New subject, new blog.

So the whole blogging thing was such a painless process last simester with real-time 3D graphcs, that (in part due to Paul's suggestion) I thougth I'd do it again for Virtual Environments. And so this blog will pick up where KIB325 left off.

Click to view Blog for Virtual Environments.

Final

Shot 1

Week 14

So I've Started developing some of the environments within unity. Theres still a not so steep learning curve I feel like I'm starting to get over with this program. At first I was kind of skeptical of unity, (the terrain builder especially) but now that I've had my hands dirty a little I have to admint I'm liking it. The one thing that changed my mind re the terrain was the plant systems. The bamboo I'd developed with only a few tweaks could be placed within unity and bending animations to simulate wind where automatically applied. For a Bamboo forrest this as you can imagine was hugely useful, as the tall slender trunks would realistically display wind effects far more obviously than a oak or cedar or the like. These wind and bend settings took a little while to tweak, and if you look at version 1 you'll see the smaller bamboo shoots behave as though in a storm and the taller trees behaving as I intended.

Previous Versions: 1 2

Feel free to click on the unity windw then navigate toward the building around the scene, keep in mind as this in Machinima scene and not a game if you move past the enclosing hills and buildings the colliders have not been created and no attention has been put into the environment.

Use W A S & D to move Forward, Strafe Left, Strafe Right and Backpedal respectively, and you mouse to change the camera in the fasion of a First Person Shooter.

Week 13

So Perhaps a little late, considering how much time I've put into this project but I've started developing the environment for the story. The opening shot I thought to have focused on a mantis in a cherry blossom. Heres the crerry blossom.



Not all of these where done within this week, many where done before it but it made sense to keep them all togeatehr. Heres the model of the temple that I'm going to use as the backdrop for the most important scene.






I've used different combinations of leaves and trunks in different configurations of sises to create the bamboo forest and there are many variations of these models. The models actually finished larger that I would have liked, and I wish I'd have looked at the unity standard asset trees before beginning modeling my ones. I could have kept the polygon count far lower than it is by having a single pollygon with a partially transparent texture applied to create the boundry of the leaf, instead now I have have about 4 pollygons per leaf, and twice that becasue I needed to duplicate faces and invert normals to get unity to show both sides of the leaves. I tried correcting this with a custom shader, but that meant I was unable to use the built in nature shader, and the trees didnt color in the way I wanted. I've intentionally left the leaves one sided to illustrate the problem I had to overcome.


Heres the bird that will fight and be beaten by the mantis in the short scene where the protagonist gets the idea to learn from the mantis. I've been having trouble getting double sided normals working for this model, have a look from beneath the model.

Week 12

Bump Mapping Issues
After getting past my issues with normals between zBrush and Maya, I'm sad to report that unity doesnt do a whole lot with them. I'm sure I'm doing nothing wrong but unity just isnt developed to make good use of the maps I've produced. None the less I'll be including all the normal maps in my final comilation, and you can see the effects of light on the model demos below.

I havent really reported anything about how my texturing has been going, or even on most of the smaller models I've produced. I'll post the plain textures and unity previews so that you can see them in action.

Chow

Texture


So heres a little walk and run cycle I did mostly to test the skin weights, I cant think where I'll be needing it in the actual animation, but it seemed to be on the list of things to do. There are enough unity previews on this page already, so if you'd like to view it on a new page.

So theres a bit of a story behind the hang-ups I had developing this model. It took about an hour to model from the base mesh I developed for the first character you see above (including high quality mesh painting). The maps however I have had a huge amount of trouble with, I wont go into detail but somewhere along the way while I was trying to fix it, zbrush lost symmetry within the right nipple of the model, and that caused a huge ammount of trouble, with me needing to remirror the mesh and losing UVs in the process. I left this model behind in week 12, and am only posting this final in week 14 where I've added cloth and painted the skin weights.

The above animation should serve
the purpose of illustrating the skin weights.

Thought its hard to see due to the animation this is the second character within the story, and is a completely different model. Here are Bump Maps and texture if your interested. The aforementioned issues I came across with the zbrush corrupting the mesh caused problems with the textureing of this model. To get around these I tried using the new 3D features in photoshop CS4. Here are the resutlts. Needless to say I persisted with zProject.

