-eg: 1: say you want a tunnel: just punch until it reaches the other side.
2: say you want a bridge: make a hill, punch a hole to the other side
-undoing that could be as simple as colapsing a donut. (pinching off a side of the donut and smoothing out the remaining nubs ) -pinch could be the undo (how's that, no need to make something to undo holes
-of course once you have what you proposed, and hopefully the crease thing I can pretty much forget poly by poly alterations since your poly reduction's a dream as is
monkey: I tried playing around a bit with your idea. Unfortunately I think it will be a little bit difficult to make it work as expected and have much real use. As you already figured out, many effects would only work well if you could manipulate their center point. In fact, I think only draw and smooth would work without that kind of control.
On the flip side, this is essentially the same thing as sculpting with textured brushes, so I basically implemented that functionality without realizing it. There are a few kinks related to how the surface angle is measured, but I should be able to solve it somehow. Results are weird but cool with the current code - check it:
This is not really what I planned to do this week, so I will refine it later at some point.
It's awesome that you even looked at my crazy idea, and so quickly! The last thing I want to do is distract you from the main task though. Between textured brushes and bumpmap painting, really looking forward to alpha 4.
Wonderful program, DrPetter. Do you realize that you've essentially made the ultimate easy-to-use 3D sculpting application? Even in alpha, it is already useful - and intuitive. I am reminded of ArtRage, which is also intuitive and deceptively simple, but for painting, not sculpting.
Perhaps you could adopt ArtRage's business model with it, which is make the application's interface minimal, its tools intuitive, and make the price very low, $20 to $40. I am sure it would sell very well.
Anyway, after two days of playing with it, I have a few suggestions.
One: it'd be great for the geometry to meld if brought close to or through itself. I.e. use inverse Draw tool long enough, and you poke an actual hole through the model. Pull out a couple tentacles with Inflate, and bring them together to join them into a teacup handle. This would be a great thing to extend the modeling range, and still work in the same intuitive fashion as Scuptris does now.
Two: can we have multiple objects per model, and interaction between them? There are things that are made much simpler by that. For example, scuplting an eye by placing a sphere into the model and then layering the eyelids on top of it is much easier than trying to sculpt a sphere with eyelids by hand, and produces a more animatable model too. Multiple objects could have their materials set to different maps for multicolor models. The best way to have it, though, would include interaction between the geometry - like literally pressing a sphere into the "clay" to leave a dent, or a cylinder through it to get a hole. You'd have what amounts to boolean operations with the same intuitive, hands-on handling. And in inverse mode, moving a second objects through the first one's surface would merge them into one object. Detailing things would be a breeze.
Three: with multiple objects, perhaps layering support would be nice? So I could sculpt a skull, and add soft tissue on top of that in a new layer. Then I could turn off the skull display, or edit the skull to have the soft tissue follow those changes - that could be great for some things, though probably a bit challenging to program. Even without that, layers would be good for turning visibility of the model's parts on and off.
Also, the Reduce Detail tool leaves little "knobs" of tesselation that just can't be made to go away, but you probably know about it already. Still, this is what makes the tool unusable, so it ought to be fixed.
This makes four: Actually, to reduce the ever-increasing polygon count, the Smooth and Flatten tools could just do it automatically. Flattening or smoothing out a patch that spans many triangles should have the little triangles merge back together when the angle between their planes crosses a certain low threshold. Right now they don't seem to work that way, and they could. That would eliminate much of the need for the Reduce Detail tool altogether.
Above all, please keep it this simple and intuitive. This accessibility is your killer feature. I can't wait till the next release!
arenhaus: Thanks. I replied to some of these questions just a moment ago in the Alpha 3 thread, so go check that out.
Interactions/collisions between meshes is definitely challenging to implement, and it's not critical enough that I will try in the near future. It would take me a lot of time, wouldn't be guaranteed to even work, and it would not make the application intensely more useful. Actually it might get in the way sometimes. I've already had people say that they like how it doesn't merge parts of the mesh when they overlap, like 3DCoat(?) does.
Here's a video of automatic detail reduction that I experimented with in December: www.youtube.com/watch?v=Svd8wQTdgY4 It was annoying to have it enabled for all tools, since you couldn't put down a very weak brush stroke and keep all the detail of it. It might be more acceptable for smooth and flatten though, since they work towards flattening and reducing detail anyway. The "knob" effect is actually a remnant from an old limitation I put in to circumvent a bug with detail reduction that I think has been fixed since then. I should be able to remove that.
Too bad about the mesh merging and tunneling through the model - I don't think it would "not make the application intensely more useful". It would make the possible range of model topology virtually infinite - that's definitely useful in my book. But you're the boss...
Hmmm...I was gonna jump in heroically and claim that poking holes into skulptures was not necessary, but come to think of it... there are a fair amount of situations when this could become really interesting. I'd - with good conscience- say "one day in the not too distant future this is going to be something to explore".
The reality is, currently it rather joins the ranks of zbrush and what ever that other thing was again, refining geometries. Although Zbrush offers the zspheres, which allow to create any kind geometry it represents its own kind of modeling section. I can see our genius here find a way to integrate that into the main modeling mode and it'll be a day of great recognition. It does not sound like a trivial thing in any way to me and I wouldn't be surprised if plenty of people would bow down to a success!
