@JKoukourakis Fully agreed. Every engine needs to have its specialty.
Though building your own engine still requires a lot of expertise and effort, even with AI. I think it's unrealistic that every team just builds their own.
But "shared" engines quickly end up being somewhat general-purpose.
@SoftEngineer Good job! Are you using sRGB formats for the diffuse textures? The contrast of the lighting feels off a bit. A typical issue when somewhere in the rendering pipeline something isn't gamma correct.
@DeltaWave0x In ezEngine we calculate the AABB only for the skeleton and then inflate it by a fixed value to make it fit the mesh. This way you only look at ~100 positions, not at every vertex. The fixed value could be pre-computed for each mesh (max vertex distance to an influencing bone).
@Dispatch_Graph Dismissal and negativity are unfortunately way too common. We should all be more appreciative of the successes of others.
LLMs are a tool, you still need to know what you are doing.
Termina is very condensed. It is great for learning, where other codebases are often too complex.
Here is a DEBUG build of #ezEngine executing 900 individual, complex animation graphs, each blending multiple animation clips, running well enough to develop in. Performance matters, not only in release builds. And in release it's 5x faster, of course.
@Dispatch_Graph Happy to hear that we have a positive impact. It is also meant for learning (both for ourselves and others). I think by now we are working for 15 years on the engine. It's definitely an open-ended endeavor :D
Release 26.3 of #ezEngine is out!
https://t.co/AZYXZpPMfZ
Among many other improvements, this release adds a water volume component, custom particle shaders, new spline tools, improved RmlUI, and better visual shader editing.
Check out the release notes for more details.
@dangreenheck That's really awesome! I did something similar many years ago, was a fun project (https://t.co/oYbStI04ts). There's definitely a need for more such tools!
@AlientrapGames That's really cool, thanks for sharing! How do you handle shapes that are very fuzzy at the edges? Does the 3D model need to be pixel-perfect there, or do you use pixel discard to make rough shape match the 2D asset?