The title of this lesson was somewhat bizarre to me since I had never before heard of NURBS and didn’t know what to expect. We were, thankfully, immediately given a definition – and I was surprised even though I shouldn’t have been.
NURBS stands for Non-Uniform Rational B-Spline. They are one geometry type that you can use to create 3D curves and surfaces in Maya. The reason this surprised me is that I had only ever worked with polygons and heard of subdivision surfaces. This third type of geometry should have been expected though since more often than not, there are multiple ways, types or versions of something in Maya so that the user can choose what works best for them.
Like polygons, NURBS are made up of three components. These however are different: surface parches (like faces), isoparms (like edges) and control vertexes (like vertices). They are mainly used in Autocap, which is another type of 3D software, and in game production, more often than in Maya anyhow. They can however be converted into polygons, which can be a quick fix for certain problems and something we went over in the lesson.
We went ahead and compared the two in real life by opening up Maya and going into the ‘create’ section. Then, in ‘NURBS primitives’ we chose ‘sphere’. Just like it normally would, a sphere popped up on the grid. At first, it seemed exactly like the poly sphere, but when selected, the geometry clearly stated that there were some differences. Side by side, with both a poly sphere and a NURB sphere selected, the linework looks like this.
When right-clicking and holding on the poly sphere, the options that I am so familiar with popped up. I chose ‘vertices’. When I did the same to the other sphere, however, the options were different, since NURBS have different components. Since the one like vertices is control vertexes, I clicked ‘control vertex’. Again, side by side, you can see a clear difference in the way the two primitives are built. We were told that while polygons are standard NURBS can be used to build certain shapes faster, therefore increasing workflow.
After clearing the grid, we clicked on ‘curved/surfaces’ which is an option directly next to the poly shelf. We would be using ‘EP Curve’ for the continuation of the lesson, which we have worked with before and I am somewhat flexible with. I went into front view since it was needed to make it, and started clicking bit by bit to make the curved shape. In a similar way to last time, I created a sort of wave, clicking in small portions so that each bend was more precise.
This is called a spline and is like the profile of the shape that is to be created. Back in perspective mode, I duplicated it and moved the copy to the side, as per the instructions. I edited the curve of the second one so that it was slightly different to the first by moving each individual control vertex where it evened out its bend. This was optional, but I also scaled it down slightly. The next part was ‘where the magic happened’ – in other words we would be seeing how NURBS could actually be applied when building in Maya. I selected both lines, went to ‘surfaces’ at the top and clicked ‘loft’. This joined the splines by generating a surface between them and creating an organic shape. It was interesting to see this happen, and how you could edit the curves even after they were joined. Any shape could be created and I began to understand the full capability of NURBS and just one of the surface tools. For example, using splines along the orthographic lines of a particular object and lofting between each section could be a much faster process than building something up from scratch using poly primitives.
Lofting – bridging between splines
As mentioned previously, NURBS can also be converted to polys, making it very convenient for the modeller to use whichever option they felt was best for the task or a combination of both. By going into ‘Modify’ with the object selected, clicking the ‘convert’ arrow and opening up the ‘NURBS to Polygons’ options panel, it was easy to change any of the settings for converting. For now, we’re sticking with the default ones, which can be obtained by going into ‘edit’ in the top right corner and hitting reset. What was essential to be selected were ‘quads’ and ‘control points’, which I clicked. After applying that, a Poly duplicate was created on top of the original, which looked a little jagged. This was fixed by moving it to the side and adding Smooth Mesh Preview.
Now I had two of the same shape, one a NURB version, and one a poly. The problem now was that if I moved the original, the poly also moved along with it. As I said, the convenient aspect of them both being connected is that they can be modified at the same time. However, what if you wanted just one or the other without any extra attachments?
To separate them, in ‘edit’ and ‘delete all by type’, I click ‘history’ and make the shapes independent. You can now delete or move anything individually if you like. Looking at my object, I could imagine using this entire process for architecture and various organic models – it is quite useful!
