Looks nice! Always enjoy seeing mechanical modeling.
Mild update. Added some more details to the frame. Working with sub-D’s and round shapes is tough. Especially when I try to get two cylinders to intersect, especially when offset. At any rate, work progresses…
Thanks! I’m most definitely a hard surface guy Organic modeling is something I think I will never wrap my head around (and just isn’t my personal area of interest).
Love it, keep it up
Have you tried the HardOps plugin? It’s designed specifically for hard surface modeling.
I had to watch a number of his getting started tutorials to get the hang of it - but that ended up making general blender skills quite a bit better as well.
Thanks! It’s a (tedious) labor of love.
I’ve seen that plugin and it looks really really nice. The purpose of this project, however, is to really teach myself modeling on a very basic level. I want to be able to build something complex without using any plugins at first. My next model, however, will probably benefit from me learning hardops.
So I spent a little time trying to teach myself the best way to merge different tubes while maintaining good Sub-D topology.
The result is that I recorded the following two tutorials (when I try to teach something, it forces me to learn it much better myself).
The upshot is that I think I am going to re-do the frame. Sigh. I think the existing frame is actually plenty fine but the purpose here is that I learn how to model correctly. So I may as well do it right the “first” time (which turns out to be the 2nd or 3rd pass in real life). So stay tuned for some more exciting pictures of the frame looking almost exactly the same but being completely different underneath.
I don’t have it in me to re-do anything at the moment. Or nothing that is already done and looking decent (I redo each piece a couple of times on average since I always seem to do it wrong the first time). I’m going to keep moving forward with the understanding that my later stuff will just be better built than the early stuff. I can always go back to redo things at the end if I feel like it is warranted.
Done is better than perfect at this point.
So… to that end. Rear shocks! (The wheel is just a duplicate of the front wheel, but will eventually have to be customized). The hardest bit today was the bracket that holds the actual rear wheel.
Quick update. Added the rear fender, hinge, and pushbar. The light was a bear. I had to do it three times to get a smooth shape that still has the two cutouts in it. If anyone can direct me to a good tutorial (paid or otherwise) for advanced sub-division modeling, I would appreciate it. I can kind of muddle my way through, but I really would like to learn more how to do it “right”.
From my point of view I cant see anything wrong with your modeling.
It is looking really good so far. If you are interested in paid tutorials, I learned a lot from cg masters Car Modeling DVD. Its pretty old by now, but you still get the basics (which you definetly dont need, judging by your renders) (http://www.cgmasters.net/training-dvds/).
But If you post the wireframe you might get some helpful tips
Thanks for the compliment
Here is the process I went through on my light (the third time through - the first two were disasters.):
I built a simple elongated cube and then added two edge loops on the side and one down the middle. I shaped the individual points till the whole shape was what I wanted more or less. Then I selected 4 faces a the top and deleted them (leaving me with a 6 sided opening). I selected those verts and used toSphere to turn them into a circle. From there it was just extruding edges to create the rest of the lip.
That worked relatively well, and a LOT better than my previous attempts which had a lot more geo to start with. Fewer points means smoother curves and is easier to work with.
But when I had to add the top shape, things got dicier. I didn’t have enough resolution up there to put in a complex cutout. Adding edge loops always puts dents or creases into the round shapes. So I set the subdiv level to 1 and applied. That way I got some extra resolution. From there I approached it the same way more or less (minus the toSphere operation). But even then I wound up pushing and pulling a lot of points, and I never got the exact shape I wanted (It’s close enough, but the shape a the top should really be a bit bigger and land closer to the circular cutout - but I don’t know how to fix this without starting over).
I feel like I need more resolution but then that would a LOT of points to manually manipulate, and I am already getting some pinching near the corners where all the points are bunching up (that I can’t fix by moving points by hand). I feel like I should be able to make this shape with fewer loops maybe? A different approach to adding resolution? A technique for adding edge loops that doesn’t put a kink into a smooth surface?
Any advice would be appreciated. Also, I will check out that tutorial you suggested. Cutouts in smooth surfaces are what cars are all about. Perhaps that would show me the correct way of doing it.
I love this kind of hard surface modeling. Watching the tread for some time and I realy like model and attention to details.
I just gave it a try myself, and this is what I came up with.
- I subdivided a Qube 2 times, and then used “to sphere” because its not perfectly round then. In the second image I forgot “to sphere”, thats why it looks a bit wobbly…
- Beveled 2 edges.
- Inset the faces.
- Slide the vertices until you have the desired form.
- Delete the obsolete faces and extrude inwards.
- For the round part, i just deleted 2 vertices and added a circle with 10 vertices.
- Rotate the circle and use “bridge edge loops” to connect to the rest.
- Subdivide one more time to smoothen it.
To get the exact form you like, you may adjust the basic form of the subdivided cube before doing all those steps.
Its not exactly what you wanted, but I am also still learning, so just take my idea as some kind of inspiration
I gave your technique a shot and it seems to hold some promise. I think the difference is that you simply use more resolution than I did, which might make it a better surface on which to start cutting holes. I generally want to keep moving forward, but I think I am going to stop and re-do the rear light once more with your technique in mind to see if it makes it any easier/better.
Thanks! I’m looking forward to seeing it complete. Of course then I need to learn how to UV map so I can move on to the next level.
Thanks! Your project is an inspiration to me and if I can get it even part way as clean and detailed I will be a happy camper.
So I tried a variation on strapazie’s technique for the tail light and it is improved. Basically I started with my existing, low poly light that just had the circular hole in it and then subdivided it twice before applying. From there I selected faces like strapazie and inset them and deleted. Then it was just pulling and pushing points. Seems kind of hit or miss for me but it does work so I’ll keep at it. Seems like the real deal is to have enough resolution, but not too much. Trying to figure out what exactly that means for future parts.
At any rate, a new tail light and the start of the fuel tank. But really the biggest change here is that I upped the sample count on the render to get a cleaner image:
Thanks, I’m flattered I think you really show great potential, just don’t get discouraged if you start over a few times when you model a specific piece, that’s part of the process and you learn a lot by doing that! Your progress from the beginning is remarkable.
Thanks! Also, thanks for showing me the wires on your project. I spent a little time studying them (and will do so some more going forward) and it really shows me how much planning you put into your base mesh shape. You use a lot less initial resolution on your pieces but you manage to get much smoother results. I spent a little time tonight just making some joined cylinders. It took me a moment to get it down, but now I think I have a better handle on it.
Here is a snapshot of my current frame. There is too much resolution in it for the base mesh and it has some areas where I should have left it as quads vs. triangulating it.
And here is some practice geo from tonight where I try to minimize my original amount of geo and avoid triangles all together.
So seeing your wireframe has been super helpful. The next step is to look through the whole frame and try to figure out the “correct” amount of base resolution that will give me spans where I need them (8 sides on the main tubes is probably too little).
I’ll just keep pushing on. I don’t really get discouraged by this since I know it is all a learning process (and since I am doing this for myself on my own schedule I have the luxury of time).