Method -- Mudbox. Sculpted a sphere with stamps. One diffuse layer, no other textures.
How long it took -- 20 minutes
What I learned -- Look at the ugly dark spots on the game ready asset. I don't think that is torn geometry, just some issue with the normal map it looks like to me. I've had issues like this before, so I expect there is some skill in knowing how your high res mesh is going to translate into a normal map.
Additionally, my level 0 rock is like 300 polys, which I expect is way to high for something like this. I'll have to research that a little as I continue to make small props.
Lore -- A small, gray rock that is smooth enough to be a little bit shiny. Found in a dry creek among many other similar instances. Throw it, perhaps, but it won't fly straight.
What I learned -- Hard to get very sharp edges like you would see on a piece of splintered granite or shale rock with the sculpting tools. Will take a bit of time later to make a base mesh with lots of flakes to simulate a layered type of rock. Also, for a rough crystalline type of surface, a bit of experimentation is in order to dial in the correct degree of specularity -- ideally I want the little bumps to shine a bit but the rest to be pretty matte.
Also, take note that with a darker colored object, any issues with the normal map are more easily hidden.
Lore -- a broken piece of hard but brittle rock with a small crystal. Your geologist friend finds this very interesting.
Lo, Med, and HiRes versions of a weapon asset for a game project in development.
Method -- Started from given concept art. Blocked it out in Maya, made finishing touches in mudbox and painted by projecting concept image and then retouching.
How Long it Took -- 2-3 hours. Most time was spent fussing with the normal map creation. Right now I spend as much time with the importing/exporting between programs as I do making the actual art. Keeping organized directories is important!
What I learned -- Get the larger shaping done with the blocking out and lower res layers first. Spent a lot of time fussing with something that should have been fixed before I got to the detail layers.
Also, after extracting normal map, it may be necessary on the lower resolution models to do some touch up between the paint layers and the normal map. In this instance, my engraved letters were not aligned properly with the paint layer on the lowRes model. But then after adjusting the paint layer to match the loRes model, it didn't match up with the highest res. That is okay because a million poly model is not going into the game engine anyway, but I need to pay attention to the layer system so that I can just affect the things I want to but not mess with the others.
Found a bug that divides my high res models into multiple pieces when importing into Unity.
Lore -- This model is for somebody else's project.
model Different resolution versions of an enemy character asset for a game in development.
Lowest version has 1.5k polys, medium is about 5k, highest has 90k.
Method -- blocked out in maya with a bit of extra detail done in there too. I treated this a little more like a hard surface because I thought it might be hard to get the hard look of an exoskeleton by sculpting alone. There is quite a bit more I would like to do with this model. For instance, I'd like to tighten up the joints on the legs, make them look more like an exoskeleton. I also wanted to make fur for key places (not the Maya xGen hair, that is not compatible with Unity, but just 2d images on transparent planes positioned around the model). But right now, I think it is more important to not work on any one project for too long, striving for perfection. Rather, at this very early stage in my development, churning out a high quantity will be the most efficient use of my time. Once I get pretty proficient with the game asset development pipeline, I'll work in some new techniques. I may revisit this model and make it the hairy menace it deserves to be, or I may just do another.
How Long it Took -- An entire day, 8-10 hours.
In mudbox, I initially had some trouble with my UV's from Maya. I should do another UV tutorial, because I am not very good at it. Part of the reason is because it is tedious and I don't like doing it, so I don't take the time to do it properly. In the end, I did the auto UV creation in Mudbox, which works good but I don't think it is the most modular friendly method. Time will tell.
What I learned -- Oh boy. A lot. This has by far been my most complex work, but also my best. Except some issues with the UV's on importing, everything worked out smoothly and my final game ready model looks just how I wanted it to in Unity. I am very please with this model. Furthermore, although this is far from AAA quality, I feel that I have identified a few deficiencies in my skillset that are holding me back from being able to create photorealistic (or close to it) models and better looking game ready versions as well. Primarily, I think I need to really get to know photoshop so that I can always be using the highest resolution images for texturing purposes, as well as be able to make my own stamps and stencils. Apart from that, it's just a matter or putting in the time and attention to detail.
Additionally: Amplify tool - basically a more powerful pinch brush. Really made shaping up some of the harder edges of the exoskeleton quick and easy. Modularity -- next time, make better use of paint layers so that I can return to my finished model and easily make different color variations.
Lore: This asset was made for an independent game developer. I refer to it as the Arachnoid.
A bunch of low poly car props for a zombie apocalypse phone game. Each model is around 300-500 polys.
Method -- I blocked out the models using side profile reference images as quickly as I could in Maya 2018. Once I got the "intact" version of a vehicle done, I duplicated it twice and messed up the vertices to get a damaged and totaled version of each. Once I got this workflow down, it really went fast and I could just zone out almost. I paid special attention to use the absolute minimum of polygons to get the desired shape. Since no smoothing is involved here, I didn't worry about n-gons or triangles at all. I also deleted every face that won't be seen from a top down perspective. With each model finished, I sent to Mudbox, auto-created the UV's in there for speed (quality isn't so important in this case), and slapped on a coat of paint to satisfaction. Next, I dragged on over into Unity, tweaked the shaders to my liking, and rinsed and repeated until I had this decent assortment of vehicles to add some variety to the upcoming mobile top down zombie shooter.
How Long it Took -- About 7 hours
What I Learned -- Compatibility between versions of Maya and Unity. Maya 2018 sends textures that Unity 5 can't read. They show up flat white. Maybe I was just doing something wrong though. What I figured out was that Mudbox and Unity work together well, so an easy workflow is to block out in Maya, paint in Mudbox, and export from there to Unity. No problems whatsoever that way. Some of the colors end up looking a bit brighter in Unity than they did in Mudbox, but a little bit of tweaking the metalness and smoothness took care of that.
Also, making a high quantity of low-quality assets can be fun! Lots of room for creativity. I know these look like utterly forgettable little props in some simple phone game, but they do take enough time to create that I like to imagine grim little stories as I paint them. The "damaged" ambulance, for instance -- notice how the back doors are bulging out. That's because a patient who was infected was back there and then turned, bit the paramedics who in turn, turned, and then chaos. The vehicle crashed and burned. The end.
Lore: A bunch of unfortunate vehicles caught in the wrong place at the wrong time with the wrong drivers.
This WIP is my project to learn Zbrush (and Substance Painter 2) with. Right now I've just begun roughing in the basic details, starting from the head down.
I am following a Pluralsight learning path for Zbrush, but instead of making the simplistic cartoon style alien in the course, I figured I'd do something that really excites me.
The cool thing about the Giger style xenomorph is that it involves hard surface and organic styles of modeling, as well as human characteristics and animal/insect characteristics. I think this project should jump start my skill in sculpting quite a bit.
What I've learned so far :
Note - since this will be such a long project, I'll jot down lessons I learn along the way otherwise I'll forget.
1. I had a lot of trouble getting the hard planar type edges on certain parts of the exoskeleton when starting from Zspheres. Although this is an organic model, much of it is actually more like hard surface modeling, like armor, and so although I want to get familiar with Zspheres, I'm going to start my initial blocking out over in Maya. I think this will benefit me down the line, too, as I will be able to duplicate certain parts more easily in Maya after sculpting in Zbrush. What I mean is the tail -- it's a series of vertebra like parts. I could just build one, duplicate, and then scale them down to the tip. I am sure there is a way to do something like that in Zbrush, but I know how to do it already and quickly in Maya. That is a hard piece to so it will be easier to do with Maya anyway.
2. Blocking it out in Maya worked well. I had a good amount of time wasted cleaning up bad geometry -- specifically with the tail -- but all good lessons learned. Only messed around with sculpting a little in Zbrush but it seems like this base mesh should be good to go. I'll get a lot done tomorrow.
****Make sure to take as much time as necessary to get proportions just perfect before doing any detail sculpting, otherwise you get stuck in an endless loop trying to get things looking right but it just isn't possible without the correct proportions. 3. A few progress updates -- learned how to make custom insert mesh brush, as well as use Dynamesh. And what a great feature Dynamesh is! A lot of creative power in one click -- wow.
And a technical note -- refer to the bottom picture -- for some reason, curve size is only updated if I adjust the brush draw size from the indicated slider. Pressing S or using the quick menu slider will not affect the curve size. I don't think that is meant to be, so I am not sure why my program is behaving this way.
Got a little done this morning. Starting to look like something. Will get in rough detail over the entire model, then Dynamesh to a high resolution and fill in little things.