This was demonstrated with an actual object this time. Following the instructions, I generated a NURB circle and scaled it up slightly. I then duplicated it, moved it upwards and repeated that for the third ring. We were essentially creating the bone structure of a barrel, which would then be filled in. I like this approach, however, it was hard to gauge how much to enlarge each spline and how far up to move them. I gave my best guess for the first three, duplicated the first two and then moved them up above in the opposite order for consistency in the model.
Now, instinctively, one would want to select everything and simply click ‘loft’. Lofting, however, needs a direction, so this wouldn’t work. I was told to select the bottom one and then shift select the others, going upwards in order. Only after that could I loft. The resulting shape wasn’t what I wanted because it didn’t look exactly like a barrel. This wasn’t surprising because I had done a lot of guesswork and I would usually use a deformer for such a simple shape, which is far more accurate. I moved the object away from the spline so that I could make some edits, before converting it to polygons and using ‘fill hole’ to close off the gaps at each end.
“NURBS can also be used to draw out and create entire scenes using curves and a range of other tools like planar, loft, extrude, revolve and birail. They are useful when constructing many types of organic 3D forms because of the smooth and minimal nature of the curves used to make surfaces. They are often involved in game production, scientific visualisation and industrial design.”
Another method when working with splines is creating half of the object and making a surface all around it from one end to the other around a 360° point. Once again, in front mode, I created an outline using EP curve. This time, it was half of a wine glass. This was the most frustrating part of the task because there was no template and I had to guess yet again, where to click next. Even moving some of the control vertexes didn’t help much and I didn’t have enough time to make the massive amount of changes needed to actually perfect the glass. So, I went ahead with the next step, which was to go into ‘surfaces’ again, with the spline selected, but this time click ‘revolve’. This tool used the spline as a profile and made the full glass. If the object being created wasn’t supposed to be hollow, a line running along the centre and connecting both ends of the spline would have been needed.
Unfortunately, the glass was inside out, hence the black colour. To add the shader, select ‘reverse’ in ‘mesh display’. Up until this point, aside from not liking the shape of the glass, I was alright with the steps. However, this was when things started to go wrong for me. We exited SMP and selected all of the top edges, which we then extruded inwards and downwards slightly. I understood that we were creating a rim and also fixing the hollow problem since a wine glass isn’t that flat in real life.
With the edges still selected, we went into front mode and clicked wireframe to see through the model. I followed along as accurately as I could with my wonky shape – extruding downwards in sections and pulling them out or in to match the outer shape. I don’t really understand where I went wrong, since none of the edges were touching and I had even used this method before for my teapot, so it wasn’t foreign to me.
What we should have done when finishing this was to fill the bottom hole left in the glass using ‘fill hole’ and fix the geometry, but I didn’t get to that part since I was so dissatisfied with the previous step. I had black splotches striped along with my glass and that, along with the fact that it looked rather ugly in my opinion, threw me off quite a bit. I rushed through the rendering process, adding a glass material and lighting quickly because I wasn’t too bothered with how it looked at this point.
To instantly fix geometry – ‘edit mesh’, ‘poke’
Still Life Modelling
For the remaining part of the lesson, we were given a modelling challenge as a fun way to apply our skills. Each person was given a real, solid object that they then had 30 minutes to create in Maya. The goal was to be as accurate as possible and pay attention to the intricacies of the design, even if the model wasn’t completed. These are some of the possible skills we could use, having learned them over the course of our 3D lesson time.
- Poly Primitives
- Extrude along a path
- Duplicate special
- Edge loops
- Multi cut
- Mirror
- Deformers
- Snapping
- Edit pivot point
- Bevel edge
- Extrude/keep faces together/apart
- Target weld
- Fencing
- Circularize
- Bridge
The first object that I had was some sort of Wi-Fi booster and had a very unique shape. I knew that it would be difficult to model as soon as I had it in my hands and got a good look at the details. I predicted that creating the curves of the main shape wouldn’t be too difficult however the way in which everything joined together was going to be the hard part to recreate – especially the side parts. The front curve stopped directly where the line was but it was also connected to the top, as well as curving round to behind the plug.
I decided to start modelling without a plan and find the best method I could for each part. My first step was to create the base of the plug by extruding from a poly cube to create an ‘L’ shape. I switched between smooth mesh preview and normal view often throughout the process, but the first time I applied it was after this point so that I could create the rounded edges. Of course, it was far too round, so I had to insert edge loops along each side, close to the edge so that the effect was just subtle enough to be realistically smooth.
