Character Can Be Seen Through 3D Objects

[NaturalGlitch]

I have a scene with a 3D low-poly town (as one object), and I'm trying to have a character walk down the sidewalk, but she's showing through the objects no matter where she is.
Any suggestions?

[Greenlaw]

Just a guess, but try enabling Sort By True Distance in the parent group. This should place items in their actual z-distance rather than by their layer stacking order.

Hope this helps.

[NaturalGlitch]

Is there a way to do that with this?

'Cause it really is just one object.

[hayasidist]

yes - as Greenlaw has said, but be aware that the layers will be sorted by true depth in the group - IOW the stacking order will change to achieve the result.

However, I **think** moho looks at (overall / average??) Layer z (and not the item / sub-item / voxel in layer z) [but it would be great if I can be shown to be wrong on this point !!]- so occlusions might not work as hoped if there's a significant spread of z depths in one or more layers.

(oh and I'm the UK so I can't see your imgur upload!)

[NaturalGlitch]

I see what the problem was; I never grouped them together in the first place to Sort By True Distance.
I've done so, but she's still in front of everything, and I think I know why.
The town is a single object, and since Moho isn't true 3D software, it treats it as 2D.
So if I want her to be behind objects like a car or a tree, I would have to break apart the town in Blender.
Of course, I could animate her in Moho and bring her to Blender, where she can be behind the objects just fine, but I have issues with the .glTF exporter, where, after clearing her alpha channel, she loses her shadow. Sometimes she's just straight up pink.
But that's a whole different thread.
Thanks anyway. This was entirely user error.

[Greenlaw]

Yes, I'm not terribly surprised; Moho has some basic 3D support, but 3D is not the program's focus or strength. I normally do my 3D object setups for Moho in a dedicated 3D animation or compositing program, which I highly recommend.

glTF should work well in Blender, but if not, I suggest rendering the footage from Moho and mapping it to a plane in Blender. This is how I did it before Moho got glTF support, but using LightWave or Maya as my 3D program. This approach should work just as well in Blender.

The downside is that you will not have bones for physical interaction and will need to track the footage. But if all you want is 3D lighting, bones or tracking might not be necessary.

FWIW, I've dabbled with Moho glTF in Blender and got pretty decent results. You might want to download the Sidequest (Moho glTF) content from the Lost Marble website to see how the files are configured.

https://moho.lostmarble.com/pages/addons

Sidenote: I'm testing Moho glTF imports in the public beta version of LightWave 2026, but currently it's not working as well as it should (doesn't import bones; only meshes with animation applied, but incorrectly). I'm working with the LW developers to see if we can improve it.

[hayasidist]

...
So if I want her to be behind objects like a car or a tree, I would have to break apart the town in Blender.
...

have you considered masking? I've done a simple test here with a largely "full" 3d background where there are a few objects that will need to be in front of the character -


group (mask - reveal all)
> character - mask this layer
> mask - subtract from mask
> 3d bg - don't mask this layer

If necessary you can have more than one masking layer - or simply multiple shapes on the one layer; and potentially (depends on how much camera movement you have) animate the points in the mask to match the action. Obviously - you only need the mask active when the character is "in range" of the thing that they need to go behind, so it ought not to need a complex masking animation.

[Greenlaw]

It's been many years since I last used Moho's built-in 3D features, so I made a simple test to see how it worked...



I followed the depth instructions described above, and also imported 3D objects as separate items in Moho.

Technically speaking, this works, but, TBH, I wouldn't bother. For something like this, it's easier to do what I suggested earlier: map your Moho renders to planes in a 3D animation or compositing program. It will look significantly better, too.

If you're using Moho 14.4 and Blender, glTF is the best option when you need bones and physical motion data. But when you don't need that, mapping renders to cards works fine.

[NaturalGlitch]

That looks great!

The issue I was having was that it was one gigantic object—the houses, cars, trees, and so on—and there is no way for me to break it apart. There is no way to import pieces/ungroup because there are no pieces—just the whole town as one object.

The masking in Moho might be a good idea, because Blender and Unreal are kicking my teeth in with how complicated their lighting is, and getting my animations to work in there is taking me a very long time.

I am sad.

[Greenlaw]

In a 3D program, a depth pass can be used to create a z-space for your character. This is something I do routinely when compositing 3D renders.

We know Moho supports depth to create DOF effects. I wish Moho also supported .EXR so that data like this can be embedded and used by compositors.

Even without .EXR support, it would be helpful to get a depth pass rendered as 16.bit PNGs. Additionally, normals (for lighting) and motion vector passes (for motion blur) would be helpful. If Moho Styles could be translated to object/material IDs, I could use this data to mask anything in the render individually.

If all of these channels could be embedded in a Moho .EXR render, I'd be in heaven.