I'm tired of looking at this guy for today, so it's quitting time. One thing I just learned -- had been Dynameshing at 1024 resolution, which is enough to get the kind of basic detail seen above. I didn't want small things to fuse together when I did the next Dynamesh, so I bumped up the resolution to 2048 (I just use these standard numbers because the guy in the tutorial I learned from did, I don't know that its really necessary).
But at 2048, some of my finer details were still lost, although the mesh is now dense enough that I can't smooth out the kind of basic detail I am still putting in. So, lesson learned, just stay at the 1024 level until I get the entire mesh ready to go up to the next level of detail. Don't be too eager to get in fine little details, but they're just going to get erased when I go up in resolution most likely.
Hey amigo. ITs great you're doing a xenmorph. I cant wait to see wht ou can do. I would suggest starting at a lower sub-d instead of with millions of polygons, it will definitely save you a lot of headache in the future. YOu can worry about the details later after you get that form blocked out.
Hey amigo. ITs great you're doing a xenmorph. I cant wait to see wht ou can do. I would suggest starting at a lower sub-d instead of with millions of polygons, it will definitely save you a lot of headache in the future. YOu can worry about the details later after you get that form blocked out.
Thanks for the advice. I was kind of learning that the hard way, but since I am working with Dynamesh, if I go too low resolution it sticks some of my smaller pieces, like the vertebrae-like pieces on the tail, together. I suppose that is something I could fix down the line, but its seems like it could be very tedious. Then again, maybe I could just make another custom insert mesh brush and draw it right out in no time...
Getting a good feel for the brushes by now. I find myself using DamStandard, pinch, move, and clay buildup 90% of the time. Ribs still lack the hard bony look I want, probably due to the slight waviness of their edges, but I think I'll wait to go up a subdivision level to really refine those edges. I like the overall positioning and at this point going in that small just moves more polygons than I need.
Simply running the pinch brush along some of the natural edges, like the light protusion on the forearm, is enough to give a nice hard, exoskeleton like look. It looks a bit doughy in the screenshot still, but alive in the viewport those hard places are starting to look how they should -- I think so anyway.
Still to do -- general detail on legs, tail, and back, and need to completely redo the hands. I had to build new geometry as the originals got messed up during Dynameshing. Touch up details on neck area. Then, go up a subdivision level and really refine and add small texturing details. A few more days I think, then it will be ready to paint.
Upper body has most of the general detail I am going to put in. I may go in and add exoskeleton joint ridges along the knuckles, but I don't know if I want to go that far with this model as its just a means to learn Zbrush with. But it is looking pretty cool, so I may spend a little extra time with it. Right now, I am doing all of this detail at 1.7 million polygons. Some area's like the fingers are pretty strung out at the ends of their resolution, but I think it's fine for now.
I don't think I will have to do any more Dynameshing. Shouldn't be adding any more geometry. I like the way the neck tubes turned out.
Right now, in retrospect, I think it may have been better to build the ribs as a insert mesh custom brush -- probably would have saved a lot of time and looked a little nicer. I decided to mold the upper ribs into kind of a solid chest plate, as if the thing was once a humanoid creature but it's slowly becoming like an insect. When I get to the finishing details, I'll put lots of little spikes on that upper chest plate like some beetles have.
Technical notes:
Dynamesh can ruin your geometry if not using 512, 1024, 2048 type resolutions. I don't fully understand this, but basically just use it when you are building your base mesh so that you can easily incorporate major pieces of geometry to your mesh.
Also, while in Dynamesh mode, be wary of accidentally tapping "D", as that turns on a process that will really chug your machine down if you have a high amount of points on screen. If you do, just go into Tool>Geometry>Dynamic Subdivision" and make sure that is turned off.
Athough the point of this project is not to be worried over the art itself, I can't help but get somewhat engrossed with it. I had been unhappy with the way the ribs were looking. You could easily see that they had been pulled from the torso itself -- they didn't look like separate pieces. But I was looking at some really badass artwork of xenomorphs a bit and decided to try and get a lot more of the biomechanical elements in. When I strung some of the coiled tubes between the ribs, it accomplished perfectly the look I wanted. It is a little difficult to see in the photo, but now the ribs appear to rest above a convoluted network of disgusting pipes.
When I go up in resolution soon -- once I finish the tail -- I will attempt to make a slimy, sticky looking skin layer that will be stretched over some of the pipes around the abdomen and neck area. I am not really sure how to do this at all. I think I'll just put little strips here and there and then later, when its time to paint the model, maybe make them slightly transparent or something.
The monster is starting to come alive now. I'm having a ton of fun getting familiar with all the different alpha's now. I think tomorrow I'll give it some teeth. That will really bring out the character I think.
This is a long way from AAA production quality, but for a beginner model I'm pretty proud of it so far. I didn't mean to get so carried away with this as I'm kind of itching to get in and start learning Substance Painter 2, but I'm just having too much fun.
I've switched projects temporarily. I think the Xenomorph could make a really nice first piece for a portfolio, but there are a few more tools I need at my disposal to really pull it off to the degree that I want. So I am working through another Zbrush tutorial to expand my library of techniques. I'll post a few pictures of the model associated with this tutorial, and then when I'm done I'll start a similar but new xenomorph project. This time, I'll be designing it myself and making a plan from the get go, so it should end up looking very nice and professional. Earth Golem creature. Not taking a lot of time to make it look nice, just learning the many different techniques for mesh creation, and perhaps more importantly how to best use grouping and subtool layers for a modular, well organized project.
This monster isn't exactly my style, but I think it is looking pretty cool nonetheless. Would go good in a fantasy game like Dragons Dogma. For the most part I am copying the tutorial model pretty closely. One thing I ought to do is make the left arm branches more angular so they appear more deadly. They were created from zspheres, and I'm not sure of the best way to go about doing that. I can go in tediously with the polish brush, etc., which is what I guess I'll do. But I imagine there is probably a better way -- there usually is.
If youre trying to to make some cool custom branches, one way you could go about it is, transpose modeling the branches, or even making your own custom IMM brush. Personally I'd just model it in another software. (Modo in my case) and then just GoZit over.
Good point. Because I am more familiar with Maya, I often think "I could do this part really quick and easy in Maya", but in this case I am just following along with the tutorial. But what you suggested, masking them off and then using the transpose to bend them at sharper angles -- I think that will be a reasonably efficient method. I'll give that a try.
So I ended up just making a new arm from zspheres, using the original transparent one as a reference. This time I just made an effort to make more sharp angles and trim the ends down a bit. Didn't take very long and was pretty easy. I guess if you want to careful detail, you can't avoid all tedium.
Technical note: If when you grab zspheres with the move transpose tool you find that you grab more than what you want and have trouble moving just one zsphere without its neighbors coming along with it, adjust the size of your brush. As with curves, the larger your brush is the larger the area of influence. Important little things like this never get mentioned in tutorials, but they should! Zbrush has a lot of little nuances like this that make it frustrating to learn, but once you do figure them out they are really handy features.
Also, if you are having trouble rotating the camera the way you want, be aware that when local transformation is turned on, wherever to click becomes to centerpoint of your rotation. So annoying at first but now that I know how to use it I really like it. Almost done with the blocking in. A few layers of fibermesh moss and and leaves and then the details will be applied. I wasn't super excited about this model at the beginning, but it's looking pretty awesome now. I could imagine fighting it in Dark Souls.
Yet another short deviation. This one was made by request, but it is nice to take a detour from the longer projects and get something finished. This is a model for a game and will be finished with textures and fur tomorrow.
Bobcat Low Poly
Bobcat Hi Poly High Poly
It looks a bit fat, but that is because I have modeled the bulk of the fur. I will add fur cards around the cheeks and a few other areas which will hopefully make that more apparent, and hopefully I can learn enough in Substance Painter 2 to make him look pretty realistic.
Having a very big fondness for mammals, this was a lot of fun to make.
Took about 3 hours to model in Zbrush, starting from zspheres, then blocking with Dynamesh, then sculpting almost entirely with the clay buildup brush (or BCB, I call it, as that is the key combination to get to it).
To get multiple pieces of a mesh into Substance Painter as separate objects -- like eyes, nose, teeth, etc -- you have to assign each it's own material. Give the material the appropriate name as it will be associated with the mesh in SP.
Also, has to have UV. But will seams matter that much? I guess I've figure this out on the way.
My wife, working up to being a post-apocalypse zombie slayer. Looks a tad creepy at the moment and she's not happy with me. I reassured her it's just the lack of hair and proper eyeballs.
