AUTOPO | Save Density (Modulator) Map, and other suggestions

[SYN7AX3RR0R]

Hey Andrew,

One thing I've been missing from the very first release of the Auto reTopo was the option to save the Density (Modulator) Map. It's crucial that we can save strokes, but it's quite a trouble not being able to save the density map. The problem gets even worse that each time you save the file you lose the density map, this means that even without closing the file during one single session I might have to repaint the map a few times. A big problem is it's hard to get back to a project after some time has passed and get very similar results as you forget how you painted the map.

These are a couple of things that would speed up my Autopo a lot:

- #1 Save Density (Modulator) Map

- Option to Clear the Map

- Option to Flood Value (if I vary the Density Map a lot I noticed that my Autopo turns out much better if I paint the whole object with a low value)

- Better Subtraction (Inverse) Paint

I'm aware that this next one might be quite a bit harder to implement and that is the reason I'm not including it above. For me it takes a number of tries to get the mesh density (Approximate Polycount) right. It is very good to have the Approximate Polycount limit, but unfortunately it does not help much in determining the denstity of a mesh as each mesh has a very different surface size.

I believe that it would be much better to have an alternate option (the 2 would not be able to work together) where you are asked to draw a quad polygon on the surface of the voxel (maybe using the Pionts/Faces tool). This polygon would be the guide polygon sizes on the zero level (unpainted) density.

Thanks for all the hard work and sweet tools

[SYN7AX3RR0R]

...while I'm at it.

This is another aspect that could benefit a lot from some improvement, I have experienced it from the first public beta.

Basically the guides do a good job of forcing edge flow, where they fail is if it's a closed guide. A closed guide should not just influence edge flow, it should force a closed edge loop. This would avoid creating edge spirals thus make the autopo meshes coming from 3DC much more production valuable.

Here's a couple of examples of what I mean.

In this image you can clearly see the spiral edge flow, and there is actually no need for this, closed edge loops would do the job ... it's 'just' about forcing one to be created and letting the algorythm work arround it.

This is the wrist guide on that Autopo:

Another issue that I experience continually on this model is that the mesh of the fingers bridges/fuses near the root of the fingers. This happens even if I draw a closed guide near the root. I believe that forcing closed edge loops by closed guides would fix these kinds of issues as well. You can see the distance of the fingers in the inset img.

This is a personal project I'm working on, but I've been dealing with all these issues in production since the first public auto reTopo release.

[digman]

Very good points made in both post...

[AbnRanger]

I agree on the issue with spiral edgeloops. I think it kills Autopo's usefulness. I rarely use the tool because of this very problem. It doesn't look messy, but trying to clean this up is a nightmare. The algorithm should close loops by default, regardless of whether you apply closed loop guides or not.

[Andrew Shpagin]

I agree with most points but spiral-like topology is really hard to avoid.

[Psmith]

In my video series for beginners, "3D-Coat From Scratch", I use a very elementary humanoid shape as a test bed for AUTOPO. In that particular video, I use both guides and density shading - and the the resulting mesh has no spiraling edges.

Non-Spiraling Guide Structure

In your realistic arm example, I don't think it is the closed guide that is the problem, but rather the spirally oriented guides that follow the muscles of the forearm, and the lack of extended guides from the fingers.

In my example, I avoided this kind of spiraling by making guides that directly parallel the edges of the arm, itself. If you try this on the model in question, I'll bet you get none of that kind of spiraling.

Also, by extending the longitudinal guides for the fingers, all the way up, through the hand, (just like the small hand bones), you can eliminate the spiraling that occurs below the closed guide at the wrist.

Give it a try.

Greg Smith

[AbnRanger]

I agree with most points but spiral-like topology is really hard to avoid.

I understand, but if a person has to "skin" (or smooth bind in Maya) a character with these spirals, they are going to "lay hands" on the modeler who gave it to them...and I don't mean in order to bless them, either.

[Andrew Shpagin]

I understand, but if a person has to "skin" (or smooth bind in Maya) a character with these spirals, they are going to "lay hands" on the modeler who gave it to them...and I don't mean in order to bless them, either.

Btw, lowering quads density decreases probability of spirals. Sometimes it is worth to make lower density at the beginning and then subdivide model afterwards.

[SYN7AX3RR0R]

Psmith | the issue here is that i don't want a grid like mesh, i actually want the edge-flow to support muscles. when i take this mesh to ZB or MB to transfer the details from the original sculptris sculpt i would need quite a bit more polys on the grid mesh to work with it as i would with the kind of mesh i get. that pretty much defeats the reason i used autopo in the first place.

Andrew Shpagin | thanks on the tip, i've noticed that subconsciously i believe. one of the major issues with a low poly mesh is that guides that are placed for a detailed region (like the ones for the eye loops on the example of the same mesh) are not going to be followed. this becomes a major issue for the same reason as descibed above.

i understand what you guys are saying and it is true for certain situations. unfortunately the situations where i've used autopo are quite different. i usually atopo a hirez mesh to be able to get more detail from the sculpt (because the initial sculpt did not have a proper base mesh or is either from 3dC / sculptris), also in most cases i will avoid using displacement maps and rather go with a mid-rez mesh and normal maps. i use autopo quite a bit, and i'm just saying ... i believe that this would greatly improve the tool for production use.

i'm not bashing the tool or such, on the contrary, it's a piece of art, a very good tool as is. doesn't mean it shouldn't be better.

[quano]

the option to save the Density (Modulator) Map

Sorry for bumping such an old thread, but was this feature ever implemented? I haven't been able to find a way to save the modulator map, and it's getting tedious to redraw it all the time.