Hard edges on curved surfaces

The first of a series of tutorial on hard surface modeling. It cover the basics of hardening edges on single and double curvature surfaces.

Stay tuned for other modeling tutorials


im sorry if i seem rude, but you do not qualify to make a video like this.

to achieve what you are trying to do, you dont need so many subdivisions, and you dont have to fiddle with merging. you simply add 2 loops that you need, and do an insertion which by default is bound to “i”.

you dont use the hotkeys, such as Alt E for aditional extrudtion options, CTRL+1-6 for subsurface, holding D for drawing,

in your subtitle you seem like one of those idiots who try to use big words to sound smart, you clearly dont know how to use techniqual terms, and you use them incorrectly. also considering this is a “basic” tutorial you cant expect beginners to understand it.

such as “first of all we have to seach the original edges of the curved surface, becayse they must remain in the modified surface” that literally made no sence at all… and so is the case for the rest of the text aswell, this is just the example i happened to see when clicking a random place on the timeline.

you use very suboptimal workflows such as merging individual points rahter than sliding and removing doubles, which is W+R as i expect you not to know…

and on your second attmept you go to TOWN with the subdivisions, which is not nessasary AT ALL! its horrible!

please take down the video, and do not attempt to make tutorials until you actually know what you are doing.

Since you slammed him and forced him to remove his videos… I hope you have a video ready showing what you feel is a better method.

what do you mean a “better method”? he is simply not qualified for making tutorials at this point, once he masters the program he may feel free to make as many as he wants.

ive even kept myself from making public tutorials cause i know even i cant do better than the once that are already out there.
i only make private tutorials targeting individuals with very specific issues.

if you want me to recreate the tutorial he did, sure i can but for what reason? i still would not wish it to be public as it covers a VERY specific issue.

Then why bully him based on your opinion? The beauty of art is that there are many ways to do one thing. So YOU dislike his method… that’s no reason to shoot him down and force him to feel a need to remove his videos. Constructive criticism helps… what you did was bash and bully. I’m thinking that wasn’t your intention, but that is the result. I am interested in how you would solve the issue he presented.

if you didnt see the video then dont comment. he was trying to show how to do a very spesific thing, but the model he ended up with had a horrible topology, and the subtitle of his literally made no sence. he tried using words he probably didnt even know what ment.

i dont usually judge how someone gets to a sertain result, but his workflow was also VERY suboptimal.
and considering he presented this as a beginner tutorial, id consider it more missleading than anything.

That wasn’t just criticism. That was a kicking. And not really warranted. I’ve seen far worse attempts. You could put a video up in response and link it, instead of taking a hammer to people. Or link something from an author you respect and cite that as an example of good practice.

I did see it.

I might not agree with the things said and how they’ve been said, including the word “sence”, but I do agree that the tutorial was something world can do without. The author can do better.

It’s not easy to make quality tutorials and the quality of the tutorial is how viewers rate it, not how well you know your stuff.
(similar critique https://blenderartists.org/forum/showthread.php?396350-Table-tutorial&p=3034848&viewfull=1#post3034848)

This one started with “as you all know, you have to add loops to make hard edges” or something like that. That’s not completely true, and it started with an assumption that the viewer already knows the things in the video and would sit through it to maybe learn another workflow, which wasn’t very efficient.

It didn’t teach much, so because of that it could’ve been more of a tip instead of tutorial. Even as that, there are videos to compare to https://www.youtube.com/watch?v=nBkwodrQq_4

It was clear the user is not comfortable with Blender. It’s not the lack of used hotkeys, but rather difficult ways of doing things. If nothing else, sequential selection of mesh elements is a dead giveaway. The tutorial was supposed to be about controlling hard edges on curved surfaces with subdivision surface modeling: what, why, and how is what one wants to learn from it, not sit through selections that has to be timelapsed.

The reason it’s a dead giveaway is because the timelapsed selection for me is under 2 second thing realtime, and I could spend half of that time for the shortest nap ever. Even if you’re not very familiar with selection tools, selection ABC would’ve been faster without the need for timelapse. Selection ABC: All/none, Box, Circle.

But the world could do without unsolicited opinions too. And yet here we are. :slight_smile:

I wouldn’t disagree with points made. Just the way they were offered. You managed to phrase your posts in such a way as to be critical and honest. Albeit a little blunt. I don’t see anything to be gained by being overzealous. Particularly with someone who does seem inexperienced.

maybe i was a bit harsh, it was primarily for him to actually take it down. considering he posted the video 2 places on this forum, who knows where else he’d post it. my point is, i did not want him to spam it to a bunch of beginners. i will keep the feedback in mind if i ever critique someone like this again and not be so aggressive, but i wont hold my tongue.

if i stumbled across the video i think id handle it a lot better but i believe the fact that he posted the video 2 places on this forum is what triggered me. also considering one of them were in the support section.

one thing is to make a video, another thing is to market it.

but i apologise for my overreaction.

