Jump to content
3DCoat Forums

I just cannot get good results from the Autopo


Tser
 Share

Recommended Posts

  • Advanced Member

I just cannot get good results from the Autopo, I use Zremesher in ZBrush and it gives very good results with the factory settings, but Autopo gives very bad results no matter what settings I use, I followed the manual to place guides but the mesh always turns out bad, what am I doing wrong? is there a trick to it?

T.

Link to comment
Share on other sites

  • Reputable Contributor

If you can, screen record the issues you are having and send a link to that + your work file to Andrew (support@3d-coat.com, of course) and see if he can't tweak it a bit. It's better than what it was, before he worked on it the last time, but I find it a bit hit and miss as well. Sometimes it works so well, I have to giggle a little bit. Other times it seems like a huge waste of time. I mentioned this to Andrew before. It's too fickle to be relied upon in a production environment. You could end up wasting half an hour or hour, just trying adjust this or that to get it to work, when using the Strokes tool would have gotten it done in a fraction of that time.

 

This is why I never use it on my own work projects, unless it's a more primitive shaped object, like a secondary part of a model (ie., Buttons, rings, belts, belt buckles, etc.). I was asking him about the ability to use the E-Panel selection modes, such as the circle, ellipse, rectangle, etc. To quickly create strokes with the Strokes tool...he said he would try to do that at some point, but also mentioned he had another type of auto-mesh generation idea. So, we'll have to wait and see. It seems ZBrush did leapfrog 3D Coat in terms of the Auto-Retopo toolset. Maybe Andrew will have a good answer, in turn.

Link to comment
Share on other sites

  • Reputable Contributor

In the mean time, I usually just try to create a clone of the object I want to retopo, and dumb it down, so to speak. I fill in any surface level holes, and smooth the object a bit. This gives it less to calculate and does help to improve the "Hit Ratio."

 

That's the point in this video:

 

And with this model, I had a bunch of trouble with the results, so I went back to the drawing board and smoothed the surface + simplified the strokes. and BAM, it worked. So, yeah...it's fickle as can be, but it can generate good results many times.

 

Link to comment
Share on other sites

  • Reputable Contributor

It's a pity really, I would like to do every thing in 3DCoat.

BTW: 3DC runs well in Centos 7.

T.

An approach you may want to take is to make 2 attempts (if the first one doesn't work). The first attempt, no guides. Just try to make a good guess as to the poly count, and go slightly above what you think it needs. Simply the surface details if possible, by choosing SMOOTH ALL a time or two. If it doesn't yield good results then try again with Strokes along major edges and contours...NO intersections....NO crossing guides over the symmetry plane (if Symmetry is enabled), and use the "Less is More" principle.

 

If after that 2nd attempt is unsatisfactory, don't waste anymore time with it. Either use the manual tools or ZBrush's Remesher. I always harp on the usefullness of the Strokes tool as an Auto-Retop tool. Heck, if one is going to apply guides, you're standing on 3rd base, as it were. The Strokes tool takes sketched topology and generates a mesh, just like Auto-Retopo does....except there is no guess work for the app. Wherever there is an intersection of 3 or 4 strokes, a polygon is created. If Andrew juiced it up with the quick shape selections (from the selection modes in the E-Panel) and some of the concepts/features from the Blender plugin, then there is much less need for Auto Retopo

 

Link to comment
Share on other sites

I think that HardSurface Retopology option is broke at last builds

 

Crease Angle dont follow model edge flow.

a 40 or 400 value dont make any difference to the final edgeflow

 

post-10142-0-20780000-1417810448_thumb.j

 

Crease Angle need to be more smart to be able to analyze the curvature of the model.

 

And talking about zremesher... adaptive-size 3DC_equivalent need a refinement upgrade too.

Link to comment
Share on other sites

  • Member

Hi Tser

Your post about autopo issues was splited here:

 

I just cannot get good results from the Autopo

 

Ty

Unfortunately current 3DC auto retopo fails miserably for snail shell shaped ionic column tops. Zbrush' retopo created flawless lowpoly mesh for this hipoly : you need a team like Verold’s QRemesher --> Zremesher so the team can solve specific problems. Manhours and engineering power of a dedicated coder team, who want to prove themselves, so they are eager, cannot be underestimated.

Edited by mercy
Link to comment
Share on other sites

  • Advanced Member

Even the Blender Strokes method would be a start, I don't expect 3DCoat to achieve the same results as zremesher, but it appears that Andrew has moved on to other things and just left Autopo as is, I think Re-topology is incredibly important if you want to get a decent, low poly mesh sculpt with its displacement map out of 3DCoat, this worked much better in previous versions. I just hope Andrew reads this and spends some time in this valuable area of the program.

T.

Link to comment
Share on other sites

  • Advanced Member

i have to agree, i think manual is the way to go. there need to be some improvements, What i would love is some semi-autopo. for example i make polygon strips of all the important loops. and then run autopo and it connects em all up along the surfaces as best it can. That would save time and give a tonne of control.

  • Like 1
Link to comment
Share on other sites

Is not an order that i get "keep it secret, dude!"

But -from my point of view- technology research must be more... confidential.

Usually competition try to punch first: "hey, look 3DC!, they are developing ____. Can we do it first?"

Tired to see this -community users that works for other companies searching this forum- I prefer to keep silence.

Sorry.

Link to comment
Share on other sites

  • Reputable Contributor

If you think about it, Pixologic came out with their own version of Auto-Retopology a few years after Andrew introduced it (Siggraph 2010). Then they hired Dr. Peters and from what I gather, they put him to work on improving it to where it could be more usable in a production setting. That took roughly 2yrs! Andrew worked to improve Auto-Retopo for roughly 1-2 months...while having to squash bugs and daily having people asking for new features. That's why I suggested to him to take his time....even if it took 3-6months.

 

We aren't used to that length of a delay from him....nor is he used to it. But, in the long run, it would be worth it, in my opinion. It would make 3D Coat a much more viable modeling toolset. Voxel Modeling = No piddling with verts/edges/polys during construction > Auto-Retopology = great looking Low-Poly result.

 

I kind of agree with Carlos on the tech info. On one hand, it gets people excited to see what's coming....like the PBR stuff. But, on the downside, it tips competing applications off and compels them to get started on an answer, themselves. I bet Pixologic and Mudbox is already considering how they can add PBR materials to their own internal toolset. If Andrew had kept it hush, then release it, that would put the competition back that much longer. But it's merely delaying the inevitable, I suppose. Nevertheless, months and months of sneak peeks and open beta testing allows competing software companies to release their version much closer to your own. So it can steal the thunder, in a manner of speaking, of software vendor's new tech.

 

A good example of that is Substance Painter. Them unveiling their new product, that could hurt companies like Pilgway....by making 3D Coat less relevant as a texture painting resource....got Andrew cooking his own PBR toolset BEFORE Substance ever officially released it. So the advanced warning does actually have an effect. It's not just paranoia.

Link to comment
Share on other sites

  • Member

I am also having Autopo problems in Version 4.1. Maybe you should try using 4.0 if you don't need any of 4.1's features. As far as I'm concerned, I'm still using 4.0 most of the time because of the Autopo problem. In 4.0, it works just fine. Maybe I have to rearrange some quads here and there, but other than that, 4.0's autopo gives very good results while still giving me the option to use a whole bunch of guiding lines to keep control over the topology process at the same time.


It's very nice to hear that retopo seems to be in development, though!


Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...