Sign up for my upcoming 3D Modeling course today: www.onmars3d.com/ And my Patreon: www.patreon.com/OnMars3D for all my project files, 3d models, bonus videos, Discord and more or just to support this channel 🙂 Discord: discord.gg/vybvJnEeya
I've been doing this for a couple of years, and just had an "aha!" moment about this the other day. I'm still getting the hang of staying organized with level of detail in general, but at least I'm getting less bad poles on edges now. Sub-d is strict! Lol. Liked and subbed!
I love this channel. I recently got back into 3D modeling and this content was excellent refresher and step by step way to re-familiarize myself. Thanks!
I was struggling with modelling for a while and I found Aunmar's car modelling series. Although it was in Maya and I'm a blender user I was still able to learn a ton load amount of stuff from that series and I applied that to one of my personal projects. A Gameboy that I was modelling and failed at least 3 times. And after watching that I literally modelled the whole Gameboy without using any booleans. Love your videos bro 🫡 .
Man, that's great to hear! I always focus teaching workflows where people can apply them to any 3D model or software. Happy to hear to it worked out for you!
I was following chamferzone's revolver tutorial and was facing this exact pinching issue with SubD enabled. Thanks for the tutorial. It's been really helpful for this specific problem I was facing.
Interesting, how did you fix the pinching when Chamferzone uses the Remesh workflow? Did you first apply subD, fix it, and then use subD again + Remesh/Corrective Smooth/Decimate?
Adding those rounded notches to the model was fasciniating. My apoproach would default to using a boolean to do the main cut-ouit work, Yours is simpler and faster. I'm not sure how you would "array" those changes around the rest of the cylinder though.
Glad it helped, to array that detail, I would separate the face loops with that specific detail since the pattern wouldn't align with the main pattern. A little extra work but it's worth it.
This is well put and I appreciate. Just learnt something new today. Another method I use is I take the base mesh into zbrush carved in the details with alphas and other sculpting techniques and retopologize that high model in Maya(I'm a Maya User). It takes some time but it has made my complex hard surface model more cleaner. But again the route to take depends on the model.
You use a very nice workflow and your patience is the key in modeling process, also as you know, do and re-do , could be the best training experience!!
Looks like an interesting technique that surely gets great results. My question would be from a practical point o view. Unless there is a shot that needs to be that close into the object, like shooting an ant crawling on that barrel and that particular shot would also be static and not animated, I personally see no practical reason to finesse that surface to that level and use additional time to do it as it will not be visible at all on medium shots or animated sequences. Is it good to know how to do it, academically? For sure, no doubt about it. Is it really practical in real life scenarios where everything is on fast forward these days? Not so sure... These kind of surface distortions tend to appear only on very shiny surfaces with almost no roughness or textures and only under certain light angles. But again, it is always good to know how to efficiently fix things.
Appreciate the comment! The practical use is as you described, primarily for Sub-d assets that absolutely need this level of surface perfection. For games, like weapon assets, it's not as necessary. It's more used for high end cgi, vfx, marketing, product visualization where you zoom in on assets at 4k, yes, you need that level of perfection. As I said in the video, I don't use this for every asset, but when I do use this workflow, it provides the results that are needed. As far as timing goes, you'd spend less time worrying about edge flow and topology so it's not like you're doubling the amount of time needed for this. Take my Datsun 240z that I modeled, probably has 100 parts, of those 100, I used this workflow on maybe 3-4 pieces that needed that level of surface perfection, like the body panels.
Do you leave the piece u detached as a separate piece or u merge it back when you are done? (not sure if you are actually working on a detached part or simply an isolated view selection)
Its just isolated. He'll attach it back when he's finished. Its just a technique. Could be for the video or just nice to isolate the area youre working in.