If something genuinely bothers you, there’s no need to keep quiet about it. And my opinion on it, is just that.

An opinion.

It just seemed a surprising approach, considering the number of very helpful replies I’ve seen from you. :slight_smile:

I don’t think it’s so much an artistic subject as it is technical. I know your intent was well meaning. But technically speaking there are more efficient ways of approaching that particular subject. The result could be considered artistic though.

I’d have left the video up just to piss you off.

And if that wasn’t enough, I would have made another one just for the sole purpose of setting you off.

Your tutorial video better than his was? Well let’s see it big mouth, show us your award-winning tutorial and really show us all how to do it the “right” way.

Then we can all bask in the glory of your tutorial-making perfection…

first, ur late to the conversation, so you lack context. second, there might be a reason i dont make tutorials? there are already better once out there.

Hey pool7n. Sorry to see the reception your video got. Thought it was ok and took a screencap of the main point for personal reference. Had me experimenting with topology and alternate potentially quicker ways to get there. It has worth. Hope you keep making and sharing them.

It didn’t teach much, so because of that it could’ve been more of a tip instead of tutorial. Even as that, there are videos to compare to https://www.youtube.com/watch?v=nBkwodrQq_4

cgcookie make nice tuts sure, but this one doesn’t cover / solve shading artefacts on single or double curvature surfaces as OPs did. If any do please share. Thanks.

Screen cap for the record:


There is sort of one in the sci-fi helmet tutorial.

Controlling subdivision surfaces with perimeter loops to make hard corners and edges on a rounded surface comes from the way the subdivision surface algorithm works. It subdivides and averages the distance between vertices, which makes an approximate smoothing on a spline:

A. Placing vertices closer together limits the distance where the subdivision and averaging happens
B. The same happens when controlling surfaces, vertices closer together makes the edges between subdivide much less further apart

On a sphere the face sizes should be equal and uniform, which means the vertices are equal distance apart and would make a perfect sphere, if the subdivision surface algorithm would make it perfect. It doesn’t though, the algorithm approximates the curvature, instead of interpolating between the control points.

A polysphere consists of quadrilateral faces and the face sizes gradually change. The change in curvature is much less noticeable that way. The higher the resolution, the less the curvature changes.

In practice the same should happen when adding detail. Face sizes on a uniform rounded surface should gradually change, otherwise the subdivision gets concentrated on a much smaller area and it makes a harder turn and moves the surrounding surface, making the change much more visible.

But it’s a compromise

  1. The hard edges on the cube form that is connected to the surface needs the vertices on its corners to make them sharp, but at the same time the vertices on the sphere form should stay equal distance apart as much as possible
  2. Having the face flow turn on the corners allows that
  3. But as much as we try keeping the vertices nicely apart on the sphere surface, the face sizes change and there are edges along which the subdivision happens in the corners, and the surface change becomes noticeable
  4. To keep the cube and sphere forms as uniform as possible they would have to be unconnected, but connecting them will change how the surfaces are subdivided, so the task is to try and hide the unwanted changes when connecting them
  5. one way to do that is to increase the resolution so much that the change is minimal between the two forms, but if you do that, there’s no need for subdivision surface modifier because you just added the subdivisions directly in the mesh.

The compromise between adding resolution and moving vertices apart hides the surface change in the corners quite well. That won’t be enough if the surfaces are made highly reflective because the reflections will get warped as the subdivision surface averages the vertex distances apart on the subdivided surface. The structure flow has to make turns on the surface or it wouldn’t be possible to have the vertices for the detail forms, but that also makes the surface move.

Getting around that would require each form have uniform surface and the placement of vertices only changing where there is a form change between them. Not a fun thing to do, but there is another way: Could use custom normals to make the rounded form appear uniform

On sphere A the reflections get pinched towards the corners because that’s where the structure turns.

Spherical form on the object B takes the surface normals from a sphere without the detail. That makes the reflections on the spherical form deform as the cube detail wasn’t even there. It uses data transfer modifier to get the custom normals, and changes them on the vertex group that is set in the modifier. The group has all the vertices from the sphere form.

That’s the theory behind it. Maybe someone else shows how to use the tools.

Thank you and apologies JA12. I didn’t intend so much effort be made afresh.

Thank you GrimZA. The topology fix is the same. Aside: In case I wasn’t clear, any author not only cgcookie would be of interest.

What’s still unclear are the L337 key combos one could sequence to arrive at the pictured form in the least amount of time and or moves. A challenge put to all.

This one:

Ordered command list will do.