Last night I took the time to mapped out the sculpting brushes so that I can easily get to it in future experiments. Blender sculpting is fun once you know the shortcuts to each individual brush. By logically organizing them, I’m able to experiment and try out more brushes on the fly.
Here’s another experiment earlier, and as you can tell, it’s based on a reference this time so it looks a bit more real, and not creaturely. In this test, I found myself using the “Clay” brush more, a lot more, because it’s a female head and things need to be round and smooth.
Clay Strips gives a rough look, and it’s a good brush to block in the form in the earlier stage of sculpting. It also gives you the clay effect. Yet at the same time, it can mess up the details that you have worked very hard on, if you’re not careful. When you already have a lot of details in place, use the Clay. If details don’t matter, use Clay Strips. If you want to be subtle, use Clay. I tend to use both the Clay/Clay Strips at 100% strength. At 100% strength, Clay Strips allow me to easily block in the form. I can still manage the strength via pressure sensitivity, but it’s at 100% strength that I use these two brushes. The SculptDraw brush on the other hand is very extreme. I would put both the Blob and SculptDraw in the same category.
I use the Grab brush a lot throughout my sculpting session. Earlier stage to fix the overall shape / proportion. Mid stage to adjust / fix the proportion of the nose, mouth, etc… When the proportion or shape doesn’t look right, the Grab brush is your friend. A lot of time we tend to be so fixated on detailing our mesh that once we zoom out, things look off! This is normal by the way. That’s why the Grab brush will be used throughout. For subtle changes, we have the Thumb and Nudge brush. Thumb and Nudge works a bit differently, but both are there for minor and subtle changes that does not affect other area of the mesh. If the jawline is too high, you can bring it down with the Thumb brush.
More can be said about Scrape, Pinch, and the Crease brush. That’s for another post.
The hand is a challenge to draw, model and sculpt (digitally). Earlier I decided to give it a try, starting with a sphere. I didn’t bother to finish it because it’s way too much work to manually move and reposition the fingers. I enjoy sculpting things manually, and for those that do, Blender really need a way to quickly bend / rotate the mesh—something like the Transpose tool in ZBrush.
Here’s my tip if you want to sculpt the hand: Let the palm face you (front view, “1” on the numpad). That way, you can use your left hand as a reference.
Modeling or sculpting creatures is a great way to learn the 3d app you’re using. Creatures allows you to make mistakes, and to be free and relaxed in your experimentation. With humanoid, you have to keep the anatomy in mind otherwise people can look at it and know that it’s off.
Sculpting this piece from a sphere was quick and easy. It’s also my first attempt at retopologizing. Blender by default has all the options and features that make this easy, but the actual process of retoplogizing (manually) can be long and tedious! I might have to look into RetopoFlow later on.
This time I slowed down and tried to be a lot more intentional with every stroke. I wanted to see if I could find a shortcut to block in a decent head form within 10 minutes from scratch (sphere). Apparently there is a shortcut. The shortcut is the “Grab” brush. After blocking in the basic planar of the head, take time and carefully reposition/adjust the shape of the head with the grab brush.
This looks like I won’t be going back to regular box modeling for organic/creature concept. Sculpt and Retoplogy is where I’m setting my eyes on. If I were to do this with regular box modeling, it would take a lot longer.
Completed in 55 minutes. No reference. Forced myself to improvise. I was a bit nervous before I hit the clock and wasn’t sure if I could pull it off because 1) My knowledge of anatomy is still lacking (the neck gives it away, I think), 2) Still learning what each brush does. Would love to experiment with other brushes, but I stuck to what I know best: Clay Strips. My left hand was hitting a lot of hotkeys and I was going as fast as I could. There were a lot of hiccups in between as I reached the numpad keys with my right hand and that kind of slowed down a bit. I’m a keyboard cowboy and haven’t fully mapped Blender’s keys to my liking yet.
Anyway, I’m very surprised by the result. Let’s just say I got lucky!
