Multi-res, subdivision sculpting with UVs - Feature Request

[AbnRanger]

1 hour ago, Elemeno said:

i personally think adding multires was a terrible move..

i use 3d coat for concept work , so ill throw together a quick example and sculpt it properly in zbrush ..

the reason i cant do proper work is because 3dcoat is constantly destroying its topology , having tris when you subd or "res+" the triangles make dense areas , becoming little nipples over the mesh ...

until quad based workflow is added is hard to do anything polished

tris are no good for sculpting

calculation time for quad based multires is much faster and doesnt destroy them mesh ...

doing anything in a lower res in multires and then going to top layer just destroys the whole topology
 

hopefully surface mode has quads soon and voxels stay tris ..

- The Multi-Level Resolution feature has it's purpose and usefulness. Mainly, to make stepping up and down levels faster. With the Proxy system, I was fine with it, but it would take several seconds to decimate each time.

- Adding Quad Mesh sculpting to the Sculpt workspace is not going to be the Holy Grail some are making it out to be. You have to always be concerned about topological flow and polygonal stretching, with Quads. It will not be interchangeable with Voxels or Dynamic Subdivision Meshes, already in 3DCoat. For this reason, I am not a big fan of Andrew spending a lot of development time on a feature that had big limitations for ZBrush users...hence the introduction of Dynamesh and Sculptris Pro, to counteract those. I would rather that development time go into improving the Paint Workspace and some UI improvements.

- The little spots mentioned can be addressed by hitting the ENTER key...basically Dynameshing the Surface mesh.

- Anything I could sculpt in ZBrush with quads, I can sculpt equally well in 3DCoat in triangles/Voxels. I have tested them side by side, model for model, brush for brush (Clay and Clay) and 3DCoat is pretty darn close in my tests. Yes, some things will work a bit better from one app to another, but the bottom line for me is this... if I am paid to sculpt a character or animal to a high detail level, there is nothing I could do in ZBrush that I couldn't also do in 3DCoat, in a similar time frame. 

 

[Elemeno]

2 hours ago, AbnRanger said:

- The Multi-Level Resolution feature has it's purpose and usefulness. Mainly, to make stepping up and down levels faster. With the Proxy system, I was fine with it, but it would take several seconds to decimate each time.

- Adding Quad Mesh sculpting to the Sculpt workspace is not going to be the Holy Grail some are making it out to be. You have to always be concerned about topological flow and polygonal stretching, with Quads. It will not be interchangeable with Voxels or Dynamic Subdivision Meshes, already in 3DCoat. For this reason, I am not a big fan of Andrew spending a lot of development time on a feature that had big limitations for ZBrush users...hence the introduction of Dynamesh and Sculptris Pro, to counteract those. I would rather that development time go into improving the Paint Workspace and some UI improvements.

- The little spots mentioned can be addressed by hitting the ENTER key...basically Dynameshing the Surface mesh.

- Anything I could sculpt in ZBrush with quads, I can sculpt equally well in 3DCoat in triangles/Voxels. I have tested them side by side, model for model, brush for brush (Clay and Clay) and 3DCoat is pretty darn close in my tests. Yes, some things will work a bit better from one app to another, but the bottom line for me is this... if I am paid to sculpt a character or animal to a high detail level, there is nothing I could do in ZBrush that I couldn't also do in 3DCoat, in a similar time frame. 

 

... its hard to compare.. hard surface 3dcoat falls to its knees but as i said , getting meshes together and making a concept is nice, i always think personally the paint needs loas of love ... especially with how they have layers etc... but apparently the paint and ui looks and feels like photoshop in 3d coat2023 ... so its a very welcome update

[AbnRanger]

2 hours ago, Elemeno said:

... its hard to compare.. hard surface 3dcoat falls to its knees but as i said , getting meshes together and making a concept is nice, i always think personally the paint needs loas of love ... especially with how they have layers etc... but apparently the paint and ui looks and feels like photoshop in 3d coat2023 ... so its a very welcome update

Have you watched any of Anton Tenitsky's videos? He has a few project series dedicated to working on Hard Surface models and I doubt he has such a jaded opinion about this kind of work in 3DCoat. There are some guys on Artstation that do killer Hard Surface modeling work, also.