Week 11

So I'm having all kinds of trouble putting the bump maps into Unity, I'll go to the pass session on friday and see if I can resolve it with some help. I'll post it online once I've had a decent chance to see whats going on.

So as I'm sure your aware, before a model have bump maps applied it requires a UV map, and the model must be neatly unwrapped. This model wont have any clothing on the upper half of his body, at least not this simester, so I've pelted the model. This screenshot shows both the unwrapped UVs and the bump map used in the screenshot below.

Bump map + UVs

Click to enlarge

applied bump map.

Bump map in unity (Just a liiiiitle bit scary)

Until then here are screenshots from Maya, where its actually working haha.

The issues zBrush was having generatin maps with I got around by using xNormals, a great freeware package.

Week 10

When modeling the head i had trouble keeping the pollycount as low as the body, especially when i got to the end of the model, at the neck, where it was to join the torso. The mesh detail I needed to get all the features and definition I wanted while maintaining quadratic topology just wasnt compatible with the torso (woah, thats a mouthfull). However, I've been toying with the idea of rigging the face eventually, so this difference in mesh resolution may have been nessecairy in the end any way. Now that I was in zBrush however, when subdeviting the mesh, the higher mesh density of the head mesh was multiplied, and the head's pollycount alone was the same as the entire body. I brough the model back into maya and tired mesh reduction, and got the mesh to a similar density, but even then, the joining of the two models was creating really ugly polys even after I'd reduced the density of head, and triying to clean up the model manually.


So this, along with the need for better edge flow and the oppertunity to reduce the polly count and add resymmetry lead me to retopologise the model. So the model was taken into maya and the undeveloped half deleted, geometry mirrored, the head's mesh reduced as described above, and taken back into zbrush.

I'm really happy with how it came out. The new retopologised mesh has only a fraction of the faces of the old, and retians all its detail. There was one significant problem I had though, and I think -as I often do- I became fixated on it. I was unable to import the new retopologised model as the subdivisional level of the high detailed mesh. I spend days looking online and trying to fix this, and in the end just subdevided and repainted the muschles onto the new mesh. If I had done this in the first place I could have saved myself ALOT of wasted time.

Heres the retopologised mesh.

So with another train trip ahead of me I started modeling the atagonist, and with the new retopologised mesh I was able to finish the model's body mesh in a single trip.

Click to enlarge.

According to the story, Monk Set a much was a man of huge stature. He thought himself the strongest of all the monks and had a short tempter and fought often. He was an unpleasant man, and a bully. I've tried to reflect this in his model. Details like a broken nose and scarred forehead, I also gave him a strong jaw like and generally masculine features..

Week 9

So I have to admit, I havent done alot of mesh painting before, so i couldnt start on the high res models without picking a software package. I've been wanting to do this for a while, ever since I saw speed mesh painting on youtube. The dicision was between zBrush and Mudbox. Paul reccomended mudbox because as it was made by autodesk, it was intuitively much the same as Maya. I could see his point, but it was probably for this reason that I went with zBrush. Its a good thing to push your limmits and try something different now and then, and oh god . . . . push limmits is exactly what zbrush did. Its nothing like any 3D package I've ever used. Instead of a virtual space, the program behaves like a canvas, and instead of having items within that virtual space, you get "Tools" with which to paint onto that canvas. I still cant see the point in this, even after having learnet the program. It serves to do little more than alienate users if you ask me.

So the first thing i did was take the model I showed in week 7, It was modeled off Da Vinci's Vitruvian Man, so I was confident that its proportions where anotomically correct. However this meant that (for the second time, grrrr) I disconnected it from the rig and lost the weight painting and positions. In retrospect, I could have saved myself alot of trouble by going straight to the mesh painting without having rigged the model first. By not doing this I'd taken the mesh into zbrush after having experimented with its animation, and resetting the neutral pose. This wouldnt be a problem, only the mesh I took into zbrush was symmetrical in topology, but the vertex positions where not. I dealt with this problem in week 10, and will post more details there.