I was going to abandon this test piece halfway, but persistence pays off. In case you’re wondering why:
It’s all about refinement, analyzing and correcting, from first to last. The mistake or mark we make becomes a point of reference. To have a nose that looks about right, it must begin by looking wrong. And that begins the refinement process.
I really like Blender 2.8’s Viewport. The wireframe, color and MatCap makes sculpting a joy. I was going through my old sculpting files back in early 2016 and imported them into 2.8’s beta just to see what they look like. And speaking of sculpting, here are a few thoughts:
At the time of doing these experiments or test back in 2016, I had no idea that I would end up with these results. I was just playing around with Blender’s GUI, exploring its features etc… and then sculpting came to mind. I had no experience whatsoever with digital sculpting or even traditional sculpting. Didn’t know what DYNTOPO was and how it could help with sculpting from scratch.
What most people usually do first before entering sculpt mode is this: They block out the form with proper subdivisions and everything. Wireframe looks clean and so on. If the resolution isn’t high enough, then you can’t sculpt in tiny details. You would have to increase the resolution of the mesh [throughout] the entire model. Things can slow down if you don’t have a fast computer, which is why a lot of hobbyists give up due to the lag/slowness.
That’s where “DYNTOPO” comes in. The entire mesh will be triangulated. In Dyntopo mode, Blender will add or remove local resolution as you sculpt (depending on your dyntopo setting). Overall, it’s a lot faster! And you don’t have to think about topology or subdivisions of your mesh. In my opinion, DYNTOPO is what makes digital sculpting fun!
You can take out a cube or any primitive object and start sculpting right away. Here was a cube I did yesterday. It was not planned. I scribbled around with the Clay Strips brush and really enjoyed the pixelated look of the mesh. Sculpting in dyntopo mode gives me the illusion that I’m working with clay. As you can see, it wasn’t that difficult to block in the face.
The first hurdle is software related. Technicality. The second hurdle is artistic. How do you go from a Cube to a finished piece of art? I’ll touch on that in my second post.
Blender has come a long way. I like its Sculpting mode/setup and this was a test to relearn the Sculpting brushes. Started off with a SPHERE.
The main sculpting brush I use is the “Clay Strips.” It’s what I used initially to block in the form/details (image: left) and it’s also what I use throughout the Sculpting process. I prefer this brush over the “Draw” brush. The “Pinch” / “Crease” brush also come in handy for refining/sharpen edges (image: right).
EDIT: Another test. I’ll probably do a lot more digital Sculpting in Blender later on.
EDIT2: Another test. I can’t seem to resist the temptation to get into Sculpting at this moment, but I must resist and focus on my 2d studies! 🙂
Blender sculpting is extremely addictive once you get a hang of it! After the first test (Skull above) I thought I could take a break from it, but did two more tests. It’s that fun! Again, my primary brush is CLAY STRIPS, 95% of the time.
Game development is more democratic than ever.
Indie game designers are putting out awesome new titles that are being consumed by gamers all over the world. A lot of these games use low poly modeling—a simple design concept that will always have a place in game design—even as graphics and performance continue to get better. However, there are a lot of moving pieces that go into designing and making a game work. It’s easy to get lost and overwhelmed. Here are five low poly pitfalls to avoid. (Link to Article)
In its simplest form, the Shrinkwrap modifier is tasked with snapping the current object onto the surface of another object. It also has the ability to only snap specific vertices if you specify a vertex group… I think the Shrinkwrap modifier was probably first created as a retopology tool, the snapping allowing you to easily create new, low-poly geometry, over your high-res model, without having to constantly think about manual snapping. (Link to Article)
Though some of the above methods did warrant you some creative freedom during the process of meshing, however, not all the methods are efficient when there is a strict requirement for you to model after a reference with precision. Therefore, I would like to share with you an alternate approach, which might be handy to you when building something that required high-level of exactness. The technique that I am about to discuss here was largely originated from the good old fine art of Silhouette Paper-Cut, as shown in the following screen grabs. You can read more about the used of Silhouette as an art form via Wikipediaif you are curious about it. (Link to Article)
I was experimenting with the Skin Modifier and in my opinion it’s not perfect. For some things it can be quick and fast, but if precision and good looking forms is what you’re after then you will have to play around a bit until it looks right. Which can be a bit time consuming since you don’t have total control over areas that bend, like the armpit for example or the neck is too short. To fix the arm, I had to change it to a different pose (T-Pose). I guess the whole point of using something like a Skin Modifier is to get the rough shape or form so that you can get right into sculpting, which is great for creatures modelling/sculpting (not saying that you can’t do that for characters). For me personally, I still prefer the old Box method. Maybe what you can try is combine both methods? Box and Skin Modifier. I find the Skin Modifier makes it a lot easier to block out the arm and hand. Perhaps you can do that and attach it to your Box form?