 

[AbnRanger]

Long Pham has some MASSIVE scenes/models almost exclusively Hard Surface modeling/Kitbashing.

https://www.artstation.com/long-pham

[Elemeno]

i think youre getting me confused... im saying that i use certain software for certain things ... one is not better than the other .

[Carlosan]

@Elemeno doing anything in a lower res in multires and then going to top layer just destroys the whole topology

Show this issue on a video will help, could you create one ? thanks

[Elemeno]

18 hours ago, Carlosan said:

@Elemeno doing anything in a lower res in multires and then going to top layer just destroys the whole topology

Show this issue on a video will help, could you create one ? thanks

will do sorry... been working alot , will try and test it now

[Elemeno]

18 hours ago, Carlosan said:

@Elemeno doing anything in a lower res in multires and then going to top layer just destroys the whole topology

Show this issue on a video will help, could you create one ? thanks

 

i have reported this before but it may have been overlooked

[Carlosan]

@Elemeno Please test latest version, thanks

3DCoat-2023-04 
- Fixed multiresolution problem when dirty noise may appear on areas not covered by the multiresolution mesh.

- Much faster preparation process for the multiresolution, especially in case of the multiresolution using the retopo mesh.

- UI tweaks in multiresolution dialog

[Tieguaili3D]

On 2/11/2023 at 6:06 PM, AbnRanger said:

Did you get a chance to watch this video? The objective when using it (CONFORM RETOPO MESH), is not to see if or how quickly you can break it, but to use it properly, as it was designed. You know ahead of time that 3DCoat has to perform a secondary calculation, at the end of each brush stroke (snapping of the Retopo mesh to the High Poly Sculpt Mesh). Let's not forget that ZBrush's Subdivision workflow, as well as its Sculpt Layers, has its limits, too. I can make a similar video, trying to quickly break it also, but what would that prove? Some ZB tools will not work with it's Subdivision levels either. Dynamesh and Sculptris Pro are 2 of them.

 

The fact that the Retopo Mesh is kept separate (ready for baking) is NOT a problem. It's actually a BIG asset, because that means you can use whatever tools you need to in the Sculpt Room without worrying. Conform Retopo mesh is not a Silver Bullet or Magic Pill, but in my opinion it provides much more flexibility for the artist and assurance that they can use their original low poly UV mapped mesh in the end (for a baking target) and have it conformed to virtually all sculpting edits made...regardless of whether I switch to Voxel mode, or sculpt with dynamic subdivision. You can absolutely forget about trying to do that with ZBrush's Quad-based Subdivision levels workflow.

 

 

saying "you're using it wrong" is extremely lazy when someone is showing you the weaknesses of the current tools and using that to show why a proper multires with proper topology is needed. I have no idea why so many people in this community are so hell-bent on this "what we have is absolutely perfect and nothing else could ever be of use so don't even try asking" mentality, there would only be benefits from giving surface mode proper topology preservation, no drawbacks, it wouldn't prevent anyone from using any one of the existing methods but it *would* allow people to use basemeshes effectively without having to go through a bunch of hacky extra steps to make an inferior tool semi-functional (yes semi-functional, it still doesn't give you the flexibility of proper multires work).

[Elemeno]

1 hour ago, Tieguaili3D said:

saying "you're using it wrong" is extremely lazy when someone is showing you the weaknesses of the current tools and using that to show why a proper multires with proper topology is needed. I have no idea why so many people in this community are so hell-bent on this "what we have is absolutely perfect and nothing else could ever be of use so don't even try asking" mentality, there would only be benefits from giving surface mode proper topology preservation, no drawbacks, it wouldn't prevent anyone from using any one of the existing methods but it *would* allow people to use basemeshes effectively without having to go through a bunch of hacky extra steps to make an inferior tool semi-functional (yes semi-functional, it still doesn't give you the flexibility of proper multires work).

nobody is saying we dont want proper topology, infact they confirmed on youtube that 3dc was getting quad topology in surface mode...

but changing from tris to quads and then changing how multires works and how RES+ works takes time ...

give them some time , they are working on paint room, modelling and sculpting as much as possible

 

they arent a big team just support and hopefully the features turn up

[AbnRanger]

2 hours ago, Tieguaili3D said:

saying "you're using it wrong" is extremely lazy when someone is showing you the weaknesses of the current tools and using that to show why a proper multires with proper topology is needed. I have no idea why so many people in this community are so hell-bent on this "what we have is absolutely perfect and nothing else could ever be of use so don't even try asking" mentality, there would only be benefits from giving surface mode proper topology preservation, no drawbacks, it wouldn't prevent anyone from using any one of the existing methods but it *would* allow people to use basemeshes effectively without having to go through a bunch of hacky extra steps to make an inferior tool semi-functional (yes semi-functional, it still doesn't give you the flexibility of proper multires work).