So now I feel is the time to praise zBrush. For all the toruble I had getting over the learning curve involved with zbrush, once passed, I cant even describe. Modeling became fun again. I somehow assumed zbrsuh would be like the brush tool in maya, I was so wrong! Working with zBrush really is like sculpting. What I was complaining about earlier, with the head taking so much time, I could now do with ease and speed. My bamboo is suddenly so much more usefull. I feel like this is how I want to model from now on.

At the time the net was down, so I took a few photos and was able to start modeling. You'll notice that -as I mentioned before- the model had at this point lost symmetry especially at points furthers from the centre such as the arms and shoulder, so I didnt bother continuing to model with symmetry, and instead worked only on his right side.

So his body now had far more detail and muschle definition, and I'm happy with what zbrush has added to the model, and am confident that the bump maps should look good. Though theres still alot of work to be done on the model.

Week 8

So I dont actually have anything to report for this week, as I haven't made any progress to speak of becides the decision to create bump maps and model with high detail with my characters. KIB220 has literally taken ALL my time, and I've done nothing but sit in my room and edit all week, wich is fair because i volenteered to edit in place of drawing. Next week I mean to begin work on high resolution models, and will post back with my progress.

Week 7

Completing the head has taken a rediculous amount of time, more so than the body and face combined. Maya started bugging and wouldnt save for a long time so I found in the end I'd done the actual head about three times before I was able to save it in a state I was comfortable with. I've resolved maya's saving problem so my work shouldnt be hindered by that any more.

The nCloth ismulation on the pants and belt have begun to really annoy me and I've hardly started animating it. Even something that should be so simple as exporting into unity from maya became difficult, as the cloth wont export and then nicely import into unity. I'll give it a little more of a chance because I really like the way that the cloth looks when animated, but I may ditch it all togeather and combine the meshes for hte belt pants and body if it becomes much more of an irritation.

So this still needs alot of work I think. Its currently rigged with FBIK but I still mean to better texture the body and increase its quality with bump maps, rig the face and at least the mouth so that I can later include expression or at least a little mouth movement. For this reason I've kept the poly count on the face and head pretty high. It should be the focus of alot of important shots in the final production so I think I can sacrifice a little more processing to cater to this. I got a bit ahead of myself when did the earlier animations, and although I knew it would happen I had to detach the mesh form the skeleton and didnt bother with plugins to save the skin weights so I'll have to create them again no that I've attached the head. I'll do this next week, after I've messed around with bump maps and decided exactly how much detail I'll include with the mesh.

Heres a link to the Maya Binary with the skeleton included.

Week 6

Well, I've started work on that face, and I'm not sure how I like it. I know one thing for sure, my polly coult went way too high. I've tried to reduce the pollys and had to introduce some tesselation, which I manage to keep out of the model until now, but I supose everything gets tesselated once imported to unity any way so it doenst really matter. None the less I feel like I lose alot of detial when I reduce the model and might just scale up the body instead of scaling down the head too much.

Original


Reduced pollys

Week 5

Well, I now have my first character more or less done. All the mantis still needs is a proper skin and I'd perhaps like to put it through a few more generic animations such as walk cycles just to test the skin weights etc.

This week I started work on my second, the protagonist, and its coming along well. I've managed to keep the polys at 3000 faces on its core mesh, even after the mesh smoothe, and the body's phesque has retained all of its most important forms and bulges for muschle groups.

I added a skeleton to the model today, and gave it a short test with smoothe bind witout setting up FBIK yet, just to see how the skin weights assigned. Luckily the weights shouldnt need to much tweaking once I want to finally bind the model.

In retrosoect, I probably should have done this through Unity, but I've made the gif now so here it is. This is the result of a quick little animation after the model was rigged with FBIK. I'm still not happy with the way the legs look, partially the groin, though I'm sure this is because the skeleton was attached while the legs where still each outstretched at 45 degrees. I think it can be fixed by adjusting the skin weights, but the finished mode will be wearing pants so I dont know if I'll take the time to fix something that may be deleted or might never be seen.

So I've added the aforementioned cloth to the model. At the moment the mesh was modeled manually and its behaviour controlled by nCloth. Maya 2010 seems to have dropped the old cloths system so I've had to relearn ncloth, but I'm quite liking it so far, with the exception of the cache system, but I suppose its needed or they wouldnt have included it.

