Sculpting continues apace.
Actually, something happened the other night that altered my trajectory a little bit.
I was looking at a rather delightful painting which I thought looked like a still from Monkey Island, La Forge at Marly le Roi by Alfred Sisley (http://www.wikipaintings.org/en/alfred-sisley/forge-at-marly-le-roi-1875).
I can’t remember exactly which image I ended up looking at after that one, but I was on wikipedia admiring a painting, and noticed that in the data block near the bottom of the page there was the name of the gallery in which it is currently hanging.
Furthermore, there was a listing of the latitude and longitude of the gallery, and a google maps link. Following the google maps link, I attempted to streetview outside the gallery (it did look like a pretty impressive building, and I wanted a better look) but instead, the streetview opened up inside the gallery!
Well, I spent ages wandering about the Musee d’Orsay and had a bloody great time (aside from some minor annoyances built into the google maps PC interface) and viewed some phenomenal sculptures.
One in particular caught my eye due to the slightly peculiar pose of the subject. That piece was Alexandre Schoenewerk’s La Jeune Tarentine (http://fr.wikipedia.org/wiki/La_Jeune_Tarentine).
So I had a shot at recreating it. Well, I had a shot at doing a low detail version. The reason for the low detail was that I’d just snagged a recent build of Blender and found that it had the new “constant detail” dyntopo mode built in. This means that the size of faces created by dyntopo is decoupled from the zoom level you’re at and specified as a percentage of a Blender Unit instead. I love this, and it makes getting a uniform detail level across a whole model an absolute cinch.
The skin modifier model I started with and some initial renders:
After fiddling with the model a bit more I had a play with a translucent material and a cloth sim and composed the following scene:
I was quite pleased with that one, although it still needs a lot of detail adding to the figure (like a face for example).
Fast forward to last night when I had a bit of a bright idea. I was wondering what I could practice sculpting that felt like it had a dynamic quality to it, some movement to the pose (regardless of it being a still) and I thought of ballet dancers.
A quick google image search later and I was sketching out another model in the same sort of way.
Another skin modifier model and set of initial renders:
And once I’d got that model looking the way I wanted it I started playing with high gloss materials and came up with this:
Which, again, I was pretty pleased with.
So all in all, I absolutely love the new constant detail mode, and I’m enjoying sculpting more than ever.
(Credit to the missus for saying “why don’t you do some full bodies rather than just heads?”)
A good few months ago I was struggling with the task of sculpting some of the more wiggly bits that people have; ears, fingers, toes, hair etc.
Each of these things are fairly easy to do in isolation, for example sculpting an ear without having the side of the head getting in the way wouldn’t be too tricky, but pulling those parts out of the main head or body object was proving to be an absolute nightmare.
At the time, lovely Mrs. Dave said something like “Can’t you sculpt them separately then stick them together?” and I thought about it briefly and said “Would be nice if you could, but stitching two sculpted objects together would be a pain in the ass.”
Fast forward to last night and I’d been thinking about it again. I sculpted two arbitrary forms and attempted to join them together. First I used circle select to select and delete a similar patch from each object. Next I attempted to alt-right click each edge loop, trying to select the vertices around the hole I’d made in each object. The trick to this was finding out which bits of broken topology halted the edge loop, healing those sections by removing edges, then trying again. Once I’d got them both selecting in a single loop, I tried using the bridge edge loops function to join them together.
This kind of worked OK, but I did have a bit of a job getting the edge counts the same to make it work nicely.
Then I had a bit of an epiphany. The only reason I’d warned myself off using boolean modifiers is that they tend to bugger up the topology of your model. But given that I’m using dyntopo, it hardly matters!
This is great! I can sculpt loads of different face/body parts and then piece them together, boolean union them, then clean up the seams in sculpt mode!
“To the face mines!”
First attempt at using the above technique (still WIP):
I’m a bit cross with myself because there’s plenty of other stuff I *should* be doing, but I can’t stop sculpting.
I fire up Blender. I see the cube. I think about vert modeling a bike, or a robot, or a building.
Then I switch to sculpt mode, turn on dynamic, and grab my stylus (ooh er…).
It’s a bit of an odd feeling. I can remember that I’ve made some decent stuff without sculpting. I can remember that there was a point when I didn’t have a stylus and barely sculpted at all, but for the life of me I can’t get “back there” and force myself to do something else.
I did try. I went looking for inspiration, but I didn’t get very far. I actually ended up back on Doris Fiebig’s site, so while there was more than enough inspiration there, it was all very… sculpty 😉
I can’t help but feel that I’m not ready to stop sculpting yet. I’ve done very few retopologies since I started this recent sculpting session (about two months now I suspect), partly because it’s dull as ditchwater, and partly because I don’t seem to be able to work in multi-res with quite the same flexibility as I can in dyntopo. “But what about texturing?” I hear you cry. Simples. I don’t care 😀 All through this, I’ve been saying to myself “If I can eventually produce fine sculpture, with all the right bumps and grooves in all the right places, it won’t *need* a texture, and if I ever specifically want to texture it, I’ll retopo it at that point, but not now.”
So there it is. I’ve carried on regardless. I may be damned to sculpting for the foreseeable future, but it’s an enjoyable damnation indeed, the most recent results of which follow.
Ps. Dynamic topology is lovely. I hear there’s some work being done to add a “detail flood fill” option. I can’t wait. It’d be super great also if you could crank the detail up and down and have it show you (non-destructively) what the model looks like detail filled at each level.
Jammy that the displace map made a nice base for the diffuse
Wrinkles and pores (although hard to see) added
Make it so-and-so…
I think green is his colour
Naff hairdo, and composed out the hands and feet, but otherwise happy with this I think.
Since the last sculpting update I’ve been practicing quite a bit, at one point almost two hours a night, and I’m not sick of it yet! Hooray!
I feel that the bit of learning/work I did on skin shading was a little premature so I’m going to keep it NPR as much as possible for the time being, looking for merit in just the form or an appropriate stylistic shading technique other than photo-texture semi-realistic stuff.
Looking back on that last paragraph it sounds a bit pretentious. How about this: I can’t do skin yet, so I ain’t gonna 😉
So here we go, some of my more recent sculptitudes:
Old geezer – sad about his rubbish ears
1992 was so long ago…
About as angry as it needed to be, and no more
How many fingers am I holding up Smith?
I suppose five, what does it matter?
Probably needs a hat