It's just interesting while I usually always try to use as less as geometry as possible, I just had the issue last week when I had a client asking for very small details to be included and I went the way applying a subd which I usually don't do and I wondered whether this is a legit way😅
Maybe i miss something but with that method you get the detail on a high subdivided geometry, so wont get a lower poly version right? The point is to just start modeling with more subvidisions? Thanks
Just a quick question on the shapes of the faces - on the left hand side of the final gun barrel, all of the faces are near enough square, but as we travel to the lower density on the right, they elongate into rectangles. Is this acceptable for UVs? I was thinking that preserving the shape of the square is best as it cuts down on distortion? Is it just a case of using your discretion? Thanks
You are absolutely correct. To make your life easier for uv unwrapping you would add edges to the right hand side to make the rectangles similar in size and shape to the faces on the left. However, this would be a metal component so it wouldnt really require a super accurate uv unwrapping so this would be more than enough.
That's correct, as @fullyleaded explained, for Sub-D, and even distribution of quads is the best for subdivision as well as UVs -- Glad you caught that :) I was originally using this asset with a remesh workflow but for sub-d, you'd want a more square density.
@@OnMars3D thanks, I wasn't trying to be pedantic or anything. I see it a lot on forums where people are telling others that their meshes are too dense, but computers these days are well up to dealing with heavy models, and as you have said, you could always bake them anyway. Speaking of baking it would be great to see some videos taking these same files and how you approach it. Many thanks for such invaluable work
@chrisjhart all good! That's one of my pet peeves too, when people say a model of too high or low poly with no context. And yeah, I'll be covering the baking process in full detail.
I'm interested in your topology course but I'm not a Patreon member. Is there any solution to get informed as soon as it will be available? Newsletter?
In fact, you don't need perfect Sub-D for this kind of hard surface modeling unless you need deformation effects. For game engines, such resources usually have too many unnecessary vertex data.
@@OnMars3D For hard surfaces, what we need are perfect and accurate normals rather than perfect surfaces. As long as the normals are correct, whether the topology is perfect or not doesn't really matter in my opinion.
@@UML649 We're saying the same but different things here. For a game ready asset, I'm talking about texture baking, so one way to use sub-d model with perfect surfaces is to bake it down into a texture map for a low poly mesh. For a VFX/CGI workflow that requires a quad sub-d workflow, the Sub-D model is perfectly fine as is. Different workflows for different use cases.
@@OnMars3D Thank you for your input. As a technical artist with over ten years of experience, I find this approach somewhat inefficient. I prefer to use CAD tools (such as Rhino or Plastisity) to create hard surface models, and then use Blender to create low-poly versions. In my opinion, except for character models (which require deformable meshes for animation), other types of models do not necessarily need to strictly follow quad sub-d, as it can be inefficient and unnecessary.
@@UML649 I agree, for game ready assets, CAD models work great since it can give you a high and low poly model from one source. I use CAD modeling tools on a regular basis and will cover the Plasticity > Blender > Unreal Engine in future videos. But for non-game workflows, like VFX/CGI where they rely on a quad/sub-d workflow, the model needs to be sub-d ready. For example, once you have a CAD model in Maya/Blender, you can't increase the level of detail if the camer pushes too close. I talk about this exact topic here where I discuss quads vs tris/ngons here: ruclips.net/video/wRPtxydqDuE/видео.html That's what I mean by different workflows for different use cases.
![Jack Harlow - Hello Miss Johnson [Official Music Video]](http://i.ytimg.com/vi/Q86_nlRoIGw/mqdefault.jpg)
Sign up for my upcoming 3D Modeling course today: www.onmars3d.com/
And my Patreon: www.patreon.com/OnMars3D for all my project files, 3d models, bonus videos, Discord and more or just to support this channel 🙂
Discord: discord.gg/vybvJnEeya
I've been doing this for a couple of years, and just had an "aha!" moment about this the other day. I'm still getting the hang of staying organized with level of detail in general, but at least I'm getting less bad poles on edges now. Sub-d is strict! Lol. Liked and subbed!
I love this channel. I recently got back into 3D modeling and this content was excellent refresher and step by step way to re-familiarize myself. Thanks!
I was struggling with modelling for a while and I found Aunmar's car modelling series. Although it was in Maya and I'm a blender user I was still able to learn a ton load amount of stuff from that series and I applied that to one of my personal projects. A Gameboy that I was modelling and failed at least 3 times. And after watching that I literally modelled the whole Gameboy without using any booleans. Love your videos bro 🫡 .
