Saturday, April 11, 2015
Railguns FTW
Thursday, April 9, 2015
Toying with Maya LT - Sticking with Blender
So, I've been working in Maya LT 2016 a bit since Autodesk has offered a 3 year student license and I've been frustrated by the workflow from Blender to UE4. Namely, I couldn't get anything to import properly even after following multiple posts on forums and watching Epic's video on using Blender for UE4. Everything was importing jacked up. Even when I had my model coming in properly, the physics asset was completely wrong, etc. Sockets were attaching at enormous proportions.
After getting Maya LT to properly import into UE4, with about 10% of the asspain that Blender has been, I briefly decided that maybe I'd make the switch to Maya LT full time.
After playing around with it some, I changed my mind. Two reasons. The first being that after playing with Maya's import settings, I got some ideas for Blender's import settings...and found a solution that works. Yay, because I love Blender. The key was scale of .01 for scene units and ASCII export at scale 1.0.
Secondly, Maya just feels a lot like 3DS Max did...inefficient. And it's the controls. Whereas the seemingly impossible task of getting my work into UE4 with Blender was making the tradeoff worth it, it simply wasn't the case once I figured that out.
It's clear that Maya is just as capable and powerful as Blender, moreso probably, even using the LT version. I like the cleanliness of the interface and the hotbox. The problem is that it makes everything that's simple in Blender is so damn difficult in comparison in Autodesk's products. And trust me, I WANT to use the industry standard..but not when I feel like I'm running a race with a pulled hamstring. I'm going to make it to the finish line, but it's going to take me longer.
For instance, having to use the manipulator really irritates me. At first, Blender's system of pressing keys seemed, a little archaic in a system that's used to design visual creation, but I quickly realized how efficient it was to hit A to select all, S to scale and type in 100 on the keypad and press enter to lock in. This is a lot more tedious in Maya.
Also, the most useful tools in Blender are a simple keystroke away. Extrude, bevel, edge loops that auto multiply and snap etc. In Maya, some of these are somewhat easy to use, but you have to interact with the interface a LOT more in order to use them.
Mirroring seems way easier in Blender. I bring in an object, delete one half, apply a mirror modifier and then edit away instantly. In Maya, even getting the object cut it half and put on the axis line was painful. I had to bring in a cube, size it, select the edge vertices, set their position to 0 manually, etc. God forbid you want to change anything once you've mirrored it.
I'm going to continue to use both products, mainly to get some experience with Maya. I'm hoping that eventually its what seems to be less intuitive control systems start to become more logical to me. I've read that it's animation and rigging system is superior, so worst case I can model in Blender and animate in Maya, but I'd really like to get experience in both products with the whole workflow.
After getting Maya LT to properly import into UE4, with about 10% of the asspain that Blender has been, I briefly decided that maybe I'd make the switch to Maya LT full time.
After playing around with it some, I changed my mind. Two reasons. The first being that after playing with Maya's import settings, I got some ideas for Blender's import settings...and found a solution that works. Yay, because I love Blender. The key was scale of .01 for scene units and ASCII export at scale 1.0.
Secondly, Maya just feels a lot like 3DS Max did...inefficient. And it's the controls. Whereas the seemingly impossible task of getting my work into UE4 with Blender was making the tradeoff worth it, it simply wasn't the case once I figured that out.
It's clear that Maya is just as capable and powerful as Blender, moreso probably, even using the LT version. I like the cleanliness of the interface and the hotbox. The problem is that it makes everything that's simple in Blender is so damn difficult in comparison in Autodesk's products. And trust me, I WANT to use the industry standard..but not when I feel like I'm running a race with a pulled hamstring. I'm going to make it to the finish line, but it's going to take me longer.
For instance, having to use the manipulator really irritates me. At first, Blender's system of pressing keys seemed, a little archaic in a system that's used to design visual creation, but I quickly realized how efficient it was to hit A to select all, S to scale and type in 100 on the keypad and press enter to lock in. This is a lot more tedious in Maya.
Also, the most useful tools in Blender are a simple keystroke away. Extrude, bevel, edge loops that auto multiply and snap etc. In Maya, some of these are somewhat easy to use, but you have to interact with the interface a LOT more in order to use them.
Mirroring seems way easier in Blender. I bring in an object, delete one half, apply a mirror modifier and then edit away instantly. In Maya, even getting the object cut it half and put on the axis line was painful. I had to bring in a cube, size it, select the edge vertices, set their position to 0 manually, etc. God forbid you want to change anything once you've mirrored it.
I'm going to continue to use both products, mainly to get some experience with Maya. I'm hoping that eventually its what seems to be less intuitive control systems start to become more logical to me. I've read that it's animation and rigging system is superior, so worst case I can model in Blender and animate in Maya, but I'd really like to get experience in both products with the whole workflow.
Sunday, April 5, 2015
Further Work on Heavy Scorpion Mech
Little bit more work on this today. Did some cutting of excessive polygons and expanded upon some shape. Went ahead and built the basic rig for it, but it's not properly set up yet for animation. Will likely wait for a day when I'm really motivated to get the skeleton setup up with Inverse Kinematics and such.
Haven't figured out where I want the anti-aircraft missile batteries to go yet.
Saturday, April 4, 2015
Spitballing for the Heavy Mech
The heavy mech in my game will be a defensive role mech. Thinking of going with a scorpion look to it. Started work on it a little bit today. This mech will considerably larger than a medium, and will have some armor plates slapped on it at some point.
