IK? 
Would be woke af
1 Like
For R15 to R6, you wouldn’t use IK because breaking joints with R6 is considered normal.
I would suggest it for when the end of the arm is closer to the shoulder than 2 studs. This would leave room for the elbow to bend.
The plugin will be public, so you can add the modification if you want. I still need to figure out arm rotation before I can finish it.
Very Nice.
2 Likes
After two weeks of optimizing, running tests, and more optimizing, here are 256 NPCs running simultaneously:
21 Likes
Wow! Love the character models btw.
2 Likes
Very realistic characters.
2 Likes
haaawwwwttttt 
2 Likes
I’ve worked on something similar recently, I could run 1000 custom r6 zombies with no lag. But R15 was quite tricky. Was hard to run more than 100 of them, just because of all the moving joints with R15. How many polys did you manage to knock down with your custom rigs?
2 Likes
Spent 14 hours learning how to sculpt… still got a lot to learn but, here’s to my first sculpted model!
I have plenty of poly’s to work with, however I do admit I’m not sure where to go to get textures for the finer details. and major details like pockets and buttons
7 Likes
Congratulations looks dope 
! Any video recommendations on sculpting? I plan on learning how to sculpt soon as well.
1 Like
One without @nurgenius in the way >.>
Finishing up the hubworld for that thing I’ve been working on for a while ya dig.
10 Likes
Done another logo for a game. This time I had to build a part of the logo in studio which is the blocky ball so I would be able to render it as a fitting image.
Didnt had a good start though. The template I searched for doesnt look round when I followed it so I had to use individual blocks for me to adjust a bit for myself. In the end it was really worth it and I loved how it turned out!
Tweet
https://twitter.com/PUMPKlNWHITE/status/1021590450770075648
cc: @HowTonyDev @0xfab1e
28 Likes
Looks amazing! I had an idea for a very similar game earlier today… which ig is not happening now. 
1 Like
The poly count wasn’t really the bottleneck here, actually. Mesh instancing took care of that pretty well. Where I started to notice issues was with physics, having all those moving assemblies really messed with that. Changing CollisionFidelity as well as CanCollide on certain parts solved most of that, as well as the NPCs only having one part on the server and the character models being attached locally.
In terms of animations and updating/interpolating the NPCs, there are many distance and visibility based optimizations being made here. If all the NPCs were in view and close to the camera, it wouldn’t be nearly as performant, but realistically that’s not going happen in an actual game scenario. You can see some of that in the video if you look at the NPCs in the distance. They are somewhat choppy, but you don’t really notice it unless you look for it.
2 Likes
How did you go about attaching the characters to the server part? I tried that once a long time ago with welds and things went sour. Something about network ownership getting messed up, I think.
For my system I just tweened the local character model to the server root part.
I have sort of an aversion to CFraming parts because I don’t quite know the expense of CFraming vs. physically simulating parts. If I could weld client parts to a server part, that would be ideal for me.