All thats really left to do for this model is its face and of course the textures. I'm not sure what I should make this model look like. I'm in two worlds as to weather it should have hair or not. I know how it can really increase the size of a project to include, and the character could be a monk or something of the like, and they wouldnt have hair at all.

Project Proposal.

After reading the assignment outline for this subject i immediately knew what i wanted to make. There is an old Chinese story that I've always liked, it goes as follows:

Synopsis.
After being badly beaten by his giant rival, a man creates a method with which small can overcome big, by observing a strange insect he finds in the forest and defeats his larger, stronger opponent.

Sound Information
This project should produce a good finished product without the need for particularly complex sounds as much of the narrative will show the Protagonist learning from the insect. I imagine most of the required sound will fall in the little bit dialogue and perhaps the creation of ambience. If there were to be music, I'd imagine it would need to be somewhat esoteric, to stay in line with the style of the rest of the animation. I'm very open to suggestion and if you're a sound student reading this feel free to contact me with ideas/questions/suggestions.

This project will only require one, maybe two environments and three characters (Antagonist, Protagonist, Mantis). The majority of the work will be within the animation rather than the creation of other assets. I've spoken with Paul during tutorials and he seems to be ok with the idea, at this point it just needs more development before its clear exactly what I'll need to do to make it happen, but this will happen once I start developing shortlists, storyboards etc for this project. I'll post them when they are done.

Week 4

- Tue 9/3 -

The lecture slides for week three haven't appeared online yet, and again I couldn't get work off to attend. I've handed in my resignation now thought so this should be the last one I miss. I went to the pass session again but I figure its still not running. During the tutorial I heard for the first time about this blog and it was part of assessment and started posting reflections from weeks 1 and 2. I'd read over the assessment already and had have a pretty good idea of what I want to do for my project. I started modelling on the train on the way home and have the mesh and a rough texture for the first of my characters completed. Here's a few quick renders and screenshots:

I have to start being careful with the way i model though, I've done a bit of reading into unity's engine and a few times its been stressed how important it is to keep the polly count down. Before the mesh smooth my model was looking pretty good in that respect, and I've kept copies of the low polly model and figure if the polly count blows out I can always go back to those models. The head is probably the worst part of the model, as it has both internal and external geometry, and before the model had been smoothed it was pretty big. I figure this is ok though, as there will be a fair bit of focus on this mart of the model for a fair bit of the animatic, at least that's what I have in mind at this point. The textures need a whole lot of work, especially if I'm going to include use bump maps. At this point its pretty much just a planar projection of my reference pictures haha.

- Fri 12/3 -

So Another long train trip to uni meant I had time to work on the last model agian, and I managed to get it rigged and the weights painted. I opted to go with IK instead of FBIK as its sometimes simpler with a model like this thats neither a Biped or Quadraped. Its looking pretty good and behaving well, I ran it through a short animation when I got home and liked the results. Here are some pictues I used as reference for movement.

Week 3

Again I wasnt able to attend the lecture due to work. This week the lecture notes weren't as useful to me as those of the week before. That said I'll undoubtedly refer back to them, but they didn't contain a whole lot that I hadn't already heard or learnt modelling for mods etc. I hope that there wasn't too much more discussed during the lecture that I might have missed but hope I'll make it back up in tutorials. I attended the pass session on friday from 1 -> 3 to try and see if i could get any more information from monday's lecture but it seems I was the only one who turned up, and I guess they'll start running as soon as people need more help.

During the tutorial we where introduced to unity, and I quickly threw together a low polly model in Maya and exported it to unity. I've not used unity before, but so far it looks like a great package, I can't wait to work with it. Moving files between software packages is a breeze, and I was surprised as to how many different mediums you could quickly and easily build to and run. The real-time feedback the game window gives was another thing that really blew me away. I've worked mostly with the valve's HL and Source engines, so its really refreshing to be able to drag and drop files form Maya to unity without needing to reinitialise the whole scene.



So this is a web build from the work I did in the first tutorial in week two. You'll need the unity web player installed to view this however. I'm really impressed with unity's ability to just build and run like this. I'll be using this in future to show my progression as I start producing more media for assignment three.