It's not lazy. It's telling the truth. What is lazy is not taking the time to watch a video that shows how the tool is supposed to be used, and yet still arguing about it's effectiveness. There is a right way and a wrong way to use any tool; whether it is in 3DCoat, ZBrush, Blender, etc. I know how this toolset works because I am the one who requested it from Andrew, as a solution (for preserving the original quad mesh with UV's intact, yet have it conform to all the changes to the High Poly Sculpt object). It may not be perfect, but neither is ZBrush's Subdivision workflow. It has its own drawbacks and limitations, also. That is why Pixologic added their own "bunch of hacky extra steps/tools (Dynamesh + ZRemesher)" to work around those limitations. Still, if you insist on using ZBrush's SubDivisions workflow, you WILL lose your original mesh + UV's the moment you use Dynamesh. If you don't use Dynamesh, then you have to deal with the limitations of the SubDivisions workflow...primarily polygonal stretching...which more subdivisions doesn't really solve.

Personally speaking, I like that 3DCoat keeps the original mesh separate from the high poly sculpt, and that the changes to the high poly sculpt can be reflected in the original, low poly quad mesh that has UV's. Whatever small issues I might have to deal with (using CONFORM RETOPO MESH), seem easier to handle and is more flexible than it is in ZB. An example is the fact that I can use whatever mode or tool I need to, in the Sculpt Workspace, without worrying about losing my original low poly mesh and it's UV's. I can use Voxels or Surface mode (with dynamic subdivision built into/available for every brush). I can paint on the model and sculpt simultaneously and bake later if I want. I can use Voxel Paint if I want, as well.

[Tieguaili3D]

4 hours ago, AbnRanger said:

It's not lazy. It's telling the truth. What is lazy is not taking the time to watch a video that shows how the tool is supposed to be used, and yet still arguing about it's effectiveness. There is a right way and a wrong way to use any tool; whether it is in 3DCoat, ZBrush, Blender, etc. I know how this toolset works because I am the one who requested it from Andrew, as a solution (for preserving the original quad mesh with UV's intact, yet have it conform to all the changes to the High Poly Sculpt object). It may not be perfect, but neither is ZBrush's Subdivision workflow. It has its own drawbacks and limitations, also. That is why Pixologic added their own "bunch of hacky extra steps/tools (Dynamesh + ZRemesher)" to work around those limitations. Still, if you insist on using ZBrush's SubDivisions workflow, you WILL lose your original mesh + UV's the moment you use Dynamesh. If you don't use Dynamesh, then you have to deal with the limitations of the SubDivisions workflow...primarily polygonal stretching...which more subdivisions doesn't really solve.

Personally speaking, I like that 3DCoat keeps the original mesh separate from the high poly sculpt, and that the changes to the high poly sculpt can be reflected in the original, low poly quad mesh that has UV's. Whatever small issues I might have to deal with (using CONFORM RETOPO MESH), seem easier to handle and is more flexible than it is in ZB. An example is the fact that I can use whatever mode or tool I need to, in the Sculpt Workspace, without worrying about losing my original low poly mesh and it's UV's. I can use Voxels or Surface mode (with dynamic subdivision built into/available for every brush). I can paint on the model and sculpt simultaneously and bake later if I want. I can use Voxel Paint if I want, as well.

as i've explained countless times....the point at which you're dynameshing and the point at which you want your clean topology are very different parts of the workflow, dynameshing is for the sketching phase when you don't know the overall form of the model, and what 3d coat currently has is fantastic for that loose sketchy clay feel, personally i agree that coat's voxels are better than dynamesh in zbrush. Subdivision on your pre-UVed basemesh is for blendshapes and making a character that fits your basemesh but isn't just a minor detail alteration, the whole point is to be working within the framework of the basemesh's capabilities, you're not suoopsed to suddenly go "oh what if i add another arm and put some holes in the chest and make the head into a dragon instead of a human" that isn't what it's for. Conform does not allow smooth working for either of the jobs you need from your basemesh and working with a clean quad basemesh lets you do everything conform allows without any of the ballache of the conform tool constantly ****** up forcing you to go over to the retopo room to fix it.