She'll end up with a coat and gloves, as well as a rifle and backpack, and my handsomest looking dog will join her in wasteland scene.
I'm supposed to be working on the Bobcat model, but I needed to go through a Substance painter tutorial first. That tutorial involved texturing a post-apocalypse female character, and as I was working on it I said, "Hey, I'll just make my own model real quick."
Well, real quick has already grown into about 10 hours of work and I've still got a long ways to go.
The face was mostly okay so far because I was working from a reference, but the body I had just been winging. Looking at it, things just seemed off, and man were they. So I took some reference photos and learned how to use this incredible reference system in Zbrush using the grids and transparency. Really nice!
Much better. Using a better reference and taking more time to pay attention to subtleties really helps. Now I can actually recognize the model despite not having hair. Now ready to get into the little details.
Ignore the Trump hands. I just started forming them. I can't believe it, this really looks pretty accurate. I can't wait to turn my wife into a zombie slaying badass.
I'm happy with the face. Made some small but crucial tweaks this morning. With this model, I have the benefit of studying everyday, so I noticed a few things like the depth of the eye orbital area that I missed before. You know, having reference images really helps, but it seems with 3d you still can't get everything just from the 2d images. Sometimes you have to kind of imagine how certain parts will be shaped if you cant discern it from a reference image.
It's hard to see well in this image but there is a layer of skin bumps/pores that catch the light nicely and add some realism. I did have a subdivision level at 14 million about, but this layer at half that looks fine so I got rid of that highest level to keep things running smoothly. I think I'll need to take each subtool pretty high to get the details I want.
As much as possible with hand painting and my current level of skill, I'm going to try to make this model look photorealistic. The materials on their now are just because I like differentiation between the pieces, but I'll be texturing this with Subtance Painter 2 when the modeling is done.
Now I'm moving down the arms and hand. The arm has some shape already, but I find it difficult to maintain orientation when rotating around int. The way the muscles wrap around the arm and the very soft, subtle nature of my wife's female arm make it challenging to get just right, particularly in the elbow area.
keepin up the good work bro. a few things generally -the underlying structure of th for m is missing. specifically the land marks of the human body.
These really help figure out whereyour proportions are goin off the rails, in my experience
-body parts are way too long, that neck is far longer than it should be. I'd go back and check that. -your proportions of the head could use work. -the ears are sunken in to her head. -youre also missing quite a bit of muscle definition.
my best sugestion att his point would be to just go to your first subdivision level, and use your move brush and move around the forms until you get this model looking correct
You know, I was feeling the same way about some of those things. The reason I haven't modified them however was because my model lines up almost exactly with my front and side reference images.
But, as you have noticed, the neck does seem a big long. The ears and head -- to be honest I didn't put much time into them because they will be covered up. So I expect the problem might have something to do with the perspective differences between the camera and working in 3d or something like that.
*** Edit : Actually, as I was typing this, I was closely reviewing a ton of pictures of my wife. Turns out, she actually just has a pretty long neck. It gets covered by her hair, but I actually measured it out against other parts of her body and the proportions were consistent.
Also, for future reference regarding this model, she does have pretty long arms and legs too. In fact, although I am 4 inches taller than her, her arms and legs are the same measurements as mine.
Also -- thanks for the reference image Paul! I have seen some videos covering these landmark points, but I haven't modeled any humans since watching that video some time ago. So I'm actually going to save that picture and set it up on my second monitor.
My primary resource will still be the actual photos of my wife, and she will be clothed so many points won't be seen, but I think it will benefit my professionalism to kind of have those landmarks memorized.
The funny thing is that this is my first human model, and although I have no other experience to compare to, I think my wife is pretty difficult to model. I mean, she doesn't have a ton of wrinkles or anything, but in real life -- and of course in my opinion -- she's really beautiful, but it seems that there is a fine line between beautiful and otherwise and it is very subtle. If I miss one minute detail, the model just doesn't look like her. I mean, it may still look fine to most people, but for me this is to be an exercise in attention to detail (well, and just general workflow methods and skills)-- I want to make it look exactly like the reference.
Got along this far, felt pretty happy with everything. Realized she has no ears.
Solution:
*create an ear and make it an Insert Mesh Brush (and save it for future use)
*Duplicate head subtool
*On duplicate, freeze or delete subdivison levels. If it is very high, freezing and then unfreezing may lock up Zbrush or crash it. In my case, I just had to delete the highest level and then redo those finer details again.
*Create polygroup where you want the ear to go on the duplicated head mesh
*Drag out the Insert Mesh ear and position it as perfectly as you can
*CTRL-click-drag two times and the new ear will be bridged to the group you created. Touch up as needed.
Lesson learned, don't forget shit! And before you go up to do any details, review review review to make sure everything is absolutely perfect.
Wanted to write down a few notes about making the fabric folds and wrinkles:
I like to use a minimum of brushes. If I'm constantly switching tools I get confused.
What I found is that at a medium to low subdivision level, you can do almost all the wrinkles, large and small, with the DamStandard brush. Just make it big and dig in for the larger folds, then smooth, and small for the harder edges. The details aren't quite done but I think it is looking pretty good. For subtle noise, I just tapped the crumple brush around. Worked as good as any alphas.
Finished with the undershirt and pants. Moving along to the coat now.
Boots seem like something I could sink a lot more time into, but for now these will suffice. I think they would be fine for a game. But perhaps later I'll take some time to make some very detailed boots. I'd like to make the soles and laces very realistic, but this just seems to take a lot of time. Right now I want to keep moving forwards.
Hoodie is mostly done. The hood could use some touching up in places, but I'll come back to that. Onwards to the leather vest, then the hair planes before final touch-ups.
Not sure why my screen snips come out so low resolution. I thought they were supposed to be the same resolution as the screen.
Although I'm happy with results so far on this project, I'm going to take some time to learn a few different techniques that I think I may be able to bring into the project.
Onto the hair, which I am still unsure about how I want to do. I think I will attempt the transparent cards style, just to do something different from sculpting. Also, this character will have very curly hair, so I think it will be better to do that with cards rather than sculpting.
I may try and make a few hanging strands with fibermesh.
I'm actually going to learn fibermesh before trying the hair cards method. I think I'll get a nicer result with fibermesh, even if it's not as performant for games.
Just thought I'd share this hilarious photo. Apparently this is part of the process.
Messed around with fibermesh enough to realize it will take a ton of dinking around to get the super tight curls this model has. I learned how to do hair cards, and had a hard time with the UV's. I hate doing UV's, but at least now I am somewhat decent at it.
I don't think hair cards will give me exactly the result I want, but I suppose it is possible with any method with prerequisite skill. I'll just have to learn as I go. If I can't get the curls I want, I'll just settle for wavy hair.
Here is the hi-res sculpt completed. Not perfect, but good enough for a first character model. Now I will retopo, make UV's, and then start texturing in the next few days. I will do the hair last, but I may have to redo the hair cards, I'll see when I get there. I am thinking that since I can easily paint the normal channels in Substance Painter, I may just make a few big cards rather than lots of thin ones.
What a tedious nightmare retopology is. Maya quad-draw is pretty efficient and easy to use, but this is still a very slow process for me. I used Zbrush Zremesher for the vest and did a few touch ups, but the face and hands and legs would normally be animated so I wanted to do those manually.
I've done some research and I believe my edge flow should be good for animation, but there is a big high valence vertex on the inner thigh I am worried about. Too tired to mess with it now though.
Will finish the hands and boots later this week, then onto UV's. Ugh. I just want to texture.
Also, I have a feeling the zippers won't come through the normal map baking process in tact. I guess I'll just have to wait and see.
New, low poly model weighing in at 9k quads. I am imagining this as a hero character for a AAA game, so I think that is not a bad range. I've read that many AAA characters are as much as 50k? Not really important for me right now, but just want to make a note of this for future.
I had some success with a proprietary technique (j/k, but I came up with the idea on my own at least) to use a premade generic hand model and "shrink wrap" it around my hi-res hand. I wasn't able to get Maya's shrink wrap deformer to work in this case, but using make live and quad draw I got fantastic results. Made all the time spent fiddling around worth it.
Put some extra geometry around the joints for animation.
This is how I did the hands. Positioned the lo-res over the hi-res like a glove, made the hi-res live, and then used smooth brush to snap the lo-res into place. Was pretty quick, and this way I didn't have to spend too much time fussing over good topology flow. That work was already done.
