1. #31
    New Member User Gallery
    Join Date
    Apr 2012
    Posts
    1

    Default Great Stuff

    Really nice tutorials there!!!

    But I'm got a weird behavior with my Zbrush! When I SliceCurve my model for the first time and then try to mirror and weld, it's doesnt work!!!!

    Zbrush is just not keeping the sliced group! It's like it's erasing it?? Ok I could bypass it by and but it's not really efficient and not really accurate!

    Anyone got any idea why it's doing that??????

    Thank You!!!

  2. #32
    Senior Member User Gallery
    Join Date
    Apr 2003
    Location
    CGI Realms
    Posts
    161

    Default

    Quote Originally Posted by calyston00 View Post
    Really nice tutorials there!!!

    But I'm got a weird behavior with my Zbrush! When I SliceCurve my model for the first time and then try to mirror and weld, it's doesnt work!!!!

    Zbrush is just not keeping the sliced group! It's like it's erasing it?? Ok I could bypass it by and but it's not really efficient and not really accurate!

    Anyone got any idea why it's doing that??????

    Thank You!!!
    Welcome to zBrush

    When posting for technical help.
    you always need to post your specs.
    For example
    Windows vista quad 64 bit ,4 RAM
    GeForce 950 2 Ram
    Wacom intous5 touch medium.

    Always a good idea to have the latest updates.
    http://www.zbrushcentral.com/showthr...ate-2-is-Ready

  3. #33
    New Member User Gallery
    Join Date
    Aug 2012
    Posts
    4

    Default

    Can I watch the videos in Youtube?

  4. #34
    New Member User Gallery
    Join Date
    Mar 2004
    Location
    Michigan, USA
    Posts
    12

    Default

    Quote Originally Posted by RorrKonn View Post
    On my home PC I personally only consider DynaMesh for zBrush only.
    I personally would not use DynaMesh for C4D on my home PC.
    If I had a super computer I might think other wise.

    As you see in zBrush gallery there are mostly Characters.
    Not a lot of buildings or cars.

    You could do this for a building or car or any thing.
    One way you could model a Character for C4D.
    Model a low poly count 2,500 character mesh in C4D.
    If you want 100% quads,Divide once 100% quad.
    The low poly count 10,000 character mesh is for C4D rigs.
    Send to zBrush ,Divide to millions.Add Details.
    Save Vector Map for C4D.

    Now you have a millions poly count detailed Character but you only half to deal with the low poly count character mesh of 2.500 or 10,000.
    in C4D untill you render.

    With Vector Map's or Displacement Map's Topology is not a major issue with me personally.
    If you use zBrush Brushes for detail ,Best to use a Wacom tablet ,Wacom takes some skill.

    You half to make sure your app supports Vector Map's or Displacement Map's.
    Well this was exactly my point the tutorial author when modeling the radio specifically references the requirements for accurate game models and the proceeds to create an accurate model of a radio by booleaning out bits in a very similar way to how parts are created in mechanical cad packages except instead of nurbs it's in in super high density polygons.

    So in the end he does have a very accurate looking model and it's great to twirl it around in zbrush and be proud of how cool and accurate it looks but what the hell good is it if it can only be manipulated inside zbrush? If I can't get it back into Maya or Max or some other traditional 3D packages without my system bursting into flames then it serves little purpose!

    The point is he says he is building the parts for game modeling but does not go to the next step of showing how he would then get that high density poly model back into some low to medium poly model that could be used for a game. The final model would be to high poly for even a hero model in a high budget movie!

    So why not demonstrate the entire round trip workflow of something like maya -> zb dynamesh -> something to make the model usable in a game or even a hero model -> maya?

    I have attached images of a complex part I modeled in maye. Building the base shapes was EASY! I'd say less then 20 minutes of work. Then I spent two days combining them into a unified part with proper topology!

    I see a lot of potential in dynamesh to reduce the time to combine the base components and created the unified shape but what I don't see is how to get the end result from a dynamesh model to something I can bring back into maya???
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	01p_AO_FG_DIFF.jpg 
Views:	231 
Size:	48.2 KB 
ID:	332420   Click image for larger version. 

Name:	bike_part_exploded.jpg 
Views:	195 
Size:	243.7 KB 
ID:	332421  
    Last edited by ctbram; 11-16-12 at 03:38 AM.