Man, that's great to hear! I always focus teaching workflows where people can apply them to any 3D model or software. Happy to hear to it worked out for you!
wow man, that is a great explanation of how to get nice geometry, there is no short cut for it, it takes effort. thank you for this tutorial
I was following chamferzone's revolver tutorial and was facing this exact pinching issue with SubD enabled. Thanks for the tutorial. It's been really helpful for this specific problem I was facing.
Glad this helped you out!
Interesting, how did you fix the pinching when Chamferzone uses the Remesh workflow? Did you first apply subD, fix it, and then use subD again + Remesh/Corrective Smooth/Decimate?
Adding those rounded notches to the model was fasciniating. My apoproach would default to using a boolean to do the main cut-ouit work, Yours is simpler and faster. I'm not sure how you would "array" those changes around the rest of the cylinder though.
Glad it helped, to array that detail, I would separate the face loops with that specific detail since the pattern wouldn't align with the main pattern. A little extra work but it's worth it.
Literally blew my mind. Never thought of doing this. Cheers.
Happy to share!
This is well put and I appreciate. Just learnt something new today. Another method I use is I take the base mesh into zbrush carved in the details with alphas and other sculpting techniques and retopologize that high model in Maya(I'm a Maya User).
It takes some time but it has made my complex hard surface model more cleaner. But again the route to take depends on the model.
Yeah, there are a bunch of ways of doing the same thing, I've used the Zbrush sculpt method too.
You use a very nice workflow and your patience is the key in modeling process, also as you know, do and re-do , could be the best training experience!!
Appreciate it, more to come!
Thank you so much for making this video!
Thanks for watching! I'm glad you enjoyed the video.
you explain it so succinctly
Appreciate it, happy to help!
Looks like an interesting technique that surely gets great results. My question would be from a practical point o view. Unless there is a shot that needs to be that close into the object, like shooting an ant crawling on that barrel and that particular shot would also be static and not animated, I personally see no practical reason to finesse that surface to that level and use additional time to do it as it will not be visible at all on medium shots or animated sequences. Is it good to know how to do it, academically? For sure, no doubt about it. Is it really practical in real life scenarios where everything is on fast forward these days? Not so sure... These kind of surface distortions tend to appear only on very shiny surfaces with almost no roughness or textures and only under certain light angles. But again, it is always good to know how to efficiently fix things.
Appreciate the comment! The practical use is as you described, primarily for Sub-d assets that absolutely need this level of surface perfection. For games, like weapon assets, it's not as necessary. It's more used for high end cgi, vfx, marketing, product visualization where you zoom in on assets at 4k, yes, you need that level of perfection. As I said in the video, I don't use this for every asset, but when I do use this workflow, it provides the results that are needed.
As far as timing goes, you'd spend less time worrying about edge flow and topology so it's not like you're doubling the amount of time needed for this.
Take my Datsun 240z that I modeled, probably has 100 parts, of those 100, I used this workflow on maybe 3-4 pieces that needed that level of surface perfection, like the body panels.
I see that you use ngons in some areas, it could produce pinch or artifacts?
Thank you as always!!
No Problem!
Is creasing the edge same as adding edge loop?
Do you leave the piece u detached as a separate piece or u merge it back when you are done? (not sure if you are actually working on a detached part or simply an isolated view selection)
Its just isolated. He'll attach it back when he's finished. Its just a technique. Could be for the video or just nice to isolate the area youre working in.
It was just done that way for demonstration purposes, I originally modeled it into the main model.