Spent a long time reviewing other models and figuring out how to get these loops just right, and then routing the extra geometry efficiently. I think this is a good result and I'll be more efficient with this in the future.
Spent better part of the morning trying to find out why my face was unfolding like this. Thought I had some bizarre geometry or something. Finally I found out that the Unfold3d plugin was not enabled. Turned that on, then everything worked. Got the UV's done without any trouble after that, and pretty quickly I think too. Maybe an hour or a bit more.
Not sure about making seams down the cheek lines. I have seen a couple models like this, so I figured I'd try it. It made the unfold easy anyway. I think with 3d painting applications like Substance Painter dealing with seams shouldn't be a big issue. Guess I'll find out.
After the holidays I'll get the ID, normal, and AO maps baked in Maya, then, FINALLY, I'll get to work in Substance Painter.
When all that is finally done, I'll do the hair from scratch. I want to learn how to do pretty well, so I'm going to treat that as a whole project in and of itself.
Although this project is far from finished, I'm already getting excited about my next one. I will do a few smaller projects for a friend first, but I think my next character will be based on this fantastic looking Dani Tribesmen I saw in a Jared Diamond book. That will give me lots of practice with anatomy and wrinkles.
Just got back from a short vacation. Didn't have the energy to jump back into the ongoing project so I played around with some sculpting for the next big one I want to do.
I couldn't find a good reference for what I had in mind initially. But, ca you guess who this is?
Jumping back into the female apocalypse survivor project....
I've picked up a few valuable lessons so far. Currently, I'm having to do a lot of retrofitting as I work to bake the normal, AO, and ID mask as I prepare for texturing. Because I am spending so much time fixing up errors I made earlier in the project, I decided just to go ahead and do a reboot.
This is beneficial in a couple of ways. One, I wasn't happy with the clothing, so I will be making some changes and also working on improving my cloth sculpting techniques. Two, I felt I could do better with some of musculature sculpting, as well as just generally making the model more interesting and visually appealing.
A good thing I did was save all the different pieces out as separate OBJ files, so I could re-use them later. Glad I did that.
An important thing I am making sure to do is not jump ahead to any detailing. I wasted a lot of time previously by not adhering to a stricter process of finishing everything completely before going up to the next level of detail. This not only meant that I ended up redoing things, but I also had organizational issues due to too many versions of the same pieces being made.
Pictured is the Version 2 model as it currently is.
I don't like the chest pockets, but I'm not going to waste too much time there. I'll have time to get better with clothes in the future. Besides, once the whole thing is a dingy olive drab they will be hard to see anyway, so this will do for now.
Working on constructing a Vietnam era, US style LBE system. Built the pouches in Maya, and just used IMM brushes that come with Zbrush for the straps and waist belt. Looks pretty good I think. Tomorrow I'll sculpt the pouches to get them looking realistic, then I may either do a small rucksack or go down to the boots. Or maybe a hat. Who knows.
I wont be doing the pants though. Tired of doing folds for now.
Technical note: Ended up redoing the shirt again. Purely sculpting in Zbrush kept giving me lousy looking creases and pockets. Lots of waviness and just didn't look very good at all. So I used the body as a live surface in Maya to draw out a new shirt on, and then made the collar and whatnot nice crisp edges. This worked much better.
Also, I took some time to study how people work topology around the shoulders for animation, and although I will end up retopologizing again, it was good to practice that.
And here is an update with some custom pieces I made. I had just been sticking with the freebies that come with Zbrush, but I decided to take some time to make some hardware like the belt buckle and friction adjuster pieces for the shoulder straps. I figured its good modeling 101 practice and I can put them away in a library.
The straps at the junction are a bit ugly. I don't know a good way to manipulate them without getting ugly twist. This character may end up wearing a backpack, but I should still try and figure this problem out.
Also, I wanted to get these backstraps nice and tight looking, so that they look like they are holding weight. They are a bit wavy, but I've given them enough time already. In the future, I need to spend some time messing around with IMM curves so I can get things like this looking just right.
Fixed the Y-strap. Was no need to actually try and route it through the buckle. Just make two different straps. Duh.
I figure this is about the distance character would be from screen in 3rd person game (which is the prototype project I'm hoping to put this character into). So when I start getting carried away with little details, I try to come back about this far and make sure things read fine. If they do, it's time to move on.
LBE is finished. Would be more interesting if wasn't symmetrical gear on waistbelt, but I will make a similar male character in the future, and if I make some different gear for him then I can swap stuff out later.
Also note that the folds suggest a very soft fabric. I actually just got my wife to wear an old BDU jacket of mine. Should of done that earlier, now I have a clearer image of what the folds need to look like to properly suggest the stiffer kind of fabric.
Finally, I figured out a technique to get crisper looking fabric.
These pants are not done yet, but already they are looking more like how I envisioned. Instead of using fold, standard, pinch and smooth brushes to make the creases and folds, I just draw the shapes of the folds in with mask lasso, and then use hard polish. Just make sure you know which side is above and below. Works pretty good.
Also, don't bother making laces yourself. That's for chumps. Haha. I just found some laces for free, as well as some eyelets. Ain't nobody got time for that!
For the boots, I watched a tutorial that involved blocking them out with Zmodeler. I hadn't used Zmodeler yet. Seems like it could be faster to work with than Maya, but it is a lot of new stuff to learn and I am pretty comfortable in Maya already. Since I am trying to get this model done so my friend can animate it, I'll just stick to what I know.
Once I get these pants finished today, I'll be almost finished with the hi-res sculpt. I'll touch of the hands and forearms and then make hair cards. I hope that goes well because so far I am pretty proud of this model.
BUt perhaps for the next character I'll try working solely in Zbrush.
12//3/2017
Taking a little time to get familiar with Marvelous Designer. I learned a lot about Zbrush sculpting these clothes, and I am very happy with the LBE, but I think I can do better and faster with Marvelous Designer, so I'll give it a go after a few tutorials.
12/4/17
I am abandoning MD for now. I just can't take the frustration right now. Something about the sewing -- I just fuck it up every time. I am normally saintly in my patience, but something about this just raises my blood pressure to no end. I'm going to finish this character out as is. I'll see what I can do with the texturing to get the fabric looking a bit better, but major improvements will come with the next iteration.
I was having a bad day while trying to learn MD. No patience to pay attention to details.
Wasn't no big deal today. Got these cool pants done. Next up, the jacket, then I'll bring them onto the character in Zbrush and be rid of those lesser garments. Marvelous!
MD result is much better, and waaaay faster too. I found a step by step tutorial at patternmakerpro.com that shows how to make these military BDU style pants. I used a different avatar to build them, so some sizing and fit is in order, but overall this is really good.
So glad I took the time to get to grips with Marvelous Designer. This looks so much better. Now I've just got to do the hair, which will just be a sculpt -- not alpha cards -- for now. Want to get this model in engine as soon as possible so the animator can bring it to life.
Note to self -- sculpt next character model using one of the MD avatars as a base. This will give closer proportions to the clothing and require less tweaking to fit. In the future, see about how to rig my own models so I can use them as avatars in MD.
Also, take this outfit and make several texture variations of it.
Bandana and hair, and the high poly is officially done. Now lets get it ready for a game and texture it too. Can wait to dirty this character up and do the hair/eyes.
A technical note for anybody doing retopo with Maya quad draw.
If you have a highly bumpy area like the rolled up sleeve here, Maya gets confused and wants to drop the vertices down to the nearest point, so if you are running the smooth brush over it the mesh is popping in and out. No way to keep the faces uniformly over the top. So I think what needs to be done is to come back to this area and use soft select to bring the low poly mesh just out along the surface of the high poly, rather than cutting in at some places because the vertices are diving down into the low places.
Oh, and also, when you are using quad draw and you want to extrude a loop -- don't use the quad draw method as it will pull the loop in one direction only. Switch out to your regular selection tool, and in 2018 you can shift click your scale tool to bring the loop out in all directions equally. This means you have to switch tools, but it seems faster than dragging out one side, then another, and another, and then connecting the sides, etc.
OD Green variation. Again, this isn't textured yet, just putting materials on in Zbrush to get an idea.
Retopo so far. Very happy about the face topology, I think it is pretty good. The hands have a few triangles in places that make me nervous, but we'll find out how big a deal this really is. I've been hung up on topology a lot lately. I feel like I'm getting better, but certain parts still frustrate me. Extra ring around the top of the wrist probably isn't necessary, but the animator can decide for himself. Since this hand will be holding a gun, I figure some extra space around the knuckles to stretch out will be good. Couldn't find a good way to move around the thumb. But these hands won't be doing THAT much or be front and center, so hopefully this will do.