[AbnRanger]

On 6/5/2023 at 6:56 PM, Tieguaili3D said:

as i've explained countless times....the point at which you're dynameshing and the point at which you want your clean topology are very different parts of the workflow, dynameshing is for the sketching phase when you don't know the overall form of the model, and what 3d coat currently has is fantastic for that loose sketchy clay feel, personally i agree that coat's voxels are better than dynamesh in zbrush. Subdivision on your pre-UVed basemesh is for blendshapes and making a character that fits your basemesh but isn't just a minor detail alteration, the whole point is to be working within the framework of the basemesh's capabilities, you're not suoopsed to suddenly go "oh what if i add another arm and put some holes in the chest and make the head into a dragon instead of a human" that isn't what it's for. Conform does not allow smooth working for either of the jobs you need from your basemesh and working with a clean quad basemesh lets you do everything conform allows without any of the ballache of the conform tool constantly ****** up forcing you to go over to the retopo room to fix it.

I understand your point, but I think you are missing something, still. One of the issues you raised was that one cannot import their base mesh (into 3DCoat's Sculpt Workspace) with UV's, and preserve the topology & UV's. I am saying that would also be the case in ZBrush when you use Dynamesh or Sculptris Pro. In both situations and in both applications, you are using that UV mapped mesh as the starting point....correct?

We clearly disagree about "Conform Retopo Mesh" and some of this is because you used it improperly in the video examples you showed on the first page of this thread. Please, instead of continuing to debate about its usefulness, watch the videos and then do some more testing. I think it really could change your views about it. 

 

[Hickz]

39 minutes ago, AbnRanger said:

I understand your point, but I think you are missing something, still. One of the issues you raised was that one cannot import their base mesh (into 3DCoat's Sculpt Workspace) with UV's, and preserve the topology & UV's. I am saying that would also be the case in ZBrush when you use Dynamesh or Sculptris Pro. In both situations and in both applications, you are using that UV mapped mesh as the starting point....correct?

We clearly disagree about "Conform Retopo Mesh" and some of this is because you used it improperly in the video examples you showed on the first page of this thread. Please, instead of continuing to debate about its usefulness, please watch the videos and then do some more testing. I think it could change your mind. 

 

Heya! Let me chime in: when you’re using a base mesh with uvs, you simply are not in a position where you need to use either dynamesh or sculptris pro. Those are used in the concept phase, where you’re still figuring out the form and general design, proportions etc. Base mesh has that established already. Your point of not being able to use dynamesh and sculptris is valid, yes, but inapplicable in the case of using a base mesh with proper uvs and topology (in which case you’d simply subdivide the base mesh with multi res levels for a denser mesh)

[AbnRanger]

41 minutes ago, Hickz said:

Heya! Let me chime in: when you’re using a base mesh with uvs, you simply are not in a position where you need to use either dynamesh or sculptris pro. Those are used in the concept phase, where you’re still figuring out the form and general design, proportions etc. Base mesh has that established already. Your point of not being able to use dynamesh and sculptris is valid, yes, but inapplicable in the case of using a base mesh with proper uvs and topology (in which case you’d simply subdivide the base mesh with multi res levels for a denser mesh)

Okay, understood. This is making my point for me, actually. I am saying that it is an unfair complaint to say that 3DCoat cannot keep a base model's topology, when it absolutely can. Then the complaint morphs into, "well, it doesn't work very well (when Conform Retopo Mesh is used), even after multiple tutorials demonstrate that it in fact does work well. In 3DCoat you CAN use a low poly, UV mapped base mesh, to start sculpting with and keep it all the way through the sculpting project. You cannot do this in ZBrush, unless you try to go without using Dynamesh and Sculptris Pro. The truth is a user has to jump through a lot more workflow hoops than 3DCoat does. This is why I asked Andrew to add this Conform Retopo feature. You never have to worry about scrapping your original, UV mapped base mesh and building another one (ZRemesher/Auto Retopo) to replace it. In Zbrush you do. 

