It's my first post of 2020! I'm going to take a look at something that's been bugging me, which is the practical applications of the cloth and garment modifiers. There are, of course, times when you will need to animate cloth. Cloth exists in real life; of course it's going to show up in video games and animations. However, it can be glitchy and has some tendency to lag badly when rendering something complicated.
For things such as a tablecloth or bedcover, cloth modifiers make sense, especially when the object is then converted to an editable poly or something similar, leaving only the end product of the simulation. This allows for realistic folds. For 3D animations, animating complex movements of cloth can be complicated. Even a simple act of pulling back a bed cover (with the end goal of a nonmoving, pulled-back cover) with a dummy object proved incredibly complicated to get right; the simulation tended to glitch a lot. While this might be worth it for, say, an animated movie or even a cutscene in a game, having a cloth simulating in realtime in a video game, especially one the player could move and interact with, would be a terrible idea in most cases. It would probably get caught on something and turn into an unpleasant mess of polygons sticking out at odd angles (I've seen it happen). Even a flag blowing in the wind is probably best as a fully rendered loop rather than allowing it to simulate continuously. I'm having more trouble finding use for the garment maker. So far I've used it to make a sack, which turned out somewhat mediocre, and I understand using it for props like this, including clothing hanging on a rack. For character clothes, I also understand using it to create the object, but as I said before, realtime animation of the cloth is a bad idea. This leads to a trend I notice in some video games of characters' clothing stretching awkwardly as they walk. This is especially noticeable with skirts, especially when they avoid unfortunate up-the-skirt camera angles by connecting the bottom edge of the skirt with a flat plane which the character's legs poke through; I've seen it with figurines as well. I'm not sure why, but I find this particular visual to be incredibly uncomfortable. I just flat out don't like seeing characters' legs sticking out through a flat plane, and I don't like how uncomfortably solid it makes the skirt look. I think a flat plane a little higher in the skirt, where it's less likely to be seen, could fix this weird visual. I don't have a solution for the awkward stretching of clothing as characters walk; I understand it's just easier to animate it that way. That's the price we pay for less computer strain.
0 Comments
I made a model of a fire hydrant today out of a bunch of primitives; I had considered lofting, but it likely would have taken too long. That's got me thinking. While attaching or booleaning a bunch of primitives might be quick, it can also look clunky. My original idea for making a fire hydrant had been to loft the main body, then loft the bits that stuck out to the sides and attach them. I ended up just using attached primitives, which worked fairly well considering that fire hydrants do have some angular features, but I feel it might have looked smoother as a loft. I did, however, use the extrude tool on some of the primitives to add detail (I remember a time when I would have made terrible use of the boolean tool for the same purpose). What are the benefits and drawbacks of these different techniques? Lofting creates smooth transitions between cross sections–mostly–and eliminates the seams that are common with using multiple primitives or even extruding. Transitioning from cross sections of a different shape, such as a circle and square, can get messy. In addition, because of the smoother curves, lofting generates a lot of polygons, which is generally something to be avoided to keep load times low. Lofting also takes a lot of time, as one needs to have the cross sections laid out beforehand, and it can take a lot of fiddling around to get things right. Overall, it looks nice but can be difficult. The extrude tool is useful for creating details quickly–I used this to add rings around parts of the fire hydrant, some of which were sunken in, others of which stuck out. It does tend to result in a lot of sharp edges, but has a lower poly count and is fairly quick. Sticking a bunch of primitives together is also quick, but tends to end up in a lot of polygons if they are simply attached rather than booleaned (due to the overlap). It doesn't require a lot of forethought like lofting, but is often angular and rough. So, to summarize:
The techniques I've been using to make simple house models in 3ds Max have given me a few thoughts. In order to save time, I've been using the extended primitives for house features such as doors and windows. I'm starting to wonder if I would have been better off making my own. In the Wild West Ravine project before this, I made my own simple four and six pane windows using little more than box primitives and boolean techniques, possibly in less time than it's taking me to fix these pre-made windows. There are a huge number of parameters, and editing is difficult because it often requires repeated small changes to each one for the desired result rather than inputing each value once. Making windows myself requires, surprisingly, less fiddling around. In addition, the pre-made windows are clunky even with reduced depth values, with the frame sticking out beyond the thin walls of the houses we're modeling, and unfortunately most of the detail is lost when the depth is reduced. This is not to mention that many of the options look the same, and certainly don't look like most windows I've seen in real life. They also consist of a huge and often unnecessary amount of polygons, something that's better off avoided. The pre-made doors have a similar problem of being clunky and having too many parameters, for a result that, in this case, could be easily recreated with a couple of box primitives and simple boolean. Overall, these extended primitives are decent at best and don't really suit the project. The Attach function, however, is incredibly useful because it allows for selecting each element separately, making texturing much easier and allowing for more detailed editing even after the objects are combined. This combined with some of the other tools can be incredibly powerful. For example, I duplicated the model I had for a house, deleted the windows, patched up the holes, shrunk it with the scale tool, and voila: two of the little structures that pop out from the roof (my lack of knowledge about architectural terminology is showing itself here). All things considered, this has been an interesting project so far. So, to summarize:
We've been working with a lot of compound objects recently, but which is the most useful–and for that matter, which is the most fun? The most useful compound object is, in my opinion, the ProBoolean. Cue the tomatoes! Yes, I know it's a simplistic answer, but I have my reasons. It's easier to use (and more stable) than regular Boolean, and has a bigger variety of functions than ProCutter, which is essentially a specialized boolean operation. It can produce a much wider variety of shapes than the relatively specialized Blobmesh, which isn't useful for any angular or completely smooth surfaces. Of everything we've used so far, I'd say the closest competition is Loft, seeing as it can allow for incredible convenience in things that might require a lot of manipulation of a lot of objects to create–I recently made a lofted screwdriver, for instance. However, Loft objects can be difficult to edit if you don't like your first result, and the transitions between different cross sections can be awkward. The fact that lofts follow a single spline limits the possible results. ProBoolean has the widest range of possibilities, and even using other compounds objects, a boolean will probably be necessary for a great many creations. A simple union can also greatly reduce the number of polygons, which is key for making games that load quickly, making ProBoolean vital for models which connecting objects. That said, I find Blobmesh to be the most fun to work with. There's just something amusing about goop; the slime craze of the earlier 2010s stands testament to that. As I've said before, the specialization of the Blobmesh makes it somewhat tricky to find uses for, but that's part of the fun. It's satisfying to know I've found a project where a relatively niche artistic tool is useful. So, to summarize:
We got back into 3D modeling pretty recently, and one of the key things that's stuck with me is how to work with splines. Where has this been all my life? Splines are basically 2D lines and shapes in a 3D program. As far as 3ds Max goes, they're incredibly easy to create–easier, ironically, than working with a pen tool in most 2D programs I've used–and can be converted into 3D objects using modifiers like lathe and extrude. This has given me so much more freedom to create unique shapes; I honestly really wish I had worked more with splines earlier. Lathe, especially, has been incredibly useful. It's essentially spinning a 2D shape and around an axis to make a 3D shape. This means I can make any kind of rounded object with ease. So far that's included bowls, candlesticks, plates, wineglasses, cups, and even a table. I've currently got a model of a flan (or a gelatin mold, depending on your point of view) in the works, and I'm working out in my head how I'm going to make a character I dreamed up when I was younger–the clothes and hat are essentially completely rounded, which will make it easy to draw a cross section and lathe it. A few modifications to the newly created 3D meshes should take care of whatever the splines can't. Unfortunately the normals on lathes tend to freak out, and I've ended up with huge dark patches on some of my models where the shading didn't work. The extrude modifier has also been helpful, though it does have its downsides. The sides of an extrusion are always flat, and as I haven't yet discovered how to get more vertices on the inside of a spline shape, I ended up with some awkward looking spoons and forks, as well as an unfortunately blocky coffee cup handle that looked like it would be uncomfortable to hold. The flat sides should be pretty easy to fix with a simple chamfer modifier. As for the internal spline vertices, I'll have to experiment and see if I can find a way. I'm still dissatisfied with that spoon; it looked like a bent spatula, because the only way I could get the middle to sink in was to lower points on the sides of the spoon as well, creating a tool that might help you dig through dirt–or maybe beans–but couldn't hold soup without spilling it. In addition to these two modifiers, I seem to remember something else I used a year or two ago, something similar to extrude but where you have to pick a spline shape that will determine the circumference of the extrusion. I've been digging through the modify panel, so far with no luck, but I'm sure I'll find it. So, to summarize:
I was going to try to get better at using Blender, but my family just got a new computer with an absolute monster of a graphics card, so finally I can run 3ds Max. I took the opportunity to finally create a 3D model I've been planning since last year but never had a chance to create: a skyscraper. This required heavy use of the skills I've learned in using the material editor, object manipulation, and modifiers, and tested my skills with lighting as well.
Then I reset the groups for the material editor: the windows, the doors, and everything else. I created a multi/sub-object material and used the brick pattern as a bitmap for the main building, then created my own from scratch using standard materials for the windows and doors. Afterwards I tried for quite some time to use lights to illuminate some of the windows, but I couldn't figure out how I'd used to do that and was unable to find any answers from google, so I ended up creating another group for the materials after reading about self-illuminating materials, hence the lit windows. The flag on top was easy: I stuck a sphere on top of a cylinder for the pole, used a boolean to combine them, and quickly created a material to look like shiny metal. The flag itself was the result of a very thin box primitive, another quick material, and some fiddling around with the wave modifier. So, to summarize:
Citations:
Khalid, Hassan A. “Free Stock Photo of Brick Texture, Bricks, Building.” Free Stock Photos, www.pexels.com/photo/brick-texture-bricks-building-cement-1200862/. We're finally wrapping up with 3D modeling. Now that it's all said and done, here are my thoughts on the matter. Besides the programming I did earlier in the year in the 2D game project, 3D modeling has probably been the easiest thing I've done all year. That's not to say it's actually been easy. The biggest challenge is that a lot of the work is meticulous, and small changes can have big effects that might not be obvious until later and are difficult to troubleshoot. During the column modeling project, for example, I noticed several of my maps had not turned out correctly. I still have no idea why, but they were definitely warped, and it was noticeable in some places on the model. I followed the tutorial meticulously, but for some reason mine didn't turn out correctly. It wasn't really that drastic, though, so I still count the column as a success. The meticulousness caused other challenges, though. A lot of projects ended up half-finished due to the amount of time they took being longer than the amount of time there was available to complete them. UVW unwrapping was also complicated, or at least some aspects of it were, and it often ended up messing up without a clear cause. I did enjoy figuring it out, though. It was almost like a puzzle. In general, that's what I like about 3D modeling: you want to make a certain object, and you have to figure out the necessary steps to make that object come into being, and it's a lot like solving a puzzle. As for the program ads Max, I have mixed feelings. I like the functionality, but it's very resource intensive and tends to have difficulty opening–or that might just be the subpar computers. Sometimes the menus are confusing to navigate, especially when icons get changed over updates. Also, a few useful features were changed since the time the tutorials were made for no apparent reason. One of these was the ability to lock viewports, which would have been helpful. So, in conclusion:
|
AuthorI'm moving on to my 4th (and final) year as a Game Art & Design student at Durham School of the Arts. I'd like to call myself an artist, but I'm a programmer at heart. Archives
February 2020
Categories
All
|