Face topology is probably too much for my purpose, but I wanted to make it something thaht could be animated as if it were going to be in a cutscene.
My 2017 and 2018 versions of Maya have a bug that won't allow the Unfold3d plugin to work normally. I've spent another four hours this time going through all the suggested troubleshooting with no solution.
Finally I've switched over to an application called UV layout, which is really good but this meant I had to spend most the day learning a new application.
Although the UV's are free of distortion, I do have a lot of seams that could probably be avoided. Most of them are in incognito places, and will be part of a camouflage uniform and thus hard to see anyway. I hope that by texturing in 3d with Substance, these seams won't be a glaring mistake. But the fact is, I am to done with UV's right now to give a damn.
If there is serious flaws, I'll keep this in mind for my next model. And I really ought to go through an intensive character UV layout course.
One good lesson learned though is that if you carefully make the topology on your hands/fingers so that you have a continuous loop running along the insides of the fingers, cutting that UV seam is as simple as one click.
Ok, some hard lessons learned from the previous character. I think the root of my problems stemmed from my workflow. All of the tutorials I have followed go in a high-to-low manner, which is probably better suited to people with lots of experience and solid understanding of the many factors that make a game model good.
So I had been focused all on the sculpting, and when it came time to turn the mesh into a game model, I had problems. So I am approaching things different now. I am going low-to-high. I've just finished box modeling out this male guerilla fighter character. This will be the game model, and I will do the sculpting from this. I hope that by doing it this way, 1. my UV's will be easy because I'm really familiar with this guys topology by now, and 2. I shouldn't have any issues with the low and high poly meshes not matching up well.
Furthermore, there won't be an annoying retopology stage to do later. Well, we will see how things turn out soon enough.
Sculpting is done for this guy. Now once the UV's are finished, we'll see if my new workflow will make map baking easier/better. Because this is sculpted directly from the low poly, everything should match up perfectly. I've also taken care not to put excessively large folds in the clothes. I didn't use MD this time, just sculpted in Zbrush.
Finally! It feels great to get over that daunting hurdle of UV'ing and map baking. I still don't really have a solid understanding of what the hell I'm doing in those stages, but these results are satisfactory for now.
This is my first complete (almost, just needs textures) game character.
Update:
Got some of the texturing on the face close to done. Filled in other area's with base colors. Having a lot of fun playing around with Substance Painter. This is definitely the most user friendly application I've used so far.
Replies
Level 7 rock_small_smooth
Method -- Mudbox. Sculpted a sphere with stamps. One diffuse layer, no other textures.
How long it took -- 20 minutes
What I learned -- Look at the ugly dark spots on the game ready asset. I don't think that is torn geometry, just some issue with the normal map it looks like to me. I've had issues like this before, so I expect there is some skill in knowing how your high res mesh is going to translate into a normal map.
Additionally, my level 0 rock is like 300 polys, which I expect is way to high for something like this. I'll have to research that a little as I continue to make small props.
Lore -- A small, gray rock that is smooth enough to be a little bit shiny. Found in a dry creek among many other similar instances. Throw it, perhaps, but it won't fly straight.
rock_small_porous level 0
rock_small_porous level 7
Method -- mudbox, start from circle
How long it took -- 15 minutes
What I learned -- Hard to get very sharp edges like you would see on a piece of splintered granite or shale rock with the sculpting tools. Will take a bit of time later to make a base mesh with lots of flakes to simulate a layered type of rock. Also, for a rough crystalline type of surface, a bit of experimentation is in order to dial in the correct degree of specularity -- ideally I want the little bumps to shine a bit but the rest to be pretty matte.
Also, take note that with a darker colored object, any issues with the normal map are more easily hidden.
Lore -- a broken piece of hard but brittle rock with a small crystal. Your geologist friend finds this very interesting.
Method -- Started from given concept art. Blocked it out in Maya, made finishing touches in mudbox and painted by projecting concept image and then retouching.
How Long it Took -- 2-3 hours. Most time was spent fussing with the normal map creation. Right now I spend as much time with the importing/exporting between programs as I do making the actual art. Keeping organized directories is important!
What I learned -- Get the larger shaping done with the blocking out and lower res layers first. Spent a lot of time fussing with something that should have been fixed before I got to the detail layers.
Also, after extracting normal map, it may be necessary on the lower resolution models to do some touch up between the paint layers and the normal map. In this instance, my engraved letters were not aligned properly with the paint layer on the lowRes model. But then after adjusting the paint layer to match the loRes model, it didn't match up with the highest res. That is okay because a million poly model is not going into the game engine anyway, but I need to pay attention to the layer system so that I can just affect the things I want to but not mess with the others.
Found a bug that divides my high res models into multiple pieces when importing into Unity.
Lore -- This model is for somebody else's project.
Different resolution versions of an enemy character asset for a game in development.
Lowest version has 1.5k polys, medium is about 5k, highest has 90k.
Method -- blocked out in maya with a bit of extra detail done in there too. I treated this a little more like a hard surface because I thought it might be hard to get the hard look of an exoskeleton by sculpting alone.
There is quite a bit more I would like to do with this model. For instance, I'd like to tighten up the joints on the legs, make them look more like an exoskeleton. I also wanted to make fur for key places (not the Maya xGen hair, that is not compatible with Unity, but just 2d images on transparent planes positioned around the model). But right now, I think it is more important to not work on any one project for too long, striving for perfection. Rather, at this very early stage in my development, churning out a high quantity will be the most efficient use of my time.
Once I get pretty proficient with the game asset development pipeline, I'll work in some new techniques. I may revisit this model and make it the hairy menace it deserves to be, or I may just do another.
How Long it Took -- An entire day, 8-10 hours.
In mudbox, I initially had some trouble with my UV's from Maya. I should do another UV tutorial, because I am not very good at it. Part of the reason is because it is tedious and I don't like doing it, so I don't take the time to do it properly. In the end, I did the auto UV creation in Mudbox, which works good but I don't think it is the most modular friendly method. Time will tell.
What I learned -- Oh boy. A lot. This has by far been my most complex work, but also my best. Except some issues with the UV's on importing, everything worked out smoothly and my final game ready model looks just how I wanted it to in Unity. I am very please with this model. Furthermore, although this is far from AAA quality, I feel that I have identified a few deficiencies in my skillset that are holding me back from being able to create photorealistic (or close to it) models and better looking game ready versions as well. Primarily, I think I need to really get to know photoshop so that I can always be using the highest resolution images for texturing purposes, as well as be able to make my own stamps and stencils. Apart from that, it's just a matter or putting in the time and attention to detail.
Additionally: Amplify tool - basically a more powerful pinch brush. Really made shaping up some of the harder edges of the exoskeleton quick and easy.
Modularity -- next time, make better use of paint layers so that I can return to my finished model and easily make different color variations.
Lore: This asset was made for an independent game developer. I refer to it as the Arachnoid.
A bunch of low poly car props for a zombie apocalypse phone game. Each model is around 300-500 polys.
Method -- I blocked out the models using side profile reference images as quickly as I could in Maya 2018. Once I got the "intact" version of a vehicle done, I duplicated it twice and messed up the vertices to get a damaged and totaled version of each. Once I got this workflow down, it really went fast and I could just zone out almost.
I paid special attention to use the absolute minimum of polygons to get the desired shape. Since no smoothing is involved here, I didn't worry about n-gons or triangles at all. I also deleted every face that won't be seen from a top down perspective.
With each model finished, I sent to Mudbox, auto-created the UV's in there for speed (quality isn't so important in this case), and slapped on a coat of paint to satisfaction. Next, I dragged on over into Unity, tweaked the shaders to my liking, and rinsed and repeated until I had this decent assortment of vehicles to add some variety to the upcoming mobile top down zombie shooter.
How Long it Took -- About 7 hours
What I Learned -- Compatibility between versions of Maya and Unity. Maya 2018 sends textures that Unity 5 can't read. They show up flat white. Maybe I was just doing something wrong though. What I figured out was that Mudbox and Unity work together well, so an easy workflow is to block out in Maya, paint in Mudbox, and export from there to Unity. No problems whatsoever that way. Some of the colors end up looking a bit brighter in Unity than they did in Mudbox, but a little bit of tweaking the metalness and smoothness took care of that.