[AbnRanger]

On 8/27/2022 at 6:28 AM, Tieguaili3D said:

in the first video you can see the lumpy surface you get when subdividing a low poly mesh, that makes sculpting cleanly an issue, you can also see the random distortions in the "conformed" low poly when smoothing, and the fact that you can't control where the low poly goes when you're working on the high poly. Last part of the video you can see the absolute mess you end up with when using the decimate and reduce functions to get something "low res" to work on, and you can also see that even at 64x decimation the result is both less detailed and higher resolution than the proper low poly mesh (this is also why it's completely pointless trying to bring in a pre-subdivided mesh from another program).

In the second video you once again see how the auto-triangulation makes an unusable mess of your high poly when subdivided, and shows how if you don't want a distracting and obstructive hologram of your low poly on top of your work the low poly no longer conforms, so you're stuck working with the UV shells and wireframe obstructing your view.

 

Third video is just a comparison with zbrush, it's faster to start working since you can just import and get working rather than import your low poly as a retopo mesh, move into the sculpt room, copy your retopo mesh into the sculpt room, then start to work. It also shows how using a proper multires system allows for the export of displacement and normal maps from those subdivision levels.

 

Using 3d coat you have 2 otpions for "multires" sculpting and 0 options for effective multires sculpting:

1) you can import your low poly as a retopo mesh, go to sculpt room, copy your retopo mesh into the sculpt room, start working with a horrible mess caused by the decimation that doesn't even get you close to the detail preservation or resolution that you get from an actual low poly.

2) subdivide your low poly in blender, export, import your low poly to 3d coat as a retopo mesh, go to sculpt room and import your high poly as a sculpt mesh, start working with a horrible mess caused by the decimation that doesn't even get you close to the detail preservation or resolution that you get from an actual low poly.

 

using blender or zbrush you have 1 option for effective multires sculpting:

blender: import your low poly, add multires modifier, sculpt, generate normal and displacement maps from your subdivsion levels.

zbrush: import your low poly, sculpt, generate normal and displacement maps from subdivision levels.

 

@AbnRangerhopefully this makes it clear as to why we need proper subdivision levels and quad meshes rather than a hacky workaround.

"@AbnRangerhopefully this makes it clear as to why we need proper subdivision levels and quad meshes rather than a hacky workaround."

Let's revisit these examples, and discuss them one by one.

Video Example 1: 

The Sculpt Object (Proxy Mesh) is decimated way too much. The Proxy mesh was not designed to be used this way. Decimating that low is just to reduce the polycount to make the scene lighter. It's of no help whatsoever, if you want to make adjustments to the mesh. It's too extreme...like trying to use a Chain saw to sand a piece of furniture. Just decimating 4-8x was more than sufficient to provide the performance benefit you are seeking. Many times, you don't even need to step down to a lower level of resolution, to get good performance while using tools like Pose, Transform, or Move.

Video Example 2:

CONFORM RETOPO MESH is unchecked (disabled), therefore it cannot do its job. If you have to disable the Retopo Wireframe for some reason, you can do that in the list menu (Retopo Room) right above the Tool Panel, in the Tool Bar. That lets the user choose how the Retopo mesh is displayed. But, you cannot simply turn it off and expect it to work after you have made your sculpting edits. It simply will not work that way.

Video example 3: 

When using Large scale Edit tools like Move, Transform, or Pose, the performance is actually quite good also. Especially if you use the Proxy or new Multi-Level Resolution system, to step down a bit in resolution. The only "lag" one might notice is when Conform Retopo is used. It performs a secondary, follow-on calculation to snap the retopo mesh to the sculpt mesh. It's not a performance killer by any stretch of the imagination, as it is barely noticeable in most cases, as you can see in the sample video below. FWIW, the video is in real time. Not sped up, so viewers can see the actual performance when using these tools.

 

[Tieguaili3D]

On 6/7/2023 at 4:39 PM, AbnRanger said:

I understand your point, but I think you are missing something, still. One of the issues you raised was that one cannot import their base mesh (into 3DCoat's Sculpt Workspace) with UV's, and preserve the topology & UV's. I am saying that would also be the case in ZBrush when you use Dynamesh or Sculptris Pro. In both situations and in both applications, you are using that UV mapped mesh as the starting point....correct?