Brush Demonstration (BD for short from now on) is a short series that I’ll be putting together to show how each brush in Blender’s sculpting mode is used. I am basically learning and discovering them as I go, and today it’s the Inflate/Deflate brush.
Earlier I was trying to give more definitions to my Thumbnailling experiments last night, and for this particular head I noticed that it was difficult to add form and volume to this flat piece of mesh/polys. Blob and Clay brushes won’t do it! With patience, you can actually get the look that you’re after, but why when the Inflate/Deflate brush can do it much quicker? When using this brush I imagine that I’m blowing air into a flat balloon. Very useful when you’re trying to solve cylindrical shapes/forms such as the arm, leg etc… if they’re too flat or thin then simply Inflate them! Too fat? Deflate it.
In this demonstration I used only the Snake brush and Smooth. That’s about it! The last sculpt in the video wasn’t sculpted entirely with the Snake but I used it just to show the Snake brush’s functionality.
Think of the Snake brush as an Extrude tool with the ability to Move and Rotateproportionally all in one brush. Now that is cool!
For me personally, when it comes to humanoid and creature modelling I can visualize and improvise as I go, even with very little references. This is because I see faces and body shapes everyday in the real world, but when it comes to hard surface modelling—let us use Iron Man as an example here—I run into problems (though I haven’t really given hard surfacing a try yet) and have a hard time visualizing. It takes skill (I think) to decode a blueprint, so for me, a front and side view aren’t enough! I believe the best way to improve in this area is to really get a hold of, for example, an actual Iron Man figure. The best references are the ones that you can touch, feel and can see from all angles.
I was at a toy store today looking for a Stormtropper but couldn’t find the head and ended up taking a few shots of something else.
Edit (March 3rd): Just found out that Blender 2.77 RC has an improvement for the Snake Hook brush: “Improvements for snake-hook that allow you to drag out long extruded segments, and rotate the snake-hook brush using the new rake option.” Didn’t know about the rotate option and haven’t tried it in any of my experiments yet.
The head with horns was sculpted mostly with the Snake Hook and Smooth brushes alone. Form sculpting is a great way to explore sculpting brushes.
Was going to save this post for a later date, but since I touched on N-gons yesterday here it is. One detailing method that requires n-gons is this:
Point-by-Point modellers have these things (flows) taken care of from the start. Whereas for Box modellers, they focus first on the shape and form (Extrude and tweaking), and once that is out of the way, they start the refining process—laying out the topology. I’m getting a little ahead of myself here, but speaking of laying out the topology this way, uv unwrap’s wireframe of the face (hint hint) can be used as a reference to practice laying out your topology in this way since the map is 2d.
Another topology method is Extrusion and thinking in term of groups or zones. Something I’ll talk about later on.
Currently skimming through Blender For Dummies, 3rd Edition by Jason van Gumster and came across this helpful thought on N-gons. Definitely something to keep in mind if you’re new to the whole thing:
“…it’s best to think of ngons as a process tool. On any mesh that’s likely to be used in animation (like a character model) or included in real-time environment like a video game, the finished mesh should be composed of only tris and quads. An exception to this rule of thumb might be for architectural models or models intended to be rendered as still images. Because those meshes won’t be deformed by something like an armature or a lattice and they don’t have to work in a game engine, often you can get away with leaving ngons in them.” — Jason van Gumster (Blender For Dummies, 3rd Edition)
What does it mean to think of ngons as a process tool? Simply this (my understanding of it): with n-gons support, you’re able to do things a lot quicker—especially when it comes to cleaning up the mesh to redirect flows and so on. Back when Blender didn’t have n-gons it was a little tedious to change flows on the mesh, but now it’s a lot easier and quicker. So in this context, n-gon helps a lot when it comes to Box modelling! With Point-by-Point modelling you don’t run into issue with n-gons.
