28 replies · posted

Collab2021 - WIP - splat21

This is my thread documenting my progress in the Collab2021 project.

  • splat21 Just a quick reminder. We still need a name for our team? What's your vote?

    1) Shabby Chic 

    2) All Access(ories) 

    3) Shabby Fresnel Chic 

    4) The Stan Dalones 

  • duerer, how about option 2 - All Access(ories).  

  • === Week 1 - Modeling

  • Blocked off to match the artwork.

  • Built a frame for the cart.

  • Model before adding details to the wood.  Per Vadim's guidance that the shaft should indeed be longer, have added some length to hook up a decent horse.

  • theluthier 

    Homework Submission - Week 1 - Cart

    • splat21  Don't forget, you also need to @ Kent ( theluthier ) as well so that he can find your homework. Looks great by the way.

    • splat21 Great work 👍! As far as the shafts are concerned, Vadim hharbinger_ua has it said here:

      You can make the shafts a bit longer as it is indeed for a horse. But not too long as it might ruin the overall balance of the silhouette. I would tilt at least the backside of the cart a little bit (as did for example  vincav81 here). Some hard edges (for example at the wheel rims) could be slightly beveveled (but only on segment!) so that it doesn't look that razor-sharp. Details like bolts are something that you can model in a more detailed version an then bake into a texture for the low-poly-model. I think that's what Kent theluthier also has in mind when it comes to such smaller parts.  

      splat21 I'm looking  forward to your further progresses with the cart 😀!


    • crew

      Wonderful progress splat21! You've easily earned full points 👍

      I second the feedback from duerer. Indeed small details like rivets will have a place in the normal map bake.

      My only stylistic note is that currently there's a bit too much precision for the aesthetic of the art. With olde-timey fatnasty straight lines are out; imperfect tapers are in. You could introduce this with lattices for a non-destructive implementation, same way that mona-loren tapered the glass portion of the lantern I worked on during the stream.

    • splat21  Congratulations to full points 😀! A discovery by vincav81 which has been confirmed byVadim  hharbinger_ua is that there's a slight curvature on the upper edge of the cart box backside. See here (orange frame).

    • crew


      theluthier take a break , your not in a fantasy world where R 'n' R is not required. 😴

    • adrian2301, dude - he just gave me full points.  He can be as fat & nasty as he wants!  😀

  • The cart placed in the artwork.  The downward angle will change as the shaft length has been increased.  The model was built with the cart upright.  For example, the shaft is perfectly horizontal.  The plane is used as a "Jig" for placement of the cart.  Placing the 3D cursor in the center of the right wheel, and change the view to Shift-Ctrl-Num3, rotate (pivoting on the 3D cursor) the cart until the shaft breaks the terrain (in this case, about half of the shaft).  Then Shift-Ctrl-Num7 for a top view and rotate 12 deg.  

    • splat21 You can make the shafts longer as Vadim harbinger_ua wrote ( linked in my previous post directly above). This results in a different tilt angle of the cart in relation to ground as long as we assume that the ground in that place is even or we add some unevenness to the ground there as vincav81 has thought about and illustrated here.

  • I just want to add a few comments to how the cart was constructed.  Below are screen shots showing the cart without mirror and array modifiers.  I wanted the boards in the very back to be unique, so I did not mirror them.  

    For detailing, I looked upon the Treasure Chest project.  The treasure chest was probably buried in sand, exposed to a lot of salt elements, and thrown around on a ship deck.  I view that this wagon does not need to be as beat up as that.  Does it have wear?  Absolutely, but I tried to give it some character while hopefully not putting in too much in cracks and dings.

    I also did not put in hardware, such as bolts, screws, etc.  Two reasons for this:  trying to keep the cart as low poly as I can and trying to stay true to the original concept art.  Having said that, I did model a guide to keep the axel in place underneath the cart.  No one will probably see it, but it struck me that it would be an incomplete construction without that.

    For some boards, I would detail the first, copy it and rotate it on two axis and use it as another board.  So, the second board would be the same but on its opposite side and upside down.  Just trying to minimize my work.

    • Thank you, splat21 for this insight into the construction process of your cart😀! I also like that you think about how such a  cart is working as you demonstration with the mostly hidden axel under the cart 👍!

  • Your cart is looking great. I like the details you’ve put into the wooden slats and the construction. 👍🏻

  • crew

    Any progress on your cart splat21? I know you're plenty busy with the Blog so no worries if this has fallen behind.

  • theluthier

    Homework Submission - Week 2 - Cart  

    I have made some changes to my model since the first homework submission:  revised all assets to remove accents (cracks, etc.) and reorganized to manage lo, med and hi poly versions of objects.  For example, a lo-poly board would be the modified cube with a one-segment bevel around its edges.  More complex shapes have been simplified.  Med-poly adds triangle geometry and holding edges in preparation for the multiresolution modifier.  Finally, the hi-poly is what is sculpted.  




  • And...the sculpted cart:

  • crew

    Your sculpt is looking great splat21! It should all bake down quite well. Can't wait to see it textured 🤩

    Keep up the good work, sir!

  • theluthier, I have been wondering about something.  I felt that in my sculpting I did not have enough mesh.  The Multiresolution was typically at a level 5 on the boards.  But, some of my cuts were not very clean due to the density of the underlying mesh.  Can you please tell me:

    1) Review my Med-Poly board above.  Should it have more geometry (that is, before using the Multiresolution modifier)?

    2) In general, I am trying to get my head around triangles vs. quads in the Med-Poly board.  I believe from your instruction, you advocated using triangles.  Should every feature be a triangle?  In other words, as we prepare for the Multires, should we convert quads to triangles?



    • splat21 I encountered the same problem described by you under the first point. In some areas I could easily model cracks whereas in other areas I encountered the problem that I didn't have enough geometry in order to model the cracks or other finer details in the direction and with the shape that I intended. Concerning your second point above I remember Kent theluthier simplyfying mmonaloren 's low-poly lantern in one of the livestreams creating a few triangles in order to save vertices. So, triangles are fine with low-poly objects, especially if we don't need to deform them.

    • crew

      splat21 Your questions are good in that there is a bit of a science to preparing a mesh for ideal sculpting, for which triangles are NOT the ideal; quads are ideal for multires sculpting. Triangles are ideal for an optimal real-time (game-ready) model.

      So why did I say it was ok to multires on a triangulated base mesh? Because when a project is low on time you have to brute force your way to a good-enough result! In your case your triangulated mesh is great for a lo-poly model but not great for multires:

      Even though subdivision mathematically will turn triangles into quads, the stretching of the geometry is present through each subdivision level:

      See how it's all quads but their stretched toward the pole in the middle of the board? This means sculpting behavior will be stretched accordingly.

      What's the solution? Another subdivision level! And another if your PC can handle it. In other words the quick and dirty approach is to simply increase your sculpt's poly count until the stretching problem is not visible. Of course this is not the most efficient way to work.

      The most efficient, clean way to work would be duplicating your lo poly before sculpting, converting it to even quads and add multires to that.

      Note that this is ideal prep for multires sculpting but in my opinion it's not ideal for baking textures. This way you have two separate models for baking which can work out fine. But I think the best baking comes from multires baking. I just think the math is more direct. So there's always pros and cons to each method.

      Hopefully that clarifies your questions about mesh structure for multires sculpting.