We clearly disagree about "Conform Retopo Mesh" and some of this is because you used it improperly in the video examples you showed on the first page of this thread. Please, instead of continuing to debate about its usefulness, watch the videos and then do some more testing. I think it really could change your views about it. 

 

no, in zbrush you always have your topology and UVs preserved on import until you destroy them with dynamesh, which should be how it works in coat as well, but as you seem to keep missing the point.....when you're working in subdivisions on a pre-made basemesh you do not use dynamesh/sculptris, there is no benefit to it in any way at that point in your workflow so saying "well dynamesh breaks your topology and UVs anyway so why would you ever bother having them?" is pointless.

[Elemeno]

8 minutes ago, Tieguaili3D said:

no, in zbrush you always have your topology and UVs preserved on import until you destroy them with dynamesh, which should be how it works in coat as well, but as you seem to keep missing the point.....when you're working in subdivisions on a pre-made basemesh you do not use dynamesh/sculptris, there is no benefit to it in any way at that point in your workflow so saying "well dynamesh breaks your topology and UVs anyway so why would you ever bother having them?" is pointless.

3dcoat uses triangulation while calculating the mesh ,

so whenever a stroke is made the mesh is recalculated to be evenly distributed

they said a year ago that quads were coming so i guess surface mode would preserve topology then and voxels be purely tris and how it is today ,

unfortunately though right now its not available

 

ive always used a high poly workflow anyways so its not really troubling me

 

i rarely even import uv model, divide sculpt bake down etc... i like having high poly models and baking elsewhere with those

[AbnRanger]

On 6/11/2023 at 3:13 PM, Tieguaili3D said:

no, in zbrush you always have your topology and UVs preserved on import until you destroy them with dynamesh, which should be how it works in coat as well, but as you seem to keep missing the point.....when you're working in subdivisions on a pre-made basemesh you do not use dynamesh/sculptris, there is no benefit to it in any way at that point in your workflow so saying "well dynamesh breaks your topology and UVs anyway so why would you ever bother having them?" is pointless.

I'm not missing your point, as I have repeatedly said from the beginning that a low poly, UV mapped mesh imported into ZB would lose it's UV's and topology anyway....IF/WHEN Dynamesh or Sculptris Pro were used (because they are Re-Meshing toolsets, after all). I never even implied that you lose it immediately on import. Where we differ is in the workflow once such a model is imported. Will it need significant deformation or proportional changes? If so, then Dynamesh or Sculptris Pro will probably get utilized in many if not most cases (to avoid nasty polygonal stretching issues). Using a base human(oid) model can and is often used to sculpt something very different. Why start from ZSpheres or a simple sphere, when you have a base mesh that gets you more than halfway there, and you have a tight deadline, just as the Flipped Normals guys state?

If a 3DCoat artist were to use this basic workflow (import a decent base mesh with UV's already applied), they would NOT need to Retopo and UV this model, as the author mentions Zbrush users would need to do, at this point in the video. I think this is a HUGE timesaver and Conform Retopo Mesh is a very underappreciated toolset. It lets the artist use whatever they want for the Sculpt Mesh, while preserving the UVs and Topology of the original. Nevertheless, Andrew implemented the Subdivision Level toolset that was asked for here, except for the Quad mesh. That may come, too, sooner or later.

[Carlosan]

 

[animk]

What I understand through watching zbrush videos and my 3d coat experience. 

In ZB it let artists retopo and UV half way in the sculpting process then use multires to add final detail while keeping the retopo and UV.
In 3d coat it only let artists to do retopo and UV after the final sculpt is done.

If minor adjustments are needed, such as tweaking proportions after, while keeping retopo and UV, in ZB it can tweak the lower res mesh, in 3d coat it uses conform Conform Retopo Mesh.  If substantial changes are needed, such as adding extra horns or cutting an hole, it requires rework of the retopo and UV in both ZB and 3d coat regardless which workflow is used.

The difference between these two workflows doesn't seem that huge.  It's most likely just a personal choice if the artists like to sculpt on multires or live clay...I personally love sculpting on live clay, sculpting on multires requires unnecessary subd level to the whole thing just for the details at certain areas (I feel this is quite restricted in blender but in ZB it's not an issue).  3d coat is not a 2.5 program, not sure if it can reach zb performance if quad multires is added.

