taron: Nice one, love the veins. Yeah, speed is kind of limiting in the alpha build. Actually last week I discovered a rather embarrassing bug that should never have been there. I think it's about 5x faster now. The speed is mainly limited by computation, and not rendering, so right now it wouldn't do much immediate good to hide geometry. I'm pretty sure I can speed it up another couple of rounds though, and will probably add some hiding later.
I don't have any huge plans for the app right now, can't do that or I'll buckle under the expectations. Mainly I want it to be a comfortable and quick way to play around, prototype models, paint them, have fun. It's important that beginners can just jump in and do relevant stuff without complex setup. Some of the more complex/powerful stuff will probably make it in at one point, but it won't take center stage, and I'll skip them if it causes too much trouble.
THANKS, by the way, almost forgot, hehe... yeah, just a quicky, really.
Hiding, however, should actually act like a mask and limit the computations to the area that requires updating (I know it goes beyond the borders due to the subdivision process). I would've imagined it helps a good bit. Naturally I can't know what you're doing already, though.
I'm excited to check out your update, of course! Color sounds fun, too! I definitely like where you're going with it and will see that I can help you show it off!
Mmm... Currently I'm basically auto-hiding all of the mesh that isn't immediately near the brush, at least as far as computations are concerned. This causes a minor delay at the start of a stroke though, and if you move far enough that it has to update the "hide mask".
Manual hiding would need to be really fast for the user if it's to be convenient. Selection boxes seem like the simplest and best approach - just line the model up and hide large sections by dragging boxes. Click some button to unhide it all. Alternatively in reverse - drag a box to hide everything but the stuff within the box.
You're right of course - if I implement it properly it should be just as fast as editing a mesh with the number of triangles you have visible (basically chopping them out and placing them in a temporary work set, completely forgetting about the hidden parts). Touching borders would cause artifacts, but you'd have to live with that (or automatically add some triangles to the visible set, but I doubt I'd go that far).
As for color, it'll be a simple addition now that I have the bumpmap painting working well. I'm assuming you've seen the youtube videos. The challenging part is to make it user-friendly, with texture management etc.
Cool demo pics are always awesome to have, so I'm very thankful for anything you can come up with. Hopefully the next release won't be so constrained.
Sweet, I still have to check out youtube, but you just get me even more excited!
Yup, selection boxed would do it. Lasso would be even nicer, of course, but not important. A spherical volume selection would be my favorite, of course, if the cenver would simply lock to the surface and you'd drag it out. Should be very easy to code, too. This would be best for just keeping what's in that sphere, hiding unselected.
That's exciting news considering this marvelous digital sculpting program has so much potential with how intuitive it is. I made some time ago a timelapsed (20 minutes into 5 minutes) youtube video : www.youtube.com/watch?v=WZ4q3HApxFA nothing spectacular, but just to show how user friendly Sculptris can be even to complete digital sculpting beginners like me.
About a 60 minute sculpt. I'm not giving it alot of small details as my computer slows down though I might do some more work on this one. The model was started from a sphere and is about 175,000 triangles...
The grab tool as I said before is really great on pulling out form, about 85 percent of the model is from just using the grab tool.
Yar. Automatic recovery is nice. Crashing not so nice though, of course. Symmetry switching in the alpha build is pretty much guaranteed to crash for complex models, it seems. Unfortunately symmetry has to be switched off when exporting to OBJ as well, so that will also crash.
You seem to have got some decent eyes on there, but it's hard to tell with the fixed lighting direction. I only really tried to make eyes once, and found it kind of tricky to get them round and undeformed - especially while keeping polycount down... Speaking of which, you can prevent the triangle count from building if you drag the detail slider down to zero. That lets you move vertices around without any sort of tesselation taking place, even if the brush is tiny. In case you didn't know.
I should play around with the app more than I do, but it's hard to relax and enjoy it when I'm constantly working on the code. Maybe I'll take a moment and spend more than 5 minutes on some model... in a few days.
Hehe, the trick with the eyes is not to worry about it too much. It's the eyelids that make it work and that's with masks and some gentle moves. I've made them very early on, so the chances were higher for them to be smooth.
What bothers me alot is the line that goes across the lighting, like some sort of environment map that doesn't wrap properly?! What is it and can you get rid of it?
Also, the lighting adjustable, even if only the color of it, would already help a lot, too! Worst case maybe some fixed choices (Frontlight, Portrait, Nightmare and so forth, haha).
I knew about the detail level concept and like it a lot, but I'm lazy and do things very quickly which occasionally makes me ignore setting changes.
TIPS/REQUESTS: - zoom alternative with numpad + / - or hold down some key or combo and mouse up/down - brushsize adjust photoshop style [ / ] or hold down some key and mouse left/right
I'm using a Wacom to sculpt and mousewheel is a pain with the little ribbon (intuos 3).
And, yes, you should definitely enjoy your fruits some more, it's excellent stuff and lots of fun!
taron: Sorry about the artifact. It's supposed to be a thin subtle line indicating center of model for symmetry. It's rendered incorrectly on most of the screenshots I've seen from other people though, so I must be doing something wrong. There's no way to disable it currently. One thing you could try is to upgrade your graphics drivers. Apparently one user had some success with that, enabling shaders rather than the fallback vertex lighting mode. You seem to have shaders working fine though, so I doubt new drivers would help all that much.
As you might have gleaned from new screenshots and videos, the lighting model has changed and is currently a strictly camera-locked setup with image-based light/material. This should be much more useful, and quite flexible since you can just swap out the image to get a completely different look.
I will try to remember sculpting around with my tablet a bit before I release the next build, so I can get a feel for what that's like and what could be changed in terms of input.
Pfff...I just watched the youtube video, can't believe I forgot about it. Well, how do I say this in the sweetest possible way... RELEASE SOON... I can't wait! Honestly, looks very interesting and I'm sure it'll be possible to scare some folks out there a little, haha!
By the way, GLSL shaders work fine here, but with XP64 I can't seem to go beyond DX9, I think. However, no problems here: Nvidia 7300 GT
Looks like there may be a flip or rather inversion on X then?!
The line is really a large and somewhat thick plane (basically a box). You might be able to see its outline if you rotate slowly around dead-center.
It's supposed to render using the stencil buffer, so only the intersection between box and mesh is shown. It's not rocket science, not even related to shaders, and the same technique obviously works for the brush cursor already, so hopefully I just mixed something up when rendering the line.
AAAAHhhhhh, I see... yeah, that's a wild idea! Well, I keep my fingers crossed it's just a lil' mixup. Worst case, however, I'd rather screw the middle line than keep the bothersom box. It's tough to find a proper angle without it in it. I wished GLSL would be no trou...you know, I just come to think of it, who without a proper GL card would want to work with this in the first place? Maybe you should consider just using shaders. It's so pleasant, when compatibility isn't so much an issue, hehe...