, and you can tweak for close-ups. This is the method I’m currently using in my project.
For short hair that’s supposed to just cover the head or other body part, it’s pretty easy to conceive of a workable armature system. Although, the way the cards are set up, though you could easily just animate their rotation or have them track to an empty or two. Another solution (haven’t tried it yet) would be Modron’s PET method, although I think you’d have to animate that with RVKs.
This topic is pretty much it (along with the City Block Generator and Hair Tutorial threads). We also have a project on the blender.org site. We’re still organizing that, but that’s also a place to check. Don’t worry, with the progress that’s being made, we’ll make sure NOBODY misses the release.
Echoing what quailman said, and adding this: for models that are covered in fur, there are many cases when you can practically obliterate the majority of the base mesh underneath, because it will never be seen. In some cases, this could actually reduce your poly count. I did that with my monkey model; deleted the entire base mesh except for the face, hands, and feet. I came out about even as far as poly count goes, and I actually subdivided my cards to give them a little bit more flexibility.
RVK’s will be supported in this, all it is is the proper use of a textured plane, teh onyl thing you mgiht have to worry about would be changign the orientation of the plane so that its always facign the camera for the best possible view
If you’re looking to use RVKs, then I’d definitely recommend joining the cards to a single mesh (remember the texturing issue). From there you can use RVKs just like you would on any other mesh. To be honest, I’m having trouble figuring out how RVKs wouldn’t be supported.
Fweebs right, RVK’s should be fine. Now, I posted this elsewhere in the forum but not in the actual beast thread itself, hehe!
Ok, here goes…
Beast Team Progress Report
There is now a density setting that determines the number of objects to be duplicated onto a particular face. When the density value is set to more than one, object placement on face is random. When it is equal to
one, object placement on face is centerred see image below of Suzanne:
*Note: Notice the variation in length where some faces were painted lighter at the top of the head. Not very realistic for an ape, however, was easily done. We hope this new density setting provides the user with greater control in that the mesh no longer requires subdivision for volume adjustment! Simply increase the density setting for a thicker head of hair without the need for unwanted vertices. Any thoughts on max density is welcome.
Default setting in the colour tab was changed to ‘None’ because of the
(temporary?) shortcoming of the python API concerning the second colour in Material.
Intensive testing and various bug fixes.
To do list…
At the moment the random rotation of an object is around the global axis and not the local axis of the object. Ezual thinks he will probably have to learn some more matrix math to sort this out.
Work out the bugs in the progress bar…yes there is a progress bar !
Ok here’s the scoop. The Beast forum over at the Blender projects site is now active. It has been for a while now but we are actually utilizing it now so feel free to waltz on over there for a visit. Threads have been created for discussion on particular topics (GUI, Hair and Fur, Feathers, Buildings, etc.). We have alos posted some requests for the community in the help section of our forum as well so if there are some of you out there that have been itching to help out while waiting to beta test than this is the place for you. Thats right, I said beta test! We are working to get a stable release done soon so that beta testing can begin, so the faster the other tasks get done, the faster the beta release will be unleasehed into the wild! Ok enough babling, here’s the link:
beast hair is made from mesh cards with an alpha channel, so they are animatable though it might take some doing… check the other beast thread for tutorial links. i think there is one that describes animating grass.