It's just interesting while I usually always try to use as less as geometry as possible, I just had the issue last week when I had a client asking for very small details to be included and I went the way applying a subd which I usually don't do and I wondered whether this is a legit way😅
It comes in handy for those types of requests :)
Maybe i miss something but with that method you get the detail on a high subdivided geometry, so wont get a lower poly version right? The point is to just start modeling with more subvidisions? Thanks
You can bake the geometry down to the low poly, I cover that here: ruclips.net/video/C_RqdNbYOjE/видео.htmlsi=vPQJC1WcNQjsGqXH
Just a quick question on the shapes of the faces - on the left hand side of the final gun barrel, all of the faces are near enough square, but as we travel to the lower density on the right, they elongate into rectangles. Is this acceptable for UVs? I was thinking that preserving the shape of the square is best as it cuts down on distortion? Is it just a case of using your discretion? Thanks
You are absolutely correct. To make your life easier for uv unwrapping you would add edges to the right hand side to make the rectangles similar in size and shape to the faces on the left. However, this would be a metal component so it wouldnt really require a super accurate uv unwrapping so this would be more than enough.
@@fullyleaded thanks very much!
That's correct, as @fullyleaded explained, for Sub-D, and even distribution of quads is the best for subdivision as well as UVs -- Glad you caught that :)
I was originally using this asset with a remesh workflow but for sub-d, you'd want a more square density.
@@OnMars3D thanks, I wasn't trying to be pedantic or anything. I see it a lot on forums where people are telling others that their meshes are too dense, but computers these days are well up to dealing with heavy models, and as you have said, you could always bake them anyway.
Speaking of baking it would be great to see some videos taking these same files and how you approach it. Many thanks for such invaluable work
@chrisjhart all good! That's one of my pet peeves too, when people say a model of too high or low poly with no context. And yeah, I'll be covering the baking process in full detail.
Cool
Can you please make another video for remesh workflow for highpoly SubD mesh like this one
That's the plan, I originally made this revolver using the remesh workflow, so I'll have more videos in the future about that!
I will look forward to it
Great, thank you!
Happy to share!
can you make tutorial on how to make doom helmet
Amazing🔥
Thank you my friend!
I'm interested in your topology course but I'm not a Patreon member. Is there any solution to get informed as soon as it will be available? Newsletter?
Yes! I'm in the process of setting up a new site, and will share on all my socials once it's up.
@@OnMars3D Thank you for your answer! Waiting for the big annoucement! :)
You are the best.
Thank you :)
A topic where I envy game artists who can use floaters.
That's definitely a good trck to use for real-time assets.
all i can say, that's right.
In fact, you don't need perfect Sub-D for this kind of hard surface modeling unless you need deformation effects. For game engines, such resources usually have too many unnecessary vertex data.
The purpose of this sub-d model was to bake down to the low poly model, that's why the high poly was created with perfect surfaces.
@@OnMars3D For hard surfaces, what we need are perfect and accurate normals rather than perfect surfaces. As long as the normals are correct, whether the topology is perfect or not doesn't really matter in my opinion.
@@UML649 We're saying the same but different things here. For a game ready asset, I'm talking about texture baking, so one way to use sub-d model with perfect surfaces is to bake it down into a texture map for a low poly mesh.
For a VFX/CGI workflow that requires a quad sub-d workflow, the Sub-D model is perfectly fine as is. Different workflows for different use cases.
@@OnMars3D Thank you for your input. As a technical artist with over ten years of experience, I find this approach somewhat inefficient. I prefer to use CAD tools (such as Rhino or Plastisity) to create hard surface models, and then use Blender to create low-poly versions. In my opinion, except for character models (which require deformable meshes for animation), other types of models do not necessarily need to strictly follow quad sub-d, as it can be inefficient and unnecessary.
@@UML649 I agree, for game ready assets, CAD models work great since it can give you a high and low poly model from one source. I use CAD modeling tools on a regular basis and will cover the Plasticity > Blender > Unreal Engine in future videos.
But for non-game workflows, like VFX/CGI where they rely on a quad/sub-d workflow, the model needs to be sub-d ready. For example, once you have a CAD model in Maya/Blender, you can't increase the level of detail if the camer pushes too close. I talk about this exact topic here where I discuss quads vs tris/ngons here: ruclips.net/video/wRPtxydqDuE/видео.html
That's what I mean by different workflows for different use cases.
using chamferzones revolver tut technique i can get the high and low at the same time in one modeling go.
I'm using the remesh workflow on the AGL Long Colt, so I will be covering videos on that workflow in future videos.