  5. #35
    Senior Member User Gallery
    Join Date
    Dec 2007
    Posts
    827

    Default

    Quote Originally Posted by ctbram View Post
    ....
    So why not demonstrate the entire round trip workflow of something like maya -> zb dynamesh -> something to make the model usable in a game or even a hero model -> maya?
    .....
    I guess because it was a video showing how to use Dynamesh's new and improved boolean features.

  6. #36
    Senior Member User Gallery
    Join Date
    Nov 2004
    Location
    Texas
    Posts
    259

    Default

    I'm certainly no expert game modeler, but if we're talking about a static mesh (one that does not deform in some crazy way), couldn't you decimate the heck out of it, retain at least most of its accuracy and use it then as an asset? Or do you think it would still be too big?

    A little off-topic, but somewhat related (albeit going in the opposite direction), I have been experimenting with bringing in SolidWorks CAD models (via a free OBJ Exporter plugin for SW). They come in heavily triangulated, but by using Dynamesh (and sometimes QRemesher) I can get a quadrangular model with a well-distributed mesh. But of course while it's then too big for a game asset, it does provide a way to manipulate the mesh for presentation purposes.
    Last edited by kevphil; 11-16-12 at 11:40 AM.

  7. #37
    Senior Member User Gallery
    Join Date
    Apr 2003
    Location
    CGI Realms
    Posts
    161

    Default

    You can decimate some meshes work better then others.
    gamers also make a high polycount mesh texture it then retopology low game mesh.
    use the high polycount texture on the game mesh.

    or you can make a low polycount mesh in C4D send to Z raze polycount ,make game mesh texture that way.
    Last edited by RorrKonn; 11-16-12 at 02:00 PM.

  8. #38
    New Member User Gallery
    Join Date
    Mar 2004
    Location
    Michigan, USA
    Posts
    12

    Default

    Quote Originally Posted by kevphil View Post
    I'm certainly no expert game modeler, but if we're talking about a static mesh (one that does not deform in some crazy way), couldn't you decimate the heck out of it, retain at least most of its accuracy and use it then as an asset? Or do you think it would still be too big?

    A little off-topic, but somewhat related (albeit going in the opposite direction), I have been experimenting with bringing in SolidWorks CAD models (via a free OBJ Exporter plugin for SW). They come in heavily triangulated, but by using Dynamesh (and sometimes QRemesher) I can get a quadrangular model with a well-distributed mesh. But of course while it's then too big for a game asset, it does provide a way to manipulate the mesh for presentation purposes.

    Yes I could come up with about a dozens guesses on how to complete the full maya to zb and back to maya workflow but hell if someone is already doing it then why not tell us how even if you can't show the complete workflow explain how you get the resulting dynamesh out of zb and back into something that cannot deal with 500 million polygons!

    And in most cases the answer is no you cannot decimate your way back to something you can use back in max or maya because all the edges will start to get tesselated and the model will generally start to look like crap (especially with hard surface things) by the time you get it to a workable poly count. Take that bike part there is no way you could combine that and make a high fidelity dynamesh model and then decimate it back down to something usable in maya that would be managable without some kind of retopologizing to create proper edge flow.

    and I was simply asking that the author possibly explain how he goes about the retopologizing, especially for complex hard surface objects.

    That bike fork brace was a bloody nightmare to retopologize in maya. I continually see tutorial and howto vids that show how to create the base shapes and it take practically no effort at all to boolean union them together regardless if I do it in maya or bring it into zb. All the real bloody work is in the step no one ever bothers to demonstrate - What workflow do they use to retopologize? Once again I am taking about HARD SURFACE objects, I have seen dozens of organics examples but there are some unique requirements on hard surface stuff like that radio or the bike part I modeled.
    Last edited by ctbram; 11-16-12 at 05:59 PM.

  9. #39
    Senior Member User Gallery
    Join Date
    Apr 2003
    Location
    CGI Realms
    Posts
    161

    Default

    There's different work flows.

    I've been in CGI since 1998.
    If I want a mesh that will work well in any app from C4D to D/S Pro.
    My work flow goes
    Modeled in C4D.
    Mapped in Blender.
    Textured in zBrush.Some of my textures for lightbox I make in gimp.
    Displacement maps or Vector maps in zBrush.

    I've never use retopology.

    -----------------------------------------------------------------------------------------
    If your just brain storming in zBrush end up with a 4 million polycount mesh you like.
    then you would need retopology.

    If you work for blizzard and they hired another Artist to make a Frankenstein.
    and they send you there Frankenstein they made in zBrush.
    then you would need retopology.