Also, making a high quantity of low-quality assets can be fun! Lots of room for creativity. I know these look like utterly forgettable little props in some simple phone game, but they do take enough time to create that I like to imagine grim little stories as I paint them. The "damaged" ambulance, for instance -- notice how the back doors are bulging out. That's because a patient who was infected was back there and then turned, bit the paramedics who in turn, turned, and then chaos. The vehicle crashed and burned. The end.
Lore: A bunch of unfortunate vehicles caught in the wrong place at the wrong time with the wrong drivers.
I am following a Pluralsight learning path for Zbrush, but instead of making the simplistic cartoon style alien in the course, I figured I'd do something that really excites me.
The cool thing about the Giger style xenomorph is that it involves hard surface and organic styles of modeling, as well as human characteristics and animal/insect characteristics. I think this project should jump start my skill in sculpting quite a bit.
What I've learned so far :
Note - since this will be such a long project, I'll jot down lessons I learn along the way otherwise I'll forget.
1. I had a lot of trouble getting the hard planar type edges on certain parts of the exoskeleton when starting from Zspheres. Although this is an organic model, much of it is actually more like hard surface modeling, like armor, and so although I want to get familiar with Zspheres, I'm going to start my initial blocking out over in Maya. I think this will benefit me down the line, too, as I will be able to duplicate certain parts more easily in Maya after sculpting in Zbrush. What I mean is the tail -- it's a series of vertebra like parts. I could just build one, duplicate, and then scale them down to the tip. I am sure there is a way to do something like that in Zbrush, but I know how to do it already and quickly in Maya. That is a hard piece to so it will be easier to do with Maya anyway.
2. Blocking it out in Maya worked well. I had a good amount of time wasted cleaning up bad geometry -- specifically with the tail -- but all good lessons learned. Only messed around with sculpting a little in Zbrush but it seems like this base mesh should be good to go. I'll get a lot done tomorrow.
****Make sure to take as much time as necessary to get proportions just perfect before doing any detail sculpting, otherwise you get stuck in an endless loop trying to get things looking right but it just isn't possible without the correct proportions.
3. A few progress updates -- learned how to make custom insert mesh brush, as well as use Dynamesh. And what a great feature Dynamesh is! A lot of creative power in one click -- wow.
And a technical note -- refer to the bottom picture -- for some reason, curve size is only updated if I adjust the brush draw size from the indicated slider. Pressing S or using the quick menu slider will not affect the curve size. I don't think that is meant to be, so I am not sure why my program is behaving this way.
Got a little done this morning. Starting to look like something. Will get in rough detail over the entire model, then Dynamesh to a high resolution and fill in little things.
I'm tired of looking at this guy for today, so it's quitting time. One thing I just learned -- had been Dynameshing at 1024 resolution, which is enough to get the kind of basic detail seen above. I didn't want small things to fuse together when I did the next Dynamesh, so I bumped up the resolution to 2048 (I just use these standard numbers because the guy in the tutorial I learned from did, I don't know that its really necessary).
But at 2048, some of my finer details were still lost, although the mesh is now dense enough that I can't smooth out the kind of basic detail I am still putting in. So, lesson learned, just stay at the 1024 level until I get the entire mesh ready to go up to the next level of detail. Don't be too eager to get in fine little details, but they're just going to get erased when I go up in resolution most likely.
Thanks for the advice. I was kind of learning that the hard way, but since I am working with Dynamesh, if I go too low resolution it sticks some of my smaller pieces, like the vertebrae-like pieces on the tail, together. I suppose that is something I could fix down the line, but its seems like it could be very tedious. Then again, maybe I could just make another custom insert mesh brush and draw it right out in no time...
Getting a good feel for the brushes by now. I find myself using DamStandard, pinch, move, and clay buildup 90% of the time. Ribs still lack the hard bony look I want, probably due to the slight waviness of their edges, but I think I'll wait to go up a subdivision level to really refine those edges. I like the overall positioning and at this point going in that small just moves more polygons than I need.
Simply running the pinch brush along some of the natural edges, like the light protusion on the forearm, is enough to give a nice hard, exoskeleton like look. It looks a bit doughy in the screenshot still, but alive in the viewport those hard places are starting to look how they should -- I think so anyway.
Still to do -- general detail on legs, tail, and back, and need to completely redo the hands. I had to build new geometry as the originals got messed up during Dynameshing. Touch up details on neck area. Then, go up a subdivision level and really refine and add small texturing details. A few more days I think, then it will be ready to paint.
Upper body has most of the general detail I am going to put in. I may go in and add exoskeleton joint ridges along the knuckles, but I don't know if I want to go that far with this model as its just a means to learn Zbrush with. But it is looking pretty cool, so I may spend a little extra time with it. Right now, I am doing all of this detail at 1.7 million polygons. Some area's like the fingers are pretty strung out at the ends of their resolution, but I think it's fine for now.
I don't think I will have to do any more Dynameshing. Shouldn't be adding any more geometry. I like the way the neck tubes turned out.
Right now, in retrospect, I think it may have been better to build the ribs as a insert mesh custom brush -- probably would have saved a lot of time and looked a little nicer. I decided to mold the upper ribs into kind of a solid chest plate, as if the thing was once a humanoid creature but it's slowly becoming like an insect. When I get to the finishing details, I'll put lots of little spikes on that upper chest plate like some beetles have.
Technical notes:
Dynamesh can ruin your geometry if not using 512, 1024, 2048 type resolutions. I don't fully understand this, but basically just use it when you are building your base mesh so that you can easily incorporate major pieces of geometry to your mesh.
Also, while in Dynamesh mode, be wary of accidentally tapping "D", as that turns on a process that will really chug your machine down if you have a high amount of points on screen. If you do, just go into Tool>Geometry>Dynamic Subdivision" and make sure that is turned off.
Athough the point of this project is not to be worried over the art itself, I can't help but get somewhat engrossed with it. I had been unhappy with the way the ribs were looking. You could easily see that they had been pulled from the torso itself -- they didn't look like separate pieces. But I was looking at some really badass artwork of xenomorphs a bit and decided to try and get a lot more of the biomechanical elements in. When I strung some of the coiled tubes between the ribs, it accomplished perfectly the look I wanted. It is a little difficult to see in the photo, but now the ribs appear to rest above a convoluted network of disgusting pipes.
When I go up in resolution soon -- once I finish the tail -- I will attempt to make a slimy, sticky looking skin layer that will be stretched over some of the pipes around the abdomen and neck area. I am not really sure how to do this at all. I think I'll just put little strips here and there and then later, when its time to paint the model, maybe make them slightly transparent or something.
The monster is starting to come alive now. I'm having a ton of fun getting familiar with all the different alpha's now. I think tomorrow I'll give it some teeth. That will really bring out the character I think.
This is a long way from AAA production quality, but for a beginner model I'm pretty proud of it so far. I didn't mean to get so carried away with this as I'm kind of itching to get in and start learning Substance Painter 2, but I'm just having too much fun.
Still miles to go before I sleep!
Earth Golem creature. Not taking a lot of time to make it look nice, just learning the many different techniques for mesh creation, and perhaps more importantly how to best use grouping and subtool layers for a modular, well organized project.
This monster isn't exactly my style, but I think it is looking pretty cool nonetheless. Would go good in a fantasy game like Dragons Dogma. For the most part I am copying the tutorial model pretty closely. One thing I ought to do is make the left arm branches more angular so they appear more deadly. They were created from zspheres, and I'm not sure of the best way to go about doing that. I can go in tediously with the polish brush, etc., which is what I guess I'll do. But I imagine there is probably a better way -- there usually is.
Technical note: If when you grab zspheres with the move transpose tool you find that you grab more than what you want and have trouble moving just one zsphere without its neighbors coming along with it, adjust the size of your brush. As with curves, the larger your brush is the larger the area of influence. Important little things like this never get mentioned in tutorials, but they should! Zbrush has a lot of little nuances like this that make it frustrating to learn, but once you do figure them out they are really handy features.
Also, if you are having trouble rotating the camera the way you want, be aware that when local transformation is turned on, wherever to click becomes to centerpoint of your rotation. So annoying at first but now that I know how to use it I really like it.
Almost done with the blocking in. A few layers of fibermesh moss and and leaves and then the details will be applied. I wasn't super excited about this model at the beginning, but it's looking pretty awesome now. I could imagine fighting it in Dark Souls.
Bobcat Low Poly
Bobcat Hi Poly
High Poly
It looks a bit fat, but that is because I have modeled the bulk of the fur. I will add fur cards around the cheeks and a few other areas which will hopefully make that more apparent, and hopefully I can learn enough in Substance Painter 2 to make him look pretty realistic.
