I realize this wouldn't be optimal for detailed meshes, but it would certainly be useful for creating others.

Sometimes I have meshes I'd like to join together, only I'd like them to interact as if they were metas.

One Meta from Mesh Exterior?

Since metas look similar to a remesh modifier with low poly, couldn't the add-on just lower the poly count as much as necessary (and only use the object outline as a guide) before converting to meta?

Multiple Metas as One?

The answer to this question: "Are there any non-standard meta objects available?" says it's not possible with python programming but suggests there may be a work around if the shape is composed using many meatballs. I feel there must be some other reason a "convert to meta" add-on isn't available.


One indirect way to achieve this will be using a grid distributed particle system, with a meta-ball as the particle object.

Here are the important particle system settings for doing this: enter image description here

And here's the meta version vs. the original: enter image description here

You can also set the overall resolution via the metaball's resolution property.

Since these are all fairly simple PS properties, it wouldn't be difficult to automate them as a script or addon.

  • 1
    $\begingroup$ Thanks, this would attain the result i'm looking for in most meshes. $\endgroup$ – 360ueck Apr 30 '16 at 15:17
  • 1
    $\begingroup$ pretty excellent concept $\endgroup$ – patmo141 Sep 30 '17 at 2:21

Metas are objects that are defined by very simple mathematical formulas and converting complex objects to metas might not be possible, or would require very complex coding and processing resources.

For example, a Meta Ball has just few variables to calculate:

  • Location in space
  • Resolution
  • Radius

Comparing this to a default sphere that has 512 faces would result in a drastic increase of parameter values and calculation time.


As an alternative solution, you could try and use dupliverts or dupli-face to duplicate a standard metaball onto every element of your source mesh, which would probably give you something -vaguely- like what you want.

You might need to scale down the original a bit, and then play with the radius of the duplicated metaball.

You should also be aware that on anything but the simplest base mesh, this will bring your computer to its knees attempting to calculate that many metaballs, so it's probably worth simplifying, decimating or generally reducing the quantity of base vertices and faces before doing this :)

  • $\begingroup$ TLousky's answer will probably generate a more useful version than this one as it copes with volume a lot better - a property important to metaballs. $\endgroup$ – Ben Apr 29 '16 at 9:29
  • $\begingroup$ This is a useful approach, between TLousky's and yours I feel nearly all meshes could be made meta. $\endgroup$ – 360ueck Apr 30 '16 at 16:03

You could use magnet and snap on faces, that will get you a result like a metaball where it extrudes when you move an object within an object. There is a proper tweak to it where you have to bevel the edges of the object touching the faces and shrinkwrap it and use data transfer to the source object.


Having said that - Could you potentially get what you want by adding a subsurface modifier to your original meshes (possibly with quite a lot of creasing so you don't lose as much of the original shape as normal) and then joining them together and manually adding faces over the join?

Subsurfs with low-poly control cage meshes often look quite similar to metaballs, joining with a rather organic looking join.

  • $\begingroup$ ooh, a -1 already. Okay, why isn't this a useful suggestion? $\endgroup$ – Ben Apr 29 '16 at 7:35
  • $\begingroup$ Reread my question and the previous answer. This answer doesn't say anything new, and ignores part of my question. Not trying to be rude, but can't help being blunt about it. $\endgroup$ – 360ueck Apr 29 '16 at 7:41
  • $\begingroup$ Nothing in your question or any of the other answers have suggested using subdivision surfaces. Remesh is similar but it generates different results to applying a subdivision surface (at least when using the default catmull-clark surfaces). I therefore believed it was a novel suggestion. You're welcome to think it's a bad suggestion and that's fine, but I disagree that it doesn't say anything new. It doesn't ignore any of your question, but I freely admit it doesn't use metaballs at all, I did state that the -results- of using a subdivision surface often look similar to metaobjects. $\endgroup$ – Ben Apr 29 '16 at 9:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.