oops... nvm, wrong thread.

Zigzter: Use more geo! :P

Originally Posted by Zigzter

Quick question: is this pinching just a result of not enough geo, or do I need to figure out some other way to get that corner to hold its sharpness?

Insets and extrusions on curved surfaces in subdivision modeling is notoriously hard to do because the support loops will break the equidistance and cause things like pinching. When doing a shape like this you are much better off getting the shape right, and then subdivide and do the inset last.

From left to right. Highpoly, Subdiv model smoothed, Subdiv model unsmoothed.

Hmmm, alright, I'll try that. Thanks guys.

Edit: Question though: if I get the shape right with the edgeflow in for an easy cut-out, wouldn't smoothing it (like actual smoothing, not the preview) smooth out everything and I'll lose the edge contours I want?

Example:




Since this surface is a little curved, manually moving the vertices back to straight edges would be a real hassle.

(Used the bottom instead of the cockpit area as an example, because the cockpit area is made up of a simple intersecting grid, so that problem doesn't happen there.)

Edit2: Figured out I can select all the edges on the mesh, then just deselect the future-cutout ones and smooth, then manually connect the edge loops. Is this okay? Although, that means I'll have to smooth again. Does that mean I'll have to go in and get rid of 90% of the resulting edge loops so I don't end up with a ridiculous poly count? This isn't a game model, so I don't think I'll be making a low poly.

Okay, I'm new here (and to 3DSMax) and I'm having a lot of issues with this nail gun:



1. How on earth am I supposed to get the trigger frame to morph into the handle??



2. How can I better topologize this? It's causing some pinching.



Here's a smooth of what I have so far:



I've been working on this thing for over 10 hours and have barely progressed and I feel awful now.

Fingus has the solution for your pinching, 2 posts above yours.

As for the other someone might come around, I haven't dabbled into any of this yet so I can't help but I am enjoying reading this entire thread.

This thread is going to help me so much! can't wait to learn new things!

This thread is the most interested for me, i learn a lot browsing this, but i have some problem with subd modeling.

So, i just want the red sketched shape, obviously with proper midpoly geometry, smooth edges, etc. But if try the triangle based main part geo (left), i have the problem with control loops. If trying some qouick quad base thing, the control loops on the corner of shape was too close together (green lines in right side shape). Manually put them right place was much time and not possible any geometry. I hope some help, and i know, this is not very difficult (for you)

Originally Posted by SPYFF

snip

Not perfect - it still need few support edges.

I'm trying to model a shotgun, I have always had problems with HP topo.
Could someone kindly show me how I can model this star shaped bolt/screw?
It's driving me


Thankyou!

Spyff i gave this a quick try. All quads. No guarantee that it's the best solution. I'm also still learning.

Originally Posted by cookedpeanut

I'm trying to model a shotgun, I have always had problems with HP topo.
Could someone kindly show me how I can model this star shaped bolt/screw?
It's driving me


Thankyou!

Use more geo and it's easy to just cut those pieces out :P

Adij & Steppenwolf - thank you very much, this two geo is enought proper for me!

@Millenia

Thankyou matey, managed to get it somewhat close to that. Could do with some refining, but atleast I have the shape!

hey guys, my first time Contr. to this. i know i can get help. I've been reading at this topic for a while, and i know that more geo could make or break a modell.

But could you guys please take your time to walk me through what i am doing wrong here. And should i redo the whole thing?






I need help with the red shapes.

Cheers in before hand.

Wirrexx

sorry for the bad Prt-Scrn. mode

oh btw. Im totally noob on doing SUB-D. Its hard. I feel like when it gets down to more advance shape i lose focus. But i want to challenge myself all the time.

Hey guys, I've been following Polycount's forums for a while now and have got a lot of great info. This thread in particular is pretty awesome for learning new sub-d stuff. Anyway, I have a question regarding welding together two cylinders of different sizes that meet in a T junction. I am primarily a Maya user and so far have not found an easy way of doing this. Any suggestions?