Having a very big fondness for mammals, this was a lot of fun to make.
Took about 3 hours to model in Zbrush, starting from zspheres, then blocking with Dynamesh, then sculpting almost entirely with the clay buildup brush (or BCB, I call it, as that is the key combination to get to it).
To get multiple pieces of a mesh into Substance Painter as separate objects -- like eyes, nose, teeth, etc -- you have to assign each it's own material. Give the material the appropriate name as it will be associated with the mesh in SP.
Also, has to have UV. But will seams matter that much? I guess I've figure this out on the way.
My wife, working up to being a post-apocalypse zombie slayer. Looks a tad creepy at the moment and she's not happy with me. I reassured her it's just the lack of hair and proper eyeballs.
She'll end up with a coat and gloves, as well as a rifle and backpack, and my handsomest looking dog will join her in wasteland scene.
I'm supposed to be working on the Bobcat model, but I needed to go through a Substance painter tutorial first. That tutorial involved texturing a post-apocalypse female character, and as I was working on it I said, "Hey, I'll just make my own model real quick."
Well, real quick has already grown into about 10 hours of work and I've still got a long ways to go.
The face was mostly okay so far because I was working from a reference, but the body I had just been winging. Looking at it, things just seemed off, and man were they. So I took some reference photos and learned how to use this incredible reference system in Zbrush using the grids and transparency. Really nice!
Ignore the Trump hands. I just started forming them.
I can't believe it, this really looks pretty accurate. I can't wait to turn my wife into a zombie slaying badass.
I'm happy with the face. Made some small but crucial tweaks this morning. With this model, I have the benefit of studying everyday, so I noticed a few things like the depth of the eye orbital area that I missed before. You know, having reference images really helps, but it seems with 3d you still can't get everything just from the 2d images. Sometimes you have to kind of imagine how certain parts will be shaped if you cant discern it from a reference image.
It's hard to see well in this image but there is a layer of skin bumps/pores that catch the light nicely and add some realism. I did have a subdivision level at 14 million about, but this layer at half that looks fine so I got rid of that highest level to keep things running smoothly. I think I'll need to take each subtool pretty high to get the details I want.
As much as possible with hand painting and my current level of skill, I'm going to try to make this model look photorealistic. The materials on their now are just because I like differentiation between the pieces, but I'll be texturing this with Subtance Painter 2 when the modeling is done.
Now I'm moving down the arms and hand. The arm has some shape already, but I find it difficult to maintain orientation when rotating around int. The way the muscles wrap around the arm and the very soft, subtle nature of my wife's female arm make it challenging to get just right, particularly in the elbow area.
a few things
generally
-the underlying structure of th for m is missing. specifically the land marks of the human body.
These really help figure out whereyour proportions are goin off the rails, in my experience
-body parts are way too long, that neck is far longer than it should be. I'd go back and check that.
-your proportions of the head could use work.
-the ears are sunken in to her head.
-youre also missing quite a bit of muscle definition.
my best sugestion att his point would be to just go to your first subdivision level, and use your move brush and move around the forms until you get this model looking correct
But, as you have noticed, the neck does seem a big long. The ears and head -- to be honest I didn't put much time into them because they will be covered up. So I expect the problem might have something to do with the perspective differences between the camera and working in 3d or something like that.
*** Edit : Actually, as I was typing this, I was closely reviewing a ton of pictures of my wife. Turns out, she actually just has a pretty long neck. It gets covered by her hair, but I actually measured it out against other parts of her body and the proportions were consistent.
Also, for future reference regarding this model, she does have pretty long arms and legs too. In fact, although I am 4 inches taller than her, her arms and legs are the same measurements as mine.
Also -- thanks for the reference image Paul! I have seen some videos covering these landmark points, but I haven't modeled any humans since watching that video some time ago. So I'm actually going to save that picture and set it up on my second monitor.
My primary resource will still be the actual photos of my wife, and she will be clothed so many points won't be seen, but I think it will benefit my professionalism to kind of have those landmarks memorized.
Got along this far, felt pretty happy with everything. Realized she has no ears.
Solution:
*create an ear and make it an Insert Mesh Brush (and save it for future use)
*Duplicate head subtool
*On duplicate, freeze or delete subdivison levels. If it is very high, freezing and then unfreezing may lock up Zbrush or crash it. In my case, I just had to delete the highest level and then redo those finer details again.
*Create polygroup where you want the ear to go on the duplicated head mesh
*Drag out the Insert Mesh ear and position it as perfectly as you can
*CTRL-click-drag two times and the new ear will be bridged to the group you created. Touch up as needed.
Lesson learned, don't forget shit! And before you go up to do any details, review review review to make sure everything is absolutely perfect.
Moving down to undershirt and visible parts of arms.
I like to use a minimum of brushes. If I'm constantly switching tools I get confused.
What I found is that at a medium to low subdivision level, you can do almost all the wrinkles, large and small, with the DamStandard brush. Just make it big and dig in for the larger folds, then smooth, and small for the harder edges. The details aren't quite done but I think it is looking pretty good. For subtle noise, I just tapped the crumple brush around. Worked as good as any alphas.
Boots seem like something I could sink a lot more time into, but for now these will suffice. I think they would be fine for a game. But perhaps later I'll take some time to make some very detailed boots. I'd like to make the soles and laces very realistic, but this just seems to take a lot of time. Right now I want to keep moving forwards.
Not sure why my screen snips come out so low resolution. I thought they were supposed to be the same resolution as the screen.
Onto the hair, which I am still unsure about how I want to do. I think I will attempt the transparent cards style, just to do something different from sculpting. Also, this character will have very curly hair, so I think it will be better to do that with cards rather than sculpting.
I may try and make a few hanging strands with fibermesh.
Just thought I'd share this hilarious photo. Apparently this is part of the process.
I don't think hair cards will give me exactly the result I want, but I suppose it is possible with any method with prerequisite skill. I'll just have to learn as I go. If I can't get the curls I want, I'll just settle for wavy hair.
Here is the hi-res sculpt completed. Not perfect, but good enough for a first character model. Now I will retopo, make UV's, and then start texturing in the next few days. I will do the hair last, but I may have to redo the hair cards, I'll see when I get there. I am thinking that since I can easily paint the normal channels in Substance Painter, I may just make a few big cards rather than lots of thin ones.
I've done some research and I believe my edge flow should be good for animation, but there is a big high valence vertex on the inner thigh I am worried about. Too tired to mess with it now though.
Will finish the hands and boots later this week, then onto UV's. Ugh. I just want to texture.
Also, I have a feeling the zippers won't come through the normal map baking process in tact. I guess I'll just have to wait and see.
New, low poly model weighing in at 9k quads. I am imagining this as a hero character for a AAA game, so I think that is not a bad range. I've read that many AAA characters are as much as 50k? Not really important for me right now, but just want to make a note of this for future.
I had some success with a proprietary technique (j/k, but I came up with the idea on my own at least) to use a premade generic hand model and "shrink wrap" it around my hi-res hand. I wasn't able to get Maya's shrink wrap deformer to work in this case, but using make live and quad draw I got fantastic results. Made all the time spent fiddling around worth it.
Put some extra geometry around the joints for animation.
This is how I did the hands. Positioned the lo-res over the hi-res like a glove, made the hi-res live, and then used smooth brush to snap the lo-res into place. Was pretty quick, and this way I didn't have to spend too much time fussing over good topology flow. That work was already done.
Spent a long time reviewing other models and figuring out how to get these loops just right, and then routing the extra geometry efficiently. I think this is a good result and I'll be more efficient with this in the future.
Spent better part of the morning trying to find out why my face was unfolding like this. Thought I had some bizarre geometry or something. Finally I found out that the Unfold3d plugin was not enabled. Turned that on, then everything worked. Got the UV's done without any trouble after that, and pretty quickly I think too. Maybe an hour or a bit more.
Not sure about making seams down the cheek lines. I have seen a couple models like this, so I figured I'd try it. It made the unfold easy anyway. I think with 3d painting applications like Substance Painter dealing with seams shouldn't be a big issue. Guess I'll find out.
After the holidays I'll get the ID, normal, and AO maps baked in Maya, then, FINALLY, I'll get to work in Substance Painter.
When all that is finally done, I'll do the hair from scratch. I want to learn how to do pretty well, so I'm going to treat that as a whole project in and of itself.
