ozz

@lawnjelly Wow, applications for these are endless. I am going to track that paper.

Wow. Super exciting. Designer babies was always on the cards, but the ability to change DNA in adult organisms is awesome. This has the potential to cure genetic disease and cancers, which is particularly useful if it happens to you.

A bit disappointed this didn’t get more attention. 😅 I was trying to announce that the FREE #NintendoSwitch port for @godotengine now supports the latest stable version: 4.3! 🎉 Let’s spread the word far and wide! 🤖💙✨ #GameDev #GodotEngine #FOSS

WTF! Did Linus Tech Tips they just used godot source code compilation to benchmark CPUs? This is so cool @godotengine 🤯 youtu.be/DWYFfzHFtM8?si…

@BCvitanic @reduzio I agree with you that for 2D, Godot is the big shot right now, but for 3d... that is a different story, Unity( for mobile games) and Unreal (for PC games) are the top engines without discussion

@reduzio I mean for 2d godot is probably the best option out there right now, but what about 3D? I would be interested in the split of 2d vs 3d for published godot projects, since you posted a picture of itch.io stats recently aswell. My hunch is that it's mostly 2D

@JyveTech @reduzio I would love to see a native terrain system implemented, too. 🥲

@reduzio A. Search filters on the 'reparent node' window. 2. Clouds node. iii. Integrated simple height based terrain node I can throw noisemaps at and it'll give me smoothed scaled terrain hundreds of miles wide. Easily sample points to add multimesh flora on later. !that's all.

A few more anoying issues to fix and my landscape system #MadeWithGodot is going to be open for testing. The extension editor is not perfect but it is usable.

@lawnjelly Oooooh, so I just need to update the llightmap module then? What do you think about the possibility of porting it to GDEXTENSION? Is it doable or do you think it might be a headache?

@OzzRC0 Nope, couple of reasons: * We are in feature freeze already for 3.6 and plan to build next beta next week hopefully, then RCs. * This is an update of LLightmap, not the core lightmapper. (It's of course possible there could end up being some improvements to core lightmapper.)

Ambient Occlusion - managed to get rid of a lot of the artifacts caused by tris that only just touched a texel. Now if the coverage < 1/4 for the sub-texels, the texel will be reclaimed, not calculated, and filled in with dilation.

@lawnjelly Are you going to add it to the 3.6 officual release?

@OzzRC0 Yup.

@wojtekpil @DYarncat In godot I have seen frame drops around 90K to 100K vertices on mobile .

@DYarncat Yeah but I think this require a bit higher poly meshes. From what I see vertices are still expensive on mobile platforms so I would be cautious with this method.

I have been working with lightmaps in #Godot and I was pretty annoyed by the "flatness" (directional lightmaps are too big for mobile). Solution? Really simple, for some surfaces you can just bake light into a texture itself (look at road at left side). Done in #MaterialMaker

@Saulcava1 I stream mipmaps on demand so there is no need to have the entire mip chain in memory. For texturing I use a very simple virtual texture, a Clipmap Texture.

@OzzRC0 how you approached streaming with that texture feature needed

This is not a pretty picture, but what you see is a 64x64 Km long terrain made with #GodotEngine. This is about 20GB of offline data splitted in 4096 regions, being streamed at runtime. The first load of the scene is painfully slow (more than 1 minute), I need to check the cause

@wojtekpil It looks amazing! What specs has the phone?

After a month of work I want to show a bit of progress of my #Godot game, targeting mobile platforms. Still looks a bit rough around the corners, but probably cool enough to show. It's quite simple concept to limit the scope. Running on Android phone using compatibility renderer.

@passivestar_ I still prefer using an outline screen shader. This will also duplicate the geometry being drawn. however, it is a very useful method to highlight small occasional quest objects

Did you know it only takes a few clicks to add an outline to a 3d object in godot?

@SpockBauru @dwarfses @TokisanGames I would suggest using the geometry of the mesh to remove the unnecessary probes. something like generate a shape and instantiate the probes only if they are inside the geometry.

Thanks to @dwarfses and @TokisanGames my light probe tool now uses GPU Raycast. This means that colliders are not needed anymore, if the camera can see, it should work! Get LightmapProbeGrid v2.0 on Godot AssetLib: godotengine.org/asset-library/… #GodotEngine #Godot #Godot4 #gamedev

@TokisanGames You are on the right track. I use a file per region for my landscape system following a suggestion made by @drugbite a while ago, kind of how the Unity Terrain system works. I am looking forward to the day that our extensions are ready. @drugbite ,@TokisanGames Great Job!!!

Re: Terrain3D design After discussing with my Out of the Ashes team, I'm now convinced multiple storage `.res` files is the way to go, one per region, with foliage transforms stored inside. Eventually we'll support region sizes other than 1024k, and more than 256 regions (16k^2), up to 2048 (45-90k), the opengl limit for texture arrays on many cards. Separate files will allow us to get under the 1gb limit for res files (#GodotEngine currently crashes above this), finally allowing us to save 16k terrains. But the biggest benefit is for team development. With only a 2k terrain in OOTA, we already have challenges handing off the baton to edit the terrain between me and my level and environment artists so we don't conflict on git. I thought about making the foliage instances its own resource file so one could paint foliage independent of sculpting/texturing. But the reality is an environment artist needs to sculpt, texture, and foliate simultaneously. It would be insanely impractical if we had a 16k terrain and several more people working on it, where one dev has the entire terrain checked out. Plus it does suck to upload the entire terrain data, when you've only been working in one small area. Currently I'm implementing a paintable foliage instancer. The next big thing I'll work on will be separate files per region. There will be an automatic upgrade process for your data as there has been on every change since 0.8.

@daihaminkey [2] Have a proper target renderer on mind from the beginning and develop the landscape for it, that will give you better performance than trying to target all backends at once. Right now my system runs on every backend but at the expense of sacrifices.

@OzzRC0 Sounds insanely cool. Building my landscape system to at the moment. Is there any chance you cloud share sources of yours? Or maybe recommend good books or other references?

Here is my #GodotEngine Landscape System running on android and using the compatibility renderer. All settings at max. In a real game, I would hide most of the grass and use a lower tree density and, of course, lods.

@daihaminkey I would recommend first looking at the existing terrain systems and using those as a base. The hardest part sometimes is getting the terrain rendering running. If you find something that already works, dont reinvent the wheel. For biomes, look at Horizon Zero Down GDC talks

@lightshaftdev The editor includes a full set of tools for editing the terrain and also supports importing external maps from terrain generators. Currently, it is only fully available in the module version, since migrating it to gdextension is more painful than I originally thought

@OzzRC0 Sorry if you already answered one question like that on other comment, I didnt take a look yet, but I have one question: Did you pretend to add terrain/landscape edit tools inside the godot editor? Or your system will only generate the terrain?

@mediapire It uses c++ and it can be compiled as a module and as an extension.

@OzzRC0 Very nice! Did you use GDScript and/or another language using GDExtension?