So the main issue is retopo and UV, how we place this process in the workflow.  I wish in the near future AI era we can get human quality auto retopo and auto UV, so it doesn't matter how many times we change the final sculpts these can always be redone with a few clicks.

Edited July 23, 2023 by animk

[Elemeno]

2 hours ago, animk said:

What I understand through watching zbrush videos and my 3d coat experience. 

In ZB it let artists retopo and UV half way in the sculpting process then use multires to add final detail while keeping the retopo and UV.
In 3d coat it only let artists to do retopo and UV after the final sculpt is done.

If minor adjustments are needed, such as tweaking proportions after, while keeping retopo and UV, in ZB it can tweak the lower res mesh, in 3d coat it uses conform Conform Retopo Mesh.  If substantial changes are needed, such as adding extra horns or cutting an hole, it requires rework of the retopo and UV in both ZB and 3d coat regardless which workflow is used.

The difference between these two workflows doesn't seem that huge.  It's most likely just a personal choice if the artists like to sculpt on multires or live clay...I personally love sculpting on live clay, sculpting on multires requires unnecessary subd level to the whole thing just for the details at certain areas (I feel this is quite restricted in blender but in ZB it's not an issue).  3d coat is not a 2.5 program, not sure if it can reach zb performance if quad multires is added.

So the main issue is retopo and UV, how we place this process in the workflow.  I wish in the near future AI era we can get human quality auto retopo and auto UV, so it doesn't matter how many times we change the final sculpts these can always be redone with a few clicks.

HERES THE PROBLEM... you wouldnt need to change a final sculpt,

before the finalize the character or prop... its critiqued , pulled apart , i want this that way etc 

so if youre using 3dcoat then you retopo the finals sculpt bake etc done ,

 

with Zbrush , your retopo mesh can be sculpted onto it , it uses quads so the calculations are easier , meaning you can flicker between divisions

once you have finished the detailing etc , you can also bake your low divisions onto your high divisions but its just workflows.. one isnt better than the other , some have pros and cons like all things , there isnt one good workflow , its what works for you and catering yourself with the tools that work for you.

Edited July 23, 2023 by Elemeno

[AbnRanger]

5 hours ago, animk said:

What I understand through watching zbrush videos and my 3d coat experience. 

In ZB it let artists retopo and UV half way in the sculpting process then use multires to add final detail while keeping the retopo and UV.
In 3d coat it only let artists to do retopo and UV after the final sculpt is done.

If minor adjustments are needed, such as tweaking proportions after, while keeping retopo and UV, in ZB it can tweak the lower res mesh, in 3d coat it uses conform Conform Retopo Mesh.  If substantial changes are needed, such as adding extra horns or cutting an hole, it requires rework of the retopo and UV in both ZB and 3d coat regardless which workflow is used.

The difference between these two workflows doesn't seem that huge.  It's most likely just a personal choice if the artists like to sculpt on multires or live clay...I personally love sculpting on live clay, sculpting on multires requires unnecessary subd level to the whole thing just for the details at certain areas (I feel this is quite restricted in blender but in ZB it's not an issue).  3d coat is not a 2.5 program, not sure if it can reach zb performance if quad multires is added.

So the main issue is retopo and UV, how we place this process in the workflow.  I wish in the near future AI era we can get human quality auto retopo and auto UV, so it doesn't matter how many times we change the final sculpts these can always be redone with a few clicks.

You can retopo/auto-retopo at any stage in 3DCoat, so there is no hard rule that says you must do this at the end. It is just a common workflow. 

 

[Elemeno]

6 hours ago, animk said:

What I understand through watching zbrush videos and my 3d coat experience. 

In ZB it let artists retopo and UV half way in the sculpting process then use multires to add final detail while keeping the retopo and UV.
In 3d coat it only let artists to do retopo and UV after the final sculpt is done.

If minor adjustments are needed, such as tweaking proportions after, while keeping retopo and UV, in ZB it can tweak the lower res mesh, in 3d coat it uses conform Conform Retopo Mesh.  If substantial changes are needed, such as adding extra horns or cutting an hole, it requires rework of the retopo and UV in both ZB and 3d coat regardless which workflow is used.

