I am learning modeling I need some advice is the topology good or need something else This is my first modeling Please correct me f I am wrong thank you all
Model is defined by vertices. Each vertex is three numbers, X, Y, and Z. Somewhere on your hardware those numbers get logged, and it takes a certain amount of bytes.
So if you have some vertices and they are not contributing something meaningful to the viewers eyes, then that is wasted memory. Just like if I insert a useless sentence here: "Dogs are great, and people who don't like dogs must have a tortured existence." It is words and maybe I like them, but given the goal of this paragraph they aren't serving a purpose. So it's wasted words.
Take a look at professional hard surface models and you can see the way how they weld off all the extraneous edge loops. The final model topology can look pretty ugly and nonsensical.
But beware destructive edits like that too early. Depending on the final goal for this model it may be useful to keep a relatively even mesh like this. Always save out your work with descriptive names, make use of incremental save feature, duplicate before experimenting, etc.
Peruse the "how the fuck do i model this thread?" for illuminations.
To get more specifics, you'd have to define a more specific goal. How is the model rendered, interacted with, and what is the larger context it fits within?
The material you have on right now is very matte. If you up the glossiness a bunch, it'll reveal most of your mistakes. Showing the references you used will help us diagnose the topology better. Showing it without turbosmooth applied also helps to show where you put the original geometry and will help us analyze the shapes. The most important thing is this: Are there any parts of the model you wish were more accurate? The level of detail and at which distance this is meant to be shown at is up to you to decide, and should dictate the accuracy and detail you put into the model.
1 and 2 should be the same object, no? Right now they're two objects clipping together. 3 shows clipping geometry from something underneath.
There is also something at the back that bothers me, but I would have to take a look at the model to understand and show you correctly. Would you mind uploading an FBX, obj or something without the turbosmooth applied?
hello i uploaded model and references Please take a look and I hope I don't waste too much of your time I hope it is clear enough Thank you for your help
I got your PM and will be responding tonight. Gotta be a dad and hang out with the GF first The main question is whether you want crits on everything I see, or just the thingy I mentioned in my previous post.
Hello and thank you for reply No problem, take all your time Yes I will be happy If you criticize everything you see It is an honor to have professional people Who has the experience to help beginners Thank you Thank you
Before I start, I noticed that you spent some real effort modeling each part separate. Fuckin' nice. Keep doing that and you'll have less issues going forward.
Alright, so what I noticed at the back was confirmed when looking at your model. When you connected the main body cylinder to the three conical nozzle parts, you altered the main body's cylinder so much in order to make them fit that you made it non-cylindrical in the process.
Look at how deflected these selected edges are in comparison to the rest of the cylinder. In these areas it's no longer a cylinder:
The way I would go about this would be to start by ligning up all the geometry you want to intersect properly. I chose the same amount of sides you did for the main body because I could see that the amount of sides you chose for everything just barely didn't line up. The only variations between mine and yours will be: -I noticed in the refs you uploaded that the three nozzles should be at the same Z-height. I lined mine up. A bonus to this is that it makes the geometry easier to line up. -The amount of sides to the nozzles will be determined by how many I need to make them fit nicely into the main body cylinder. 20 for the big one, 14 for the small ones.
The sizes for the nozzles I copied from yours by placing mine on top of yours and inserting them as floaters into the main body, ready for booleaning. Then I upped the amount of sides on the small nozzle until I could envision myself making the cuts shown in the nextnext pic when they're all grey.
TEMPORARY FUTURE TIME! These are the cuts I'm envisioning and they will be my supporting geometry. No alteration of the main body cylinder will happen outside these lines and that will ensure that the cylindrical shape of the main body cylinder doesn't get fucked. No, not proper fucked.
Back to current time, I selected the outermost faces of the nozzles and scaled them down until they matched yours in size. From this...
..To this: That gave me the same amount of conicity™ as you had.
Now it's time to clean up the boolean intersection. First I select every vertex that is only connected to two edges. These verts are generated in a (max) boolean operation and do absolutely nothing except hide behind other verts thinking you cleaned up your boolean, but you didn't, so everything looks whack for some reason. Select them and delete them (ctrl+backspace):
Now it's time for some cleanup. If this was some huge boolean operation with thousands of verts on something that wouldn't be important I would just select and weld the verts by some small distance, but because this is a practice in making the right decisions, and there's 20 verts to weld, I'm gonna do some thinking on this. Because the edges on the nozzles are shorter than on the main body cylinder, the error generated by altering the topology will be less noticable if I target weld the main body's verts onto the nozzle verts. The same logic applies to the nozzles. Big nozzle has longer edges, therefore weld those onto the small nozzle's verts. I also connected the verts that are selected here to create that new quad between the two nozzles. Now did I plan on having to create this quad in the planning stage when looking at how many sides I wanted each nozzle to be? Let's ask Mark Wahlberg.