(tip to Tomas: Some wishes appear crazy for the present, but it's always best to leave a benefit of a doubt to the requester! One of us at pmg once claimed SubSurfaceScattering was irrelevant, replying to a users inquiry. That wasn't just foolish, but we also ended up implementing it for all the right reasons. Might I use this opportunity to boast a bit about my initial invention of volumetric translucency a year before SSS was even branded?!)
I was comparing the idea of automatic mesh self-interaction against what was discussed earlier in this thread about boolean operations between different objects. They both provide the same "enabling" functionality, i.e. make arbitrarily complex topology that the current system doesn't allow, but boolean operations would be a lot more obvious and attainable in terms of implementation and usage. I think it would also feel more natural, if the object manipulation can be made comfortable.
It's safe to assume that nothing I say will mean "no, definitely never", since I'm only planning what's feasible from my current position and maybe two months ahead. I can't speculate too much on "distant future" features since it would disrupt my focus and prevent me from making rapid progress. I'd rather say "no, it probably won't happen" and then avoid thinking about it until a point where it might make sense in the context of what else is there. I don't want to lay out a path where I'd need to work for 3 years to implement all the things I "really want" to have in it.
As a closing note, I'll just repeat that none of these things are "top priority", since I mainly want to make an application where you can sculpt and paint an organic model like a creature, and almost all examples of such models can be built from a sphere without bridges and holes.
You're asking for sugar, and sugar comes after the main meal... unless the sugar is very quick to implement
Personally I'm just waiting for the multithreaded release and, naturally, the bumpmapping.
This is, however, Feature Request Thread...so I'm just pulling something out of the hat for this post: - Material painting (although I'm sure it's been requested) - Height setting for FLATTEN (relative to referenced average) - Strength of GRAB to act like "nutch" in zbrush, simply letting go based on distance to center
- Continuous line to avoid stippling during drawing, particularely noticable at crease, actually. Might be a slowing feature, but a very helpful one! - Gravity, dropping (or drooping, actually) "skin" to bunch up or some fancy algorythm to relax surface tension while maintaining volume or the likes. (could be fun!?)
I have to say that the reason I find myself using Sculptris for sculpting heads in particular is because of the non-intereaction between the surface of the object. In 3d-Coat, sculpting a human body or a creature's body is very fun and intuitive, however, in my personal opinion, the interaction of the voxel surface makes sculpting eyes, lips, and noses, very involved and technical. Whereas here in sculptris, it can be as simple as a few brush strokes.
This program is almost too good to be true it was exactly what I was looking for.
I'm also finding sculptris is extremely well suited for sculpting clothing as well. Probably more so, for me, than 3d Coat. The materials are also more impressive than 3d-Coat's materials which have always been a bit low quality (I think to preserve voxel performance). But I don't know.
Yeah, but I love sculptris a lot. It is extremely excellent. Thank you for making it. QAs I said before I'll be donating.
I think an adaptive skin feature (like zbrush's) would suffice as an alternative to poking holes and merging intersecting objects. You could create a cavity and then drag both sides of the bridge out so that they overlap and then run an adaptive skin over it.
This would also work as a refining tool. Drag out the basic form and then run adaptive skin to erase the subtle sculpting errors while retaining the model's shape. Then sculpt finer details and repeat the process as necessary.
djolar: I thought of something like this at an earlier point, but not for bridging gaps or changing topology. I was thinking you could draw an outline on the surface to put a "blanket" over it, attached to that outline. Then you could select how tight or perhaps inflated that blanket should be. In the end it would probably be very similar to what you can do with regular smooth and other tools.
I don't know how ZBrush's adaptive skin looks or works, but you'd need some way of defining where it got applied, and what topology it would have. I'm assuming it's something like a "plastic bag" that gets shrink-wrapped onto the model?
The "adaptive skin" could use the same method as layered objects, actually. They both depend on geometry beneath them. The only difference is that the "skin" modifier is run once, and layering would have to readjust the "hull" as you work on the "core", and could have varying thickness. (As an example, soft tissue is an adaptive "hull" over the skull "core".)
As for the boolean operators - they may seem obvious and easier to make, and they would do the job of taking mesh topology beyond a sphere or a plane, but they don't seem to me as fitting Sculptris workflow. Right now Sculptris is very hands-on and sculptural; booleans would add another learning curve to it and push it into the "geometry" set of mind. Not the best idea if you want to keep the intuitive sculpting feel to it.
Using a tool (or a piece of "clay") to poke a hole through the model or add another lump to it, and then massage it into desired shape, would be much more hands-on than booleans.
My programming mind tells me that the layering of "clay" and self-merging adaptive mesh could actually use the same adaptive algorithm, and be related to the geometry-reducing method for the Smooth and Flatten tools - only instead of rewriting the mesh by detection of angles, it would work by detecting proximity and/or intersection. Adding polygon merging to the current polygon tesselation would be enough to open the way to both. So while it is more challenging to make for the programmer, it is also richer in possibilities.
The "skin" modifier would then just be a shortcut to spread an even layer of "clay" on top of your sculpture - might not even be worth doing automatically.
Also, if the setting for proximity threshold for mesh merging is made an adjustable setting, then it would control how thin you can spread the "clay" before it tears. So the people who don't want the mesh to merge could simply turn it off, while the rest of us will be able to sculpt as we do with real clay. It could even be set to never merge, producing "bubbles" of threshold thickness - pliable but not tearing. There could be some uses for that.
Perhaps it's something to think of for a later version?..