Next, I started to work on the front curve, which I felt was one of the key defining features of the object. At first, I thought about extruding outwards and then scaling the face in, would have created a curve with SMP, but I decided against this since it wouldn’t be accurate enough. The technique I went for was manipulating the faces in soft select. I first had to add an extra edge loop towards the top, but after this, it was quite smooth and easy to create that curve and tweak it as necessary. The main issue is I kept coming back to it, dissatisfied with the angle. I wanted to get it as close as possible to the real thing, but without a template, sometimes you have to realise when to stop.
After forcing myself to move onto another part of the object to model (because of our time limit), I rotated the plug to the top and worked on creating that oval shape. I did this by scaling in the vertices on each end and scaling out the extra edge loop I added in the middle. With SMP, this created that curve that I wanted. Each of these steps seem very simple and quick to do, but there was a lot of editing and adjusting involved because I had to reference the object throughout to make sure that what I had made so far was accurate. Because of this, it took a lot longer than I would have liked it to, to create each part of the booster.
I think that I was also subconsciously procrastinating the aspects that I knew would be the hardest to replicate because I did every part except the back and the curved sides, both of which I was dreading getting to since I had no solid plan on how to make them. Perhaps with the back, I had an approach of some kind, but it would only get me so far. So unsurprisingly, the bottom is the next place I went to work on. Looking back at the object, I could see that it didn’t have a flat bottom but another curve instead, like the one on its front face. This curve ran from the bottom of the front face upwards to the contact section, creating a rounded cut-off that shrank the overall design. I followed the same method as the front when creating the curve, pulling out the faces in soft-select mode and changing any individual row or face as I felt was necessary.
This was where I took a step back and looked at what I had done so far. It didn’t look too bad – I felt that I had captured the basic form quite well, however it took far too long and there was still the matter of the side pieces and the back.
I was just doing the edge loops on the back as needed for the plug pins but we ran out of time and had to switch. Or rather, I wanted to try modelling something else.
My next object was a lock, which I thought was much more doable, considering it had a more regular shape and was more symmetrical and ‘standard’ in terms of structural complexity.
As with the wifi booster, I began with the base using a poly primitive. Once again, I chose a cube since it was the most similar to the lock as a starting point. I could see that it was wider in the middle, which I used an edge loop for. In the same way as my previous object (for the top), I used the multi-cut tool, holding the control button and clicking in the centre to add a new edge and making sure to hit ‘Q’ after so that I didn’t accidentally cut anything up. With the side scale arrows, I then stretched out the vertices of this new loop to widen the entire cube. Using ‘bevel’ with all of the edges selected, I then smoothed everything out and created those rounded edges.
Naturally, the handle came next. At first, I tried to create it using the ‘extruding along a path’ method, but I couldn’t select both the cylinder and the line so this didn’t work. Next, I thought about using reformers, but for some reason, I couldn’t apply ‘bend’ to a cylinder. This confused me since I remember bending a rounded object before when creating lamps in Maya, but I could only use a cube this time. I was frustrated because this was such an easy object to make, yet I was struggling. It felt like the program was against me, but I knew it was just my way of thinking. I don’t have the 3D problem-solving skills that come with practice and experience in modelling. Looking back, I think I could have tried using lofting or just looked through the Maya guide and found another method, but I was so annoyed and stuck that I couldn’t come up with a solution. At that point, I had spent so much time working on the basics of the lock, that our time ran out. I hadn’t gotten far at all and was rather disappointed in myself.
I understand that not every 3D lesson is going to be the same, and some are less productive than others – especially when coming back from a break, but it was a shame to see how much I struggled with creating objects like a lock and plug when I have created life-like teapots, chairs and flags. I didn’t utilise many of the techniques that were listed either, which reinforces the theory that I have; I find them easy to use and understand in lessons but they don’t really pop into my head when thinking of faster ways to build objects or obtain certain shapes. It was still quite interesting as a whole though because learning about NURBS gave me another possible method to build in the future, and the video we were shown of how lofting was used when making a car was intriguing.