A couple of edgeloops...
..And a symmetry modifier later:
This is why I bother, when you look at the results. When you squish the geometry of the main cylinder like you did, the reflections on it don't stay cylindrical. They can't. Some effort is required, but I'm guessing I spent less effort trying to select the appropriate amount of sides on cylinders and making them fit than you did. Yours on the left, mine on the right.
An added benefit I just noticed after completing this is that I see that the refs call for the small nozzles to intersect with the big nozzle, and that the big nozzle's conical shape should dominate that intersection. This workflow achieved this by accident by following the sizes of your nozzles. Mine left, yours right.
I have a bunch of other things I could critique, if you want me to. I will spend twice the amount of time to critique your work than you spent modeling, if you ask.
Now let's play a little game of "Find the Ngon". No, not the obvious ones on the flat faces of the nozzles.
I do not know how to thank you For this great great work I say it honestly I have reviewed and studied your explanation carefully once and twice I have found many benefits I swear to God I didn't find anyone Explains professional detail like you I was impressed And I noticed that you added Ngon Is it necessary when we modeling something
I do not know how to thank you For this great great work I say it honestly I have reviewed and studied your explanation carefully once and twice I have found many benefits I swear to God I didn't find anyone Explains professional detail like you I was impressed And I noticed that you added Ngon Is it necessary when we modeling something
Flat surface, triangle and n-gons are especially your friend in speeding everything up. If it doesn't cause any deformations in your sub-d model, you're fine.
People typically tell you to avoid Ngons and triangles because if you're not careful they can mess up the smoothing of your model. Understanding what catmull-clark does is essential if you want to put Ngons in your models to piss people off. That may or may not be my intention. https://www.youtube.com/watch?v=THiF7-QxKXk&ab_channel=UCDavisAcademics If you watch and understand that, you'll understand why I wanted you to not mess too much with the main body's cylinder. Altering the spacing or angles between it's edges messes with it's roundness.
On a flat surface that is supported by edgeloops, neither the Ngon nor triangle does anything. When you subd them, they'll be turned into quads on a plane that is defined as flat by the edgeloops, so they will also be flat. As Kanni3d pointet out, that makes them a great place to end troublesome edgeloops that want to interfere with your shapes. On a curve, a triangle will appear to pinch inwards towards it's center, while the Ngon will smooth evenly across all it's edges. If what you're looking for is a nice consistent curve, both of them are your enemy.
On the intersection where I left an Ngon, I could've turned it into quads, but it would have been time-consuming and a bit complicated. I could've turned it into a triangle if I wanted a sharper intersection, but I wanted a nice smooth transition between the three shapes that met there, so I left the Ngon to do it's thing.
Replies
So if you have some vertices and they are not contributing something meaningful to the viewers eyes, then that is wasted memory. Just like if I insert a useless sentence here: "Dogs are great, and people who don't like dogs must have a tortured existence." It is words and maybe I like them, but given the goal of this paragraph they aren't serving a purpose. So it's wasted words.
Take a look at professional hard surface models and you can see the way how they weld off all the extraneous edge loops. The final model topology can look pretty ugly and nonsensical.
But beware destructive edits like that too early. Depending on the final goal for this model it may be useful to keep a relatively even mesh like this. Always save out your work with descriptive names, make use of incremental save feature, duplicate before experimenting, etc.
Peruse the "how the fuck do i model this thread?" for illuminations.
To get more specifics, you'd have to define a more specific goal. How is the model rendered, interacted with, and what is the larger context it fits within?
Showing the references you used will help us diagnose the topology better.
Showing it without turbosmooth applied also helps to show where you put the original geometry and will help us analyze the shapes.
The most important thing is this: Are there any parts of the model you wish were more accurate? The level of detail and at which distance this is meant to be shown at is up to you to decide, and should dictate the accuracy and detail you put into the model.
1 and 2 should be the same object, no? Right now they're two objects clipping together.
3 shows clipping geometry from something underneath.
There is also something at the back that bothers me, but I would have to take a look at the model to understand and show you correctly. Would you mind uploading an FBX, obj or something without the turbosmooth applied?
The main question is whether you want crits on everything I see, or just the thingy I mentioned in my previous post.
No problem, take all your time Yes I will be happy If you criticize everything you see It is an honor to have professional people Who has the experience to help beginners Thank you Thank you
Alright, so what I noticed at the back was confirmed when looking at your model.