  10. #40
    New Member User Gallery
    Join Date
    Mar 2004
    Location
    Michigan, USA
    Posts
    12

    Default

    Quote Originally Posted by RorrKonn View Post
    There's different work flows.

    I've been in CGI since 1998.
    If I want a mesh that will work well in any app from C4D to D/S Pro.
    My work flow goes
    Modeled in C4D.
    Mapped in Blender.
    Textured in zBrush.Some of my textures for lightbox I make in gimp.
    Displacement maps or Vector maps in zBrush.

    I've never use retopology.

    -----------------------------------------------------------------------------------------
    If your just brain storming in zBrush end up with a 4 million polycount mesh you like.
    then you would need retopology.

    If you work for blizzard and they hired another Artist to make a Frankenstein.
    and they send you there Frankenstein they made in zBrush.
    then you would need retopology.
    But with dynamesh you end up with 100+ million poly models and it cannot simply be decimated. The edge flow is generally terrible if you try to decimate it. Take the radio example, I could retopo that that in zb or 3dcoat. That bike fork brace however, I would struggle with retopologizing even in 3dcoat. So I would like to know what work flow people that create parts of that complexity use.
    Last edited by ctbram; 11-17-12 at 12:46 AM.

  11. #41
    Senior Member User Gallery
    Join Date
    Apr 2003
    Location
    CGI Realms
    Posts
    161

    Default

    In your wire frame.jpg of ya bike parts.where the cylinder polycount is 16 around the diameter.
    sub divide. cylinder polycount 64 around the diameter.
    C4D has a subdividing weights for sharp edges.
    I would think Maya n Max would have subdividing weights also.
    tweak mesh .
    freeze and it's done.

    with precision mechanical meshes for Maya ,C4D a car final polycount can be 100,000 to 200,000.

  12. #42
    Senior Member User Gallery
    Join Date
    Apr 2003
    Location
    CGI Realms
    Posts
    161

    Default

    http://www.turbosquid.com/
    has all kind of meshes some have wire frames so you can see how they made there's.

  13. #43
    New Member User Gallery
    Join Date
    Mar 2004
    Location
    Michigan, USA
    Posts
    12

    Default

    Quote Originally Posted by RorrKonn View Post
    In your wire frame.jpg of ya bike parts.where the cylinder polycount is 16 around the diameter.
    sub divide. cylinder polycount 64 around the diameter.
    C4D has a subdividing weights for sharp edges.
    I would think Maya n Max would have subdividing weights also.
    tweak mesh .
    freeze and it's done.

    with precision mechanical meshes for Maya ,C4D a car final polycount can be 100,000 to 200,000.

    yeah I am fine with 100's of thousands for precision hard surface stuff but not 100's of millions and that is what you get out of dynamesh.

    The image that shows the exploded view has the base shapes that make up the part. That does not show the final number of subd's I went with. I go through a couple steps between that and the final unified part. I will add support edges, and then subdivide the individual base parts and then use the combined model of those parts as a 3D reference to retopologize the unified mesh over the top of. This was an all maya workflow. I am just looking for a more efficient workflow. Particularly for the retopolozing step.

    Also the final part shown has been unified into a single part with proper edge flow and is all quads. I do not know what it is about other packages but maya does not play well with ngons so simply booleaning things together does not produce a usable final model. I have attached an image of the retopologized bike parts wireframe.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	wheelBrace_14_wf.jpg 
Views:	97 
Size:	97.8 KB 
ID:	332553  

  14. #44
    Senior Member User Gallery
    Join Date
    Apr 2003
    Location
    CGI Realms
    Posts
    161

    Default

    It takes to long and drives you nuts figuring out how to model 100% quad character.
    So I model a character about 5000 polycount with a few nGones n tri's ,subdivide once & freeze makes it 100% quad.
    I use that trick for any complicated mesh.


    Don't know about LW 11 but Lightwave 7.5c didn't have nGones so I would model in the unstable TrueSpace.
    Guess since I started in TrueSpace ,I wont model with out nGones.
    The stable C4D works a lot like TrueSpace probably why I click with C4D.

    Those that started in app's that don't have nGones model different.
    App's that don't have nGones work flow would extrude edges a lot and weld vertices together.

    I take it your using zBrush re topology for a good topology flow.

    If you import ngone in zbrush it will make a quad n tri.
    If you sub divide a ngone in C4D 9 it will make 5 quads.
    The fastest work flow is a modeler that suports ngones and subdivides them in to 5 quads.

    Never seen any app do Boolean all that well except zBrush.

Page 3 of 3 FirstFirst 123

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •