@Dvids You got the shape of the decocking lever wrong, it's not rectangular at the top but rather one continuos curve. I've attached some references that might help. It's always good to get a lot of refs from different angles to get a better idea of the overall shape. I'd build the whole lever first and do the serrations last, boolean + cleanup is a valid approach.
(It's also worth considering if you really need to model the serrations. Those are really shallow and you'd only see the silhouette change at specific angles and extreme close-ups. Doing this detail in texture would be easier. All depends on the project scope and requirements)
You are absolutely right that I misinterpreted the shape. Was doing this at 4am last night lol, think I got sleepy. I also didn't have reference showing the backside of the lever - may I ask where you found that? Just google or a specific source?
I re-created the whole thing and started with the back transition using a cut off cylinder. This way I got the main shape right before even worrying about the serration. I also understand that they are less deep in real life - I am intending to exaggerate them a little on purpose to get a nice looking normal map. If this was for production, I would not model these but solve it in Painter, but since this is a personal project with the objective to practice good topology and sub-d modeling, I am doing as many details by hand as I can.
This is the second attempt. Still have to clean up the edgeflow but got the main shape correct this time.
okidoki Thank you for actually going ahead and modeling it. Turns out I just needed better reference
Actually went back to the base mesh and built it again. This is the cleanest I managed to come up with so far. There are still ngons along the edge which I can't find a smart way to resolve, but the shading actually seems to be working regardless, so I'm just gonna leave them for now.
Funny side note - I'm modeling this handgun because 8 years ago, I abandoned the same project. I managed to find the old project files and threw them in to compare it to what I am doing today. For this declocking lever specifically.. let's just say the version from 8 years ago acurately reflects my frustration with sub-d modeling at the time 😆
Hey, so I've been working on this drum paddle for a bit and its giving me a really tough time. I got a few attempts to show but none of em "feel right" if that makes sense. I traced out a base from the top view but I got no idea how to go forward from here. Idk if I'm biting off more than I can chew with a shape like this. Any guidance would be appreciated. Attempts and Refs below.
What would be a better way to resolve these edge loops so they don't cause that sharp of a corner edge on their way out?
Your misstake is continueing the sharp edge away from the bevel details. I would " attach" the edge of those details to the edge of the model instead of continueuing the lines and try to cut them off halfway along the edge of the model.
And in order to keep those surfaces more sharp between the bevel detail and the edge, i would just use inset instead of creating more edgeloops.
@Veer_P It's generally considered best practice to block out the forms from largest to smallest. Keep things simple at first. This makes it easier to adjust the larger shapes and place geometry where it's needed to support smaller shapes and shape intersections. Work through the block out in stages and only add smaller details once the larger shapes are accurately defined and the basic topology flow smooths cleanly.
Below is an example of what this iterative block out process could look like: Start with relatively simple geometry that captures the overall shape and contours of the object then increase the mesh density as required, by adding loops with loop cut and bevel / chamfer operations or by applying the subdivision, until there's enough geometry to support the next smallest surface features.
Adjust the topology layout as required to turn corners around shapes and redirect the loop flow that defines the narrower edges between some of the shape transitions.
Here's an additional (front) view of the topology layout around the part.
There's more than one way to approach the order of operations and topology flow. Which approach makes the most sense and how complex the base mesh should be depends entirely on how accurate and detailed the final high poly mesh needs to be.
For something like this, it probably makes sense to keep the mesh relatively simple while creating the basic shapes and setting up the loop flow around the sharper edges then apply one or two levels of subdivision before adding the smaller details like the drain and screw holes.
@Zoddo The solution provided by @Neox should cleanly resolve your mesh to all quads.
Briefly adding to that: occasionally there's a subtle smoothing artifact near pole vertices and it may be necessary to adjust the vertex's height so it's consistent with the surrounding geometry. Usually it's not an issue but sometimes star poles can leave little dimples on compound curves, even after they're resolved to quads.
On surfaces that have compound curves like this, sometimes the corner details are constrained by the number and size of the radial segments and other times it's the circumferential segments that are the limiting factor. Ideally the number of segments in the curve will line up with and support the corners and any intersecting shapes.