When you connected the main body cylinder to the three conical nozzle parts, you altered the main body's cylinder so much in order to make them fit that you made it non-cylindrical in the process.
Look at how deflected these selected edges are in comparison to the rest of the cylinder. In these areas it's no longer a cylinder:
The way I would go about this would be to start by ligning up all the geometry you want to intersect properly. I chose the same amount of sides you did for the main body because I could see that the amount of sides you chose for everything just barely didn't line up.
The only variations between mine and yours will be:
-I noticed in the refs you uploaded that the three nozzles should be at the same Z-height. I lined mine up. A bonus to this is that it makes the geometry easier to line up.
-The amount of sides to the nozzles will be determined by how many I need to make them fit nicely into the main body cylinder. 20 for the big one, 14 for the small ones.
The sizes for the nozzles I copied from yours by placing mine on top of yours and inserting them as floaters into the main body, ready for booleaning.
Then I upped the amount of sides on the small nozzle until I could envision myself making the cuts shown in the nextnext pic when they're all grey.
TEMPORARY FUTURE TIME! These are the cuts I'm envisioning and they will be my supporting geometry. No alteration of the main body cylinder will happen outside these lines and that will ensure that the cylindrical shape of the main body cylinder doesn't get fucked. No, not proper fucked.
Back to current time, I selected the outermost faces of the nozzles and scaled them down until they matched yours in size. From this...
..To this: That gave me the same amount of conicity™ as you had.
Back To The Future©, and this new cut will dictate where the melding of the nozzles will stop.
Now it's time to clean up the boolean intersection. First I select every vertex that is only connected to two edges. These verts are generated in a (max) boolean operation and do absolutely nothing except hide behind other verts thinking you cleaned up your boolean, but you didn't, so everything looks whack for some reason. Select them and delete them (ctrl+backspace):
Now it's time for some cleanup. If this was some huge boolean operation with thousands of verts on something that wouldn't be important I would just select and weld the verts by some small distance, but because this is a practice in making the right decisions, and there's 20 verts to weld, I'm gonna do some thinking on this.
Because the edges on the nozzles are shorter than on the main body cylinder, the error generated by altering the topology will be less noticable if I target weld the main body's verts onto the nozzle verts. The same logic applies to the nozzles. Big nozzle has longer edges, therefore weld those onto the small nozzle's verts. I also connected the verts that are selected here to create that new quad between the two nozzles.
Now did I plan on having to create this quad in the planning stage when looking at how many sides I wanted each nozzle to be? Let's ask Mark Wahlberg.
A couple of edgeloops...
..And a symmetry modifier later:
This is why I bother, when you look at the results. When you squish the geometry of the main cylinder like you did, the reflections on it don't stay cylindrical. They can't. Some effort is required, but I'm guessing I spent less effort trying to select the appropriate amount of sides on cylinders and making them fit than you did. Yours on the left, mine on the right.
An added benefit I just noticed after completing this is that I see that the refs call for the small nozzles to intersect with the big nozzle, and that the big nozzle's conical shape should dominate that intersection. This workflow achieved this by accident by following the sizes of your nozzles. Mine left, yours right.
I have a bunch of other things I could critique, if you want me to. I will spend twice the amount of time to critique your work than you spent modeling, if you ask.
Now let's play a little game of "Find the Ngon". No, not the obvious ones on the flat faces of the nozzles.
I edited your topic title, I hope you don't mind. You get more feedback if you use a descriptive title. Tips here.
@FrankPolygon has similar modeling tips in his Sketchbook, it is worth a look!
Flat surface, triangle and n-gons are especially your friend in speeding everything up. If it doesn't cause any deformations in your sub-d model, you're fine.
Understanding what catmull-clark does is essential if you want to put Ngons in your models to piss people off. That may or may not be my intention.
If you watch and understand that, you'll understand why I wanted you to not mess too much with the main body's cylinder. Altering the spacing or angles between it's edges messes with it's roundness.
On a flat surface that is supported by edgeloops, neither the Ngon nor triangle does anything.
When you subd them, they'll be turned into quads on a plane that is defined as flat by the edgeloops, so they will also be flat. As Kanni3d pointet out, that makes them a great place to end troublesome edgeloops that want to interfere with your shapes.
On a curve, a triangle will appear to pinch inwards towards it's center, while the Ngon will smooth evenly across all it's edges. If what you're looking for is a nice consistent curve, both of them are your enemy.
On the intersection where I left an Ngon, I could've turned it into quads, but it would have been time-consuming and a bit complicated. I could've turned it into a triangle if I wanted a sharper intersection, but I wanted a nice smooth transition between the three shapes that met there, so I left the Ngon to do it's thing.