Although this project is far from finished, I'm already getting excited about my next one. I will do a few smaller projects for a friend first, but I think my next character will be based on this fantastic looking Dani Tribesmen I saw in a Jared Diamond book. That will give me lots of practice with anatomy and wrinkles.
I couldn't find a good reference for what I had in mind initially. But, ca you guess who this is?
I've picked up a few valuable lessons so far. Currently, I'm having to do a lot of retrofitting as I work to bake the normal, AO, and ID mask as I prepare for texturing. Because I am spending so much time fixing up errors I made earlier in the project, I decided just to go ahead and do a reboot.
This is beneficial in a couple of ways. One, I wasn't happy with the clothing, so I will be making some changes and also working on improving my cloth sculpting techniques. Two, I felt I could do better with some of musculature sculpting, as well as just generally making the model more interesting and visually appealing.
A good thing I did was save all the different pieces out as separate OBJ files, so I could re-use them later. Glad I did that.
An important thing I am making sure to do is not jump ahead to any detailing. I wasted a lot of time previously by not adhering to a stricter process of finishing everything completely before going up to the next level of detail. This not only meant that I ended up redoing things, but I also had organizational issues due to too many versions of the same pieces being made.
Pictured is the Version 2 model as it currently is.
Working on constructing a Vietnam era, US style LBE system. Built the pouches in Maya, and just used IMM brushes that come with Zbrush for the straps and waist belt. Looks pretty good I think. Tomorrow I'll sculpt the pouches to get them looking realistic, then I may either do a small rucksack or go down to the boots. Or maybe a hat. Who knows.
I wont be doing the pants though. Tired of doing folds for now.
Technical note: Ended up redoing the shirt again. Purely sculpting in Zbrush kept giving me lousy looking creases and pockets. Lots of waviness and just didn't look very good at all. So I used the body as a live surface in Maya to draw out a new shirt on, and then made the collar and whatnot nice crisp edges. This worked much better.
Also, I took some time to study how people work topology around the shoulders for animation, and although I will end up retopologizing again, it was good to practice that.
And here is an update with some custom pieces I made. I had just been sticking with the freebies that come with Zbrush, but I decided to take some time to make some hardware like the belt buckle and friction adjuster pieces for the shoulder straps. I figured its good modeling 101 practice and I can put them away in a library.
The straps at the junction are a bit ugly. I don't know a good way to manipulate them without getting ugly twist. This character may end up wearing a backpack, but I should still try and figure this problem out.
Also, I wanted to get these backstraps nice and tight looking, so that they look like they are holding weight. They are a bit wavy, but I've given them enough time already. In the future, I need to spend some time messing around with IMM curves so I can get things like this looking just right.
I figure this is about the distance character would be from screen in 3rd person game (which is the prototype project I'm hoping to put this character into). So when I start getting carried away with little details, I try to come back about this far and make sure things read fine. If they do, it's time to move on.
LBE is finished. Would be more interesting if wasn't symmetrical gear on waistbelt, but I will make a similar male character in the future, and if I make some different gear for him then I can swap stuff out later.
Also note that the folds suggest a very soft fabric. I actually just got my wife to wear an old BDU jacket of mine. Should of done that earlier, now I have a clearer image of what the folds need to look like to properly suggest the stiffer kind of fabric.
These pants are not done yet, but already they are looking more like how I envisioned. Instead of using fold, standard, pinch and smooth brushes to make the creases and folds, I just draw the shapes of the folds in with mask lasso, and then use hard polish. Just make sure you know which side is above and below. Works pretty good.
Also, don't bother making laces yourself. That's for chumps. Haha. I just found some laces for free, as well as some eyelets. Ain't nobody got time for that!
For the boots, I watched a tutorial that involved blocking them out with Zmodeler. I hadn't used Zmodeler yet. Seems like it could be faster to work with than Maya, but it is a lot of new stuff to learn and I am pretty comfortable in Maya already. Since I am trying to get this model done so my friend can animate it, I'll just stick to what I know.
Once I get these pants finished today, I'll be almost finished with the hi-res sculpt. I'll touch of the hands and forearms and then make hair cards. I hope that goes well because so far I am pretty proud of this model.
BUt perhaps for the next character I'll try working solely in Zbrush.
12//3/2017
Taking a little time to get familiar with Marvelous Designer. I learned a lot about Zbrush sculpting these clothes, and I am very happy with the LBE, but I think I can do better and faster with Marvelous Designer, so I'll give it a go after a few tutorials.
12/4/17
I am abandoning MD for now. I just can't take the frustration right now. Something about the sewing -- I just fuck it up every time. I am normally saintly in my patience, but something about this just raises my blood pressure to no end. I'm going to finish this character out as is. I'll see what I can do with the texturing to get the fabric looking a bit better, but major improvements will come with the next iteration.
Wasn't no big deal today. Got these cool pants done. Next up, the jacket, then I'll bring them onto the character in Zbrush and be rid of those lesser garments. Marvelous!
Note to self -- sculpt next character model using one of the MD avatars as a base. This will give closer proportions to the clothing and require less tweaking to fit. In the future, see about how to rig my own models so I can use them as avatars in MD.
Also, take this outfit and make several texture variations of it.
Bandana and hair, and the high poly is officially done. Now lets get it ready for a game and texture it too. Can wait to dirty this character up and do the hair/eyes.
A technical note for anybody doing retopo with Maya quad draw.
If you have a highly bumpy area like the rolled up sleeve here, Maya gets confused and wants to drop the vertices down to the nearest point, so if you are running the smooth brush over it the mesh is popping in and out. No way to keep the faces uniformly over the top. So I think what needs to be done is to come back to this area and use soft select to bring the low poly mesh just out along the surface of the high poly, rather than cutting in at some places because the vertices are diving down into the low places.
Oh, and also, when you are using quad draw and you want to extrude a loop -- don't use the quad draw method as it will pull the loop in one direction only. Switch out to your regular selection tool, and in 2018 you can shift click your scale tool to bring the loop out in all directions equally. This means you have to switch tools, but it seems faster than dragging out one side, then another, and another, and then connecting the sides, etc.
OD Green variation. Again, this isn't textured yet, just putting materials on in Zbrush to get an idea.
Retopo so far. Very happy about the face topology, I think it is pretty good. The hands have a few triangles in places that make me nervous, but we'll find out how big a deal this really is. I've been hung up on topology a lot lately. I feel like I'm getting better, but certain parts still frustrate me.
Extra ring around the top of the wrist probably isn't necessary, but the animator can decide for himself. Since this hand will be holding a gun, I figure some extra space around the knuckles to stretch out will be good.
Couldn't find a good way to move around the thumb. But these hands won't be doing THAT much or be front and center, so hopefully this will do.
Face topology is probably too much for my purpose, but I wanted to make it something thaht could be animated as if it were going to be in a cutscene.
My 2017 and 2018 versions of Maya have a bug that won't allow the Unfold3d plugin to work normally. I've spent another four hours this time going through all the suggested troubleshooting with no solution.
Finally I've switched over to an application called UV layout, which is really good but this meant I had to spend most the day learning a new application.
Although the UV's are free of distortion, I do have a lot of seams that could probably be avoided. Most of them are in incognito places, and will be part of a camouflage uniform and thus hard to see anyway. I hope that by texturing in 3d with Substance, these seams won't be a glaring mistake. But the fact is, I am to done with UV's right now to give a damn.
If there is serious flaws, I'll keep this in mind for my next model. And I really ought to go through an intensive character UV layout course.
One good lesson learned though is that if you carefully make the topology on your hands/fingers so that you have a continuous loop running along the insides of the fingers, cutting that UV seam is as simple as one click.
So I had been focused all on the sculpting, and when it came time to turn the mesh into a game model, I had problems. So I am approaching things different now. I am going low-to-high. I've just finished box modeling out this male guerilla fighter character. This will be the game model, and I will do the sculpting from this. I hope that by doing it this way, 1. my UV's will be easy because I'm really familiar with this guys topology by now, and 2. I shouldn't have any issues with the low and high poly meshes not matching up well.
Furthermore, there won't be an annoying retopology stage to do later. Well, we will see how things turn out soon enough.
Finally! It feels great to get over that daunting hurdle of UV'ing and map baking. I still don't really have a solid understanding of what the hell I'm doing in those stages, but these results are satisfactory for now.
This is my first complete (almost, just needs textures) game character.
Update:
Got some of the texturing on the face close to done. Filled in other area's with base colors. Having a lot of fun playing around with Substance Painter. This is definitely the most user friendly application I've used so far.