Another quick test started with the good old Cube. I am truly impressed by Blender’s sculpting capability. Main Brushes used to block out the form: Blob and Snake Hook. Brush for defining the head and shoulder: Clay Strips. Used Blender 2.77 RC.
Here’s a quick test I did earlier (roughly 5 minutes?), starting with the good old Cube. In total only three brushes were used: 1) Blob, 2) Snake Hook, 3) Smooth (that’s it!). From what I understand, the first stage of sculpting is the form and shape. In Blender, the Blob, Grab (does not add geometry) and the Snake Hook (add geometries) brushes are there to aid you in that process.
The difficult part of starting out is not knowing where to begin! What tool do I use? What’s the setting? With traditional sculpting you don’t have to worry about the settings, but with all things digital, there’s always a technical side to it. I was going to create a video demonstration of the tools but the screencast app I’m using is demo. I’ll do a lot of demonstrations later on once I purchased a license for it, so wait for it.
Note: This was a fluke! I didn’t know that I could block out the form with a Snake Hook on a Cube, but thank God I played with it because now I know something. I was thinking to myself, “How on earth am I going to block out the head? Ok let’s pull this one out, wait… oh! This is amazing.”
If you’re just starting out then there’s a chance that you’re using references (I hope so!). Not only that, but you’re trying to match what you see. Then to your surprise, things aren’t lining up no matter how hard you tried! If you get the front view correct then the right view (profile) will be out of proportion. If you fixed the right view then the front view will be screwed and it’s an endless cycle of going back and forth. That was my experience with blocking out the form of this head with a head references in the viewports. What I learned from this is that it’s best to approximate. Don’t aim to match your references 100% perfectly. Instead, see references as guides and don’t be afraid to exaggerate or even go outside of the reference(s).
NOTE: Will fix the videos when I have a chance. Apparently it doesn’t work on iOS. I put together these videos using QuickTime player recording. Now I have a better screen recording app which will work on iOS. — March 04, 2016.
This is an example of how I approach it in Blender. In the picture and videos, what you’re looking at is the top of the nose flowing to the mouth area. There’s a thought process behind it which I’ll talk more later in the future (God willing).
Texture artists don’t like it. Rig artists don’t like it, and a lot of people just hate it. So the best thing to do is just avoid it! Modelling for yourself is different, but when you’re working with other artists it’s best to play by the rule that they all go by: No n-gons.
Since I am modelling for myself and experimenting, I can be flexible in this area. Sometimes I’m just lazy to clean up the mesh or haven’t decided on where things should go, so I end it with an N-gon to get back later. Cleaning up N-gons requires a bit of thinking, depending on your modelling experiences, but it’s very easy once you know what you’re doing.
As of this writing and from what I know, I don’t think there’s such a thing as a “one topology to rule them all,” and you won’t know how good a topology is until you start to get into deformation of the mesh (animating it). Not to mention that all modellers are different in their approach and each mesh has a unique flow or structure to it. However, there is a general guidelines that all modellers follow, such as the mouth and eyes areas and so on. If we get the part for the mouth and eyes setup correctly then the rest will slowly fall into their proper places as you try to piece them together. How you go about piecing them together will be different from how others go about it.
One of the reasons why I’m currently experimenting with creature’s heads is because it’s uncommon to see people post topology overview for monster’s heads. So I want to train myself to see topology overview for a varieties of models (human, monsters, creatures etc…). Right now I’m just guessing and learning as I go.
Digital modelling is both creative and technical but don’t let that turn you off!