Derek: Adjust the number of segments in each cylinder so that the edges meet up. Like now you have about 2.5 polies on the bottom cylinder for each one on the top cylinder.

Thanks Perna. I tried what you said but am still having problems. I totally get how it would work when the cylinders are of equal radius and you can just rotate 90 degrees and weld it together like in example C. However, when one cylinder is smaller, no matter what combination I use for the number of sides, I can never get it to match up all the way around like you can see in example B. I just can't seem to wrap my head around a way for the topology to be that doesn't break the roundness of one of the cylinders. It's driving me nuts because it seems like there should be a very simple way of doing this that I'm just overlooking.

derek: Think about it like this; if there's no possible way of lining up the two cylinders perfectly, then just use the closest match.

The only geometry that's going to experience warping is the tiny strip between the two cylinders anyway, so it's hardly going to be noticed. And even then, the error is minimized by using more cylinder segments (add more geo etc)

edit: You have a lot of unnecessary loops on that bottom cylinder.



And a quick demonstration of not requiring a lot of geometry to pull this off:

Hey guys, I'm coming from a mostly organic background but I really want to be more comfortable in hard surface modelling. I'm modelling an environment and have come across a problem. As you can see in the image below I ran into some issues with support edges causing artefacts on the curved surface. I've posted an image of my solution but it still doesn't feel right. How would you guys go about solving this problem?

Arch: The problem here is a common one: You add flat control loops to a curve.

Flat geometry is not going to look curved.

Just make the curve have enough geo in the first place to support what you need to do with it (add more geo etc)

I just gave it a try after looking at your examples. Makes much more sense now. I think I was getting hung up on being extremely accurate and unwilling to move any verts that didn't perfectly line up. My example, which required a few small vert tweaks to line up, seems to be smoothing fine but if you still see any problems or areas of concern feel free to point them out. Thanks again.

Arch, to sum it up: You could have done A or B, but you did C. C is not a curve.

Derek: Well, you asked for it




A - You need a control loop here. Control loops go on both sides of the edge, not just one

B - This loop has no purpose

C - This control loop is getting tighter towards the top. It needs to keep a consistent distance away from the corner to give a clean result. Also, both control loops are to keep the same distance, to create a clean curve. Currently C is very tight and D is wide.

D - Same as C. Gets tighter towards the top.

E - Double control loop. Again you just need one. You only ever need maximum one control loop on each side of the shape you control. When you want a melted look on purpose, you can use only one, or none, but you never need more than two. The isoline view can be misguiding. Study the actual geometry created, not the isolines, to learn how subdivision really works.

Since I see a lot of people make this mistake, I put together an illustration:



A: 3 quads next to each other, the last one taking a steep plunge. So you'd think that would cause the entire mesh to curve, right? Wrong, any one polygon can only ever affect its immediate neighbours when subdivided. The first quad shape is in fact still the exact same shape and position it was before subdivision.

B: The first quad is still fully intact. We don't need any control loops to protect it from the curvature created by the third quad.

C: Adding control loops to get a sharp corner. Both loops are at the same distance from the edge.

D: The isoline makes it look like the curve is uneven because the one control loop is pulled down so much, but if you look at the silhouette you see that's not the case. The corner curvature is perfectly even and highlights on it will look clean.

Great stuff Perna. It is very much appreciated and makes a lot of sense in terms of what is actually going on during the subdivision. The only part that I have a question about is `B - This loop has no purpose'. I originally placed that one there to ensure the cylinder's verts from that point up to the top would stay perfectly round/aligned even when I had to move the verts below it slightly to merge with the other cylinder. I was concerned if I had to move the verts very much I would get a slight taper from where they are welded going towards the top. I thought that edge would preserve the cylinder and any taper would happen beneath that edge loop. I'm sure I'm wrong but that was at least why I thought it was needed.