Has a weapon mount underneath for a railgun, on the tip of the tail, and two mounts on the top of the fuselage. Will also be able to deploy mines.
Wednesday, April 1, 2015
Sunday, March 29, 2015
Running, Jumping, Blowing Up
Obviously still very raw, but piecing together the bits that I need to get stuff working and functional. I don't intend the mechs to vaporize when hit by the slightest bit of damage in the end product :)
Thursday, March 26, 2015
Medium Mech Animated
Welp, got my mech rigged and some basic animation going. Just showing the run sequence here, but have built the walk and jump features as well. Going to use this to start prototyping my game.
Monday, March 16, 2015
Two Mechs - Practice
No guns attached. My intent is to model the guns separately and have them attach on the fly. The guns are interchangeable. The second mech is a little more thug life...but you know, sometimes ugly is effective. Look at the A-10...
Sunday, March 15, 2015
Blender Animation, One Mesh versus Multiple Meshes, and Vertex Groups
Ok, so as to hopefully prevent someone in the future the same asspain that I dedicated my entire Saturday to...
I went to learn how to model and animate this weekend, and created a sweet mech model based off of Alimayo Arango's YouTube channel. He does an excellent job at explaining how to model and rig up a mech, then goes on to show you how to animate it. By the end of his third mech video, you have a nice mech that runs around Blender.
Since I have game development in mind, I recognized I had a problem at that point however. Alimayo's method of parenting each mesh to the bone works GREAT for animation, but it won't fly for most game engine applications. I didn't want 20 meshes floating around, because that would mean 20 UV mappings, 20 textures, and 20 objects in my game engine for each "object". That would be a big performance hit come action time.
So I had to get my 20 meshes back into 1 mesh, without destroying the animation data, but more specifically, the data connecting the meshes to the bones. If you select all your meshes and hit Ctrl-J to Join them up, it destroys this relationship.
After trying about 400 ways to do this by randomly clicking shit and hoping for the best, combined with asking anyone I could find on the subject, nobody seemed to recognize my issue, either due to simply not knowing or me asking poorly.
Then, I strolled across the concept of Vertex Groups.
Essentially, what I had to do was select each mesh I had, one by one, and assign all its vertices to a new Vertex Group of the SAME NAME as the bone that had been deforming that mesh previously. This kinda "locks in" the relationship between those vertices and the bone, which is essential because when you go to Join all the meshes later back into a single mesh, the only way Blender will know how to move those vertices afterwards is if Vertex Groups are established.
So, I built all the vertex groups, named them to the bones, and Joined the meshes. You then select your new, single mesh, and apply an Armature modifier to it using your Armature as the Object. The default settings are fine aside from that.
Voila! Single mesh now working the same as 20 discreet meshes were originally.
I think I still like Alimayo's method of splitting up each section into discreet parts, but there are two ways you could approach the issue with Vertex Groups in mind:
1. Instead of splitting the mesh into separate meshes to begin with, just assign the sections into Vertex Groups to begin with rather than parenting the meshes to the bones.
2. Do it as Alimayo's video has you do it, then select each mesh afterwards, assign the vertices, and combine the meshes into one later as I did.
Now to learn how to use the damn animation editor properly....
I went to learn how to model and animate this weekend, and created a sweet mech model based off of Alimayo Arango's YouTube channel. He does an excellent job at explaining how to model and rig up a mech, then goes on to show you how to animate it. By the end of his third mech video, you have a nice mech that runs around Blender.
Since I have game development in mind, I recognized I had a problem at that point however. Alimayo's method of parenting each mesh to the bone works GREAT for animation, but it won't fly for most game engine applications. I didn't want 20 meshes floating around, because that would mean 20 UV mappings, 20 textures, and 20 objects in my game engine for each "object". That would be a big performance hit come action time.
So I had to get my 20 meshes back into 1 mesh, without destroying the animation data, but more specifically, the data connecting the meshes to the bones. If you select all your meshes and hit Ctrl-J to Join them up, it destroys this relationship.
After trying about 400 ways to do this by randomly clicking shit and hoping for the best, combined with asking anyone I could find on the subject, nobody seemed to recognize my issue, either due to simply not knowing or me asking poorly.
Then, I strolled across the concept of Vertex Groups.
Essentially, what I had to do was select each mesh I had, one by one, and assign all its vertices to a new Vertex Group of the SAME NAME as the bone that had been deforming that mesh previously. This kinda "locks in" the relationship between those vertices and the bone, which is essential because when you go to Join all the meshes later back into a single mesh, the only way Blender will know how to move those vertices afterwards is if Vertex Groups are established.
So, I built all the vertex groups, named them to the bones, and Joined the meshes. You then select your new, single mesh, and apply an Armature modifier to it using your Armature as the Object. The default settings are fine aside from that.
Voila! Single mesh now working the same as 20 discreet meshes were originally.
I think I still like Alimayo's method of splitting up each section into discreet parts, but there are two ways you could approach the issue with Vertex Groups in mind:
1. Instead of splitting the mesh into separate meshes to begin with, just assign the sections into Vertex Groups to begin with rather than parenting the meshes to the bones.
2. Do it as Alimayo's video has you do it, then select each mesh afterwards, assign the vertices, and combine the meshes into one later as I did.
Now to learn how to use the damn animation editor properly....
Subscribe to:
Comments (Atom)