One trick to keep up your sleeve is to remember that in real life, things are made of separate pieces. So your model can do the same. No need to match topologies. Just make them separate, and put them next to each other.
how to go about transition two beveled corners of a form into a third beveled corner?
This is a topology issue I have been running into a lot lately and each time I just barely manage to solve it, with inadequate results. Below is an image illustrating the corners in question, thank you for any help.
Yeah for a prop like this with a lot of ~90 hard edges that need beveled, I'd just do them all at once at the end. Lot less work and you can clean up weirdness without having to manage every bevel intersection.
Hey guys! I apologize in advance for not seeing this solution in the thread. I'm just starting to learn Blender and I don't understand how to deal with these corners and points where 4+ edges meet. How do I deal with this?
It's inevitable that there's going to be 5 and 3 edge poles, one of the easiest options is to try to move them to flat areas were distortions will be minimal. Could also leave ngons instead of having edges go to a 5 pole. Sometimes they don't cause much of an issue at all with smoothing, it really just depends, and managing them is mostly what subdiv modeling is about.
It's inevitable that there's going to be 5 and 3 edge poles, one of the easiest options is to try to move them to flat areas were distortions will be minimal. Could also leave ngons instead of having edges go to a 5 pole. Sometimes they don't cause much of an issue at all with smoothing, it really just depends, and managing them is mostly what subdiv modeling is about.
I tried to move it here, it seems correct, but at another angle 6 already converge, and I don’t understand how to move them
What does it look like when subdivided? It's important to show wireframe before, and subdivision after. Like in the first post of this thread:
Check the links in the first post, this kind of modeling issue has been covered many times before. For example, from Shared: My Technical Talk content comes this pretty step-by-step:
Like Zac and Eric mentioned: try subdividing the mesh and see if there's any smoothing errors like pinching, stretching, or unintended surface deformation. The unsubdivided wireframe is great for clarity but it's easier to identify specific problems by comparing it to the smoothed mesh and subdivision wireframe preview.
As far as the topology routing strategy goes, it really depends on what the model will be used for, but it's worth mentioning that a couple of posts up there's a conversation that about topology routing around a complex shape intersection and the answer there is probably applicable here too.
It's really easy to get bogged down moving loops around without seeing a lot of progress. Which is why it's generally considered best practice to block out all of the major forms first then solve any basic topology flow issues during the block out, before adding a lot of support loops. This way the shapes themselves and the existing geometry in those shapes will help guide the loop flow around the shapes.
Here's some links to other write-ups that cover how to streamline the block out process and solve topology flow issues so most of the final support loops can be added with a few simple bevel / chamfer operations.
Below is an example of what this approach could look like for this part of the model: Start by blocking out the larger curved shapes and the bulk of the complex form underneath, maintaining a few key loop paths off the curved shapes. Try to use the existing geometry in the curved shapes as part of the loop flow path for the support loops whenever possible.
This is typically done by adjusting the density of the curved surfaces and the placement of intersecting geometry so the shape transition lands between the existing segments of the curved surfaces. That way the intersecting shape and the support loops around it are
less likely to disrupt the segment spacing of the existing shape, avoiding unintended surface deformation.
Continue blocking out the secondary forms then join the two shapes together. Direct the loop flow around the combined shapes to define the basic loop flow around the shape transitions and any tertiary details. Sometimes it's easier to work from the outer most bounds of the shapes and push recessed areas inwards and other times it's easier to work from the inner most bounds of the shape and extrude the details outwards.
On this particular shape it seemed to be easier modeling inwards. So the basic loop path was created where it intersects the hemispherical shape then the lower section was pushed inwards to create the depth along that curved lip. Which avoids potential deformation issues on the back of the hemispheres.
Add the basic loop paths around the shapes once all of the major forms are accurately represented. Keeping things relatively simple at this stage makes it easier to solve a lot of basic topology flow issues and the loop paths established here will both hold the larger curves and provide a path for the final edge support loops that are added later.
With the basic loop routing supporting the larger shapes and providing a path for the support loops around the shapes, simply select the edges that define the outer profiles of the shapes and shape transitions then add the support loops with a bevel / chamfer operation. This process can be further streamlined using a bevel / chamfer modifier that's controlled by edge weights.
The example below highlights how the edge support loops created by the bevel / chamfer operation automatically resolve the problems with the corner intersections with a combination of grid and diamond quad topology. Here's some links to additional write-ups that cover a few different ways to approach these kinds of corner intersections using both grid and diamond quad topology.
Below is what the final model looks like. All of the smaller support loops, along the edges that define the shapes, are automatically generated using a bevel / chamfer modifier that's controlled by edge weights.
Most of the topology flow issues were solved early in the block out: the few support loops for the larger forms constrain all of the triangles and n-gons to flat surfaces between the edge support loops where they can't cause any unintended surface deformation. Which is why the mesh subdivides cleanly without being all quads.
Recap: Keep things relatively simple and block out all of the larger forms. Use the block out process to solve most of the topology routing issues before adding a lot of complex support loops. Try to use the existing geometry of curved surfaces as part of the support loop structure around complex shape intersections. Streamline the modeling process by adding final edge support loops around shape using a bevel / chamfer operation.
Some other topology write-ups that may be helpful on this project:
It might be worth to look not only on this example of an 4+ edge in an "abstract" way, but to have a closer look at the reference. For example if the actual topology is more like so..
..then the wanted to be rounded, chamfered, beveled edges has a very unproplematic topology. Here is a more extrem version of this..
Of course this may not fit the original reference but because of the lack of showing them here i can only generaliye: look at the real think. Artisans had to make those rounding somehow in the real world. So follow the real thing.
@Osoramiru I absolutely am one to beat a dead horse, so here's my best swing. Consider that there is no such thing as a perfect edge, chamfer, fillet, nor corner if you're relying on a curve smoothing algorithm to generate it. That includes catmull-clarke. For insignificant or miniscule edges such as the one you're making, it's fine. In my experiments, I found that I want to chamfer edges that are larger than 4mm radius for something rendered in full-screen at 1440p. I'm currently working on an OZ9, and on it's receiver, there are two slide rails, front and rear, that have the same geometry as yours. The edges are ~1mm radius, here zoomed way, waay in. First the front rail. The thickness in height of the slide rail is about 6mm.
The rear rails you can see I went with a slightly different flow, simply because there's a 45° edge on the front face, which taunts me to secure it first and foremost. The slight width-inconsistency you can see on the smoothed curve/edge is not due to my geometry, but due to catmull-clarke generating a wider and softer curve as the angle of the faces decrease.
While it absolutely is possible to find imperfections stemming from poles when zoomed in like this, much of the inconsistencies also stem from the fact that I'm using the catmull-clarke algorithm to create a radius from what is essentially 45° to 90° flat surfaces. No algorithm that I've heard of can create perfect chamfers or fillets from such a base.
It doesn't matter, though. Next is the complete part, at approximately the resolution on screen that it'll be rendered at. I've really tried my best to angle this to show the imperfections, and the material is much more glossy than it'll have in the end. The extra glossy material is there to highlight imperfections that I do want to fix, and help me see them early on as I make the HP. You either can't see them, or they're so insignificant or in such a tight spot that you won't notice. A.K.A. good enough. If you request it of me, I'll post some parts that according to my experiments need attention.
Replies
Update: Figured it out a bit. If anyone has tips or anything please feel free to share.
how to go about transition two beveled corners of a form into a third beveled corner?
This is a topology issue I have been running into a lot lately and each time I just barely manage to solve it, with inadequate results. Below is an image illustrating the corners in question, thank you for any help.
You may have to elaborate this to get a more suited answer.
Check the links in the first post, this kind of modeling issue has been covered many times before. For example, from Shared: My Technical Talk content comes this pretty step-by-step:
https://polycount.com/discussion/comment/2718563/#Comment_2718563
https://polycount.com/discussion/comment/2757752#Comment_2757752
https://polycount.com/discussion/comment/2800274#Comment_2800274
https://polycount.com/discussion/comment/2798366/#Comment_2798366
..then the wanted to be rounded, chamfered, beveled edges has a very unproplematic topology. Here is a more extrem version of this..
Of course this may not fit the original reference but because of the lack of showing them here i can only generaliye: look at the real think. Artisans had to make those rounding somehow in the real world. So follow the real thing.
I absolutely am one to beat a dead horse, so here's my best swing.
Consider that there is no such thing as a perfect edge, chamfer, fillet, nor corner if you're relying on a curve smoothing algorithm to generate it. That includes catmull-clarke.
For insignificant or miniscule edges such as the one you're making, it's fine.
In my experiments, I found that I want to chamfer edges that are larger than 4mm radius for something rendered in full-screen at 1440p.
I'm currently working on an OZ9, and on it's receiver, there are two slide rails, front and rear, that have the same geometry as yours.
The edges are ~1mm radius, here zoomed way, waay in. First the front rail. The thickness in height of the slide rail is about 6mm.
The rear rails you can see I went with a slightly different flow, simply because there's a 45° edge on the front face, which taunts me to secure it first and foremost. The slight width-inconsistency you can see on the smoothed curve/edge is not due to my geometry, but due to catmull-clarke generating a wider and softer curve as the angle of the faces decrease.
While it absolutely is possible to find imperfections stemming from poles when zoomed in like this, much of the inconsistencies also stem from the fact that I'm using the catmull-clarke algorithm to create a radius from what is essentially 45° to 90° flat surfaces.
No algorithm that I've heard of can create perfect chamfers or fillets from such a base.
It doesn't matter, though. Next is the complete part, at approximately the resolution on screen that it'll be rendered at.
I've really tried my best to angle this to show the imperfections, and the material is much more glossy than it'll have in the end. The extra glossy material is there to highlight imperfections that I do want to fix, and help me see them early on as I make the HP.
You either can't see them, or they're so insignificant or in such a tight spot that you won't notice.
A.K.A. good enough.
If you request it of me, I'll post some parts that according to my experiments need attention.