Steve Streeting @sinbad@mastodon.gamedev.place

I'm getting more and more permissive with my public code, current preference: https://unlicense.org/

It's fitting that this is just an official version of where I started in the 90s, pushing PD code to places like http://xoulu.fi . As always, "greetz" 😉

My new blog on using Instancing in Unity to render enemy health bars in one draw call: https://www.stevestreeting.com/2019/02/22/enemy-health-bars-in-1-draw-call-in-unity/

There was quite a lot of interest in this, so I spent far too long writing this up today 😂 Please boost to make it worth my while!

Me, too often: sitting down at my desk, putting headphones on, then getting engrossed in something and forgetting to start playing the music. Then being asked an hour later what I'm listening to, and trying to think of a way to say "nothing" without seeming like a crazy person

My first instinct was to build a simple combined vertex buffer & params array myself, until I realised Unity's instancing system basically covers this already with a few shader macros. Nice! 🤘

Wondering if I should write a quick blog post about using instancing to display enemy health bars in a single draw call in Unity. A casual search of how others were doing this led to a lot of really bad advice, including using a Canvas per object 😬

Marie has been frustrated at how long it's taking her to make 3D things, but I've been telling her she's effectively learning 3 people's jobs (modeller, rigger, animator) at once with no prior experience. I think she's doing great, her first character model for game 2 is ace.

Just ordered a second book for us specifically on rigging, because it's become clear that the semi-automatic humanoid stuff isn't going to be enough for what we need, and once you need to build your own rigs there's a lot to unpack.

I used to think I understood character animation because I could make skeletons/meshes function in a 3D engine, but ho boy there's so much more to a rig than that (as anyone who has to do it knows)

"Heaven forbid" = "heaven forbid anyone have to see my art", not that being an artist is bad BTW XD

I've been enjoying learning *just enough* Maya not to be an actual artist (heaven forbid), but to be able to connect the dots through to the bits that I mainly work on, and so I can work with Marie better.

I've done this in the past to a degree, mainly with 3DS and XSI, but am going a bit further this time by necessity. A reminder that everyone's job is hard and surprisingly deep

Found a bug, I feel much better now

I just completely redesigned a few components which took me weeks to get working right in their previous (but rather shabby) state.

Put in the last piece today and flipped over to the new system and...it worked. I had to double check I’d really switched it.

Still suspicious.

I've built some standard walk cycles for Humanoid too but they're pretty janky except as tests, Mr Cylinder turned out OK for a quick hack. It's a great case for Root Motion since forward movement is non-linear. Now I'm tempted to refine it and turn it into Purple Tentacle 😃

I think I've finally grokked the cases in which root motion works and doesn't, in both Humanoid and Generic skeleton modes, by teaching myself enough Maya to build test cases. The docs just aren't clear enough on their own

I do hope that the next Blender update works better for me at some point in the future, because it's a great tool and I wish I could get on with it. I'm not holding my breath though because people have been telling me "it'll be much easier to use in the next version" for at least 10 years, and I've learned and forgotten most of it about 4 times 😉

I love that Blender exists but my brain refuses to retain knowledge of it. I've been learning Maya LT and my poor mind seems far more compatible with it.

As only an occasional modeller I was previously happy with Softimage and others, but Blender knowledge always slid out of my brain within *days* and I'd have to constantly refer to notes/videos to the point of exasperation. I have forgotten things in Maya but managed to find them again on my own, which is much less frustrating.

A little game we made over the weekend for the Global Game Jam in about 20h https://youtu.be/xNMF-gUBlK8

Theme was a tricky one this year, most people found it a little bland but we had fun anyway

I moan about noodle graphs, but here's one I *like* (behaviour trees in Behavior Designer)

* structured layout, not arbitrary mazes
* nodes are high-level code blocks; no math expressions built from 15 "add" and "multiply" nodes in tangled spaghetti
* actually useful at runtime

First attempt with Behaviour Trees, my little placeholder enemies now patrol about, chase the player when they see them, predict where the player has gone when they lose line of sight for a second, search the area for a while when they totally lose them, and then return to their patrol afterwards.

The solution was much smaller than I expected although it took a while to arrive at it!