Below is an example of how segment spacing impacts the size and quality of rounded corners on shapes with compound curves. Increasing the segment count does tend to increase the accuracy but how
much geometry is required to hold the shapes really depends on the size
of these details and the desired sharpness or quality of the corners.
It tends to be a lot easier to work through the basic shapes first,
adjusting the segment spacing of the curves whenever possible to match
edge and corner details. This also makes it a lot easier to solve any
topology routing issues and avoid creating poles out in the middle of a
curved surface. After the basic topology flow is solved the support loops around the edge of the shape can be added with a simple bevel / chamfer operation.
Replies
You are absolutely right that I misinterpreted the shape. Was doing this at 4am last night lol, think I got sleepy. I also didn't have reference showing the backside of the lever - may I ask where you found that? Just google or a specific source?
I re-created the whole thing and started with the back transition using a cut off cylinder. This way I got the main shape right before even worrying about the serration. I also understand that they are less deep in real life - I am intending to exaggerate them a little on purpose to get a nice looking normal map. If this was for production, I would not model these but solve it in Painter, but since this is a personal project with the objective to practice good topology and sub-d modeling, I am doing as many details by hand as I can.
This is the second attempt. Still have to clean up the edgeflow but got the main shape correct this time.
okidoki
Thank you for actually going ahead and modeling it. Turns out I just needed better reference
Funny side note - I'm modeling this handgun because 8 years ago, I abandoned the same project. I managed to find the old project files and threw them in to compare it to what I am doing today. For this declocking lever specifically.. let's just say the version from 8 years ago acurately reflects my frustration with sub-d modeling at the time 😆
And in order to keep those surfaces more sharp between the bevel detail and the edge, i would just use inset instead of creating more edgeloops.
Just a quick mobile drawing. Collapse red, remove pink
Below is an example of what this iterative block out process could look like: Start with relatively simple geometry that captures the overall shape and contours of the object then increase the mesh density as required, by adding loops with loop cut and bevel / chamfer operations or by applying the subdivision, until there's enough geometry to support the next smallest surface features.
Adjust the topology layout as required to turn corners around shapes and redirect the loop flow that defines the narrower edges between some of the shape transitions.
Here's an additional (front) view of the topology layout around the part.
There's more than one way to approach the order of operations and topology flow. Which approach makes the most sense and how complex the base mesh should be depends entirely on how accurate and detailed the final high poly mesh needs to be.
For something like this, it probably makes sense to keep the mesh relatively simple while creating the basic shapes and setting up the loop flow around the sharper edges then apply one or two levels of subdivision before adding the smaller details like the drain and screw holes.
Some additional write-ups that cover the iterative block out process:
https://polycount.com/discussion/comment/2751240/#Comment_2751240
https://polycount.com/discussion/comment/2751340/#Comment_2751340
https://polycount.com/discussion/comment/2776197/#Comment_2776197
Briefly adding to that: occasionally there's a subtle smoothing artifact near pole vertices and it may be necessary to adjust the vertex's height so it's consistent with the surrounding geometry. Usually it's not an issue but sometimes star poles can leave little dimples on compound curves, even after they're resolved to quads.
On surfaces that have compound curves like this, sometimes the corner details are constrained by the number and size of the radial segments and other times it's the circumferential segments that are the limiting factor. Ideally the number of segments in the curve will line up with and support the corners and any intersecting shapes.
Below is an example of how segment spacing impacts the size and quality of rounded corners on shapes with compound curves. Increasing the segment count does tend to increase the accuracy but how much geometry is required to hold the shapes really depends on the size of these details and the desired sharpness or quality of the corners.
It tends to be a lot easier to work through the basic shapes first, adjusting the segment spacing of the curves whenever possible to match edge and corner details. This also makes it a lot easier to solve any topology routing issues and avoid creating poles out in the middle of a curved surface. After the basic topology flow is solved the support loops around the edge of the shape can be added with a simple bevel / chamfer operation.