The difference between these two workflows doesn't seem that huge.  It's most likely just a personal choice if the artists like to sculpt on multires or live clay...I personally love sculpting on live clay, sculpting on multires requires unnecessary subd level to the whole thing just for the details at certain areas (I feel this is quite restricted in blender but in ZB it's not an issue).  3d coat is not a 2.5 program, not sure if it can reach zb performance if quad multires is added.

So the main issue is retopo and UV, how we place this process in the workflow.  I wish in the near future AI era we can get human quality auto retopo and auto UV, so it doesn't matter how many times we change the final sculpts these can always be redone with a few clicks.

we will never ever get human quality auto retopology or human quality uv unwrapping .. unfortunately .. we have different uses for everything , EXO and Zremesh and autopo do a nice job of topology flow but isnt as good as manual ...

also as i said above .. changes to high poly sculpt are very rare after doing your retopo and uv unwrap etc ...

 

get your high poly right first time lol

[animk]

7 hours ago, AbnRanger said:

You can retopo/auto-retopo at any stage in 3DCoat, so there is no hard rule that says you must do this at the end. It is just a common workflow. 

 

 

 

If that's the case the final sculpt shouldn't be off too much from the retopo version, so why not just retopo when the final sculpt is done, doing retopo in the middle of sculpting process makes no much sense.  There's also no need for quad multires, because there's much more freedom sculpting on live clay...  The retopo is done at the final stage anyway.

[animk]

7 hours ago, Elemeno said:

we will never ever get human quality auto retopology or human quality uv unwrapping .. unfortunately .. we have different uses for everything , EXO and Zremesh and autopo do a nice job of topology flow but isnt as good as manual ...

also as i said above .. changes to high poly sculpt are very rare after doing your retopo and uv unwrap etc ...

 

get your high poly right first time lol

I don't know... those AI generated image really changed my perspective on the future tools.  It's not yet 100% automated for non-random works  but human input is drastically reduced.  It may come to 3d at some point.

If the final sculpt is rarely changed, why is the urge of needing quad multires, sculpting on multires kinda sucks, unless it can handle crazy poly counts like ZB, but still the freedom is limited once the multires is built.

Edited July 23, 2023 by animk

[AbnRanger]

12 hours ago, animk said:

If that's the case the final sculpt shouldn't be off too much from the retopo version, so why not just retopo when the final sculpt is done, doing retopo in the middle of sculpting process makes no much sense.  There's also no need for quad multires, because there's much more freedom sculpting on live clay...  The retopo is done at the final stage anyway.

It's a common workflow to wait until the end of a sculpt project, to retopo. However, the benefit of the "Conform Retopo" feature comes when a user has a base mesh already in their possession, from the start. For example, an experienced character artist probably has a number of low poly (quad) models, stored on their hard drive, to use as a starting point for many of their sculpting projects. They probably have a base male and female mesh, and another for a hero character, and another to use for humanoids/creatures, etc.

It's great to know how to do a sculpt from scratch, but when facing a tight deadline, why not start roughly halfway, instead? In 3DCoat, if you are happy with the topology of the base mesh, and UV's, then there is almost no retopology or UV editing to do (maybe a little tweaking or clean up). Just sculpting > bake > texture paint.

[Elemeno]

15 hours ago, animk said:

I don't know... those AI generated image really changed my perspective on the future tools.  It's not yet 100% automated for non-random works  but human input is drastically reduced.  It may come to 3d at some point.

If the final sculpt is rarely changed, why is the urge of needing quad multires, sculpting on multires kinda sucks, unless it can handle crazy poly counts like ZB, but still the freedom is limited once the multires is built.

only use for multires .. is if you start to finalise the sculpt and you realise that a part of the head is wrong , moving or smoothing wont be very effective at such a high resolution so you can drop down , move where you want the areas and then back to high detail without loosing detail ,

also the push for quads is something 3DC desperately needs .. although you can get a nice finish with the tris 3dc generates , sometimes(most) of the time tris that are elongated or narrow cause weird shading , having quads that follow a topology are more optimised and also bake better , if you notice even at 20+ mil tris .. the edges are still jaggedy , this shows through on bakes ..  calculation time of quads is alot faster . dont get me wrong i love voxels ... but for a smooth clear finish i need quads and i need catmull clark subdivions inside sculpt too..