Sorry your browser is not supported!

I am really curious to hear users experience with PBR in AppGameKit. I know that Game Guru pbr models can be imported via Preben's GG Loader, and also I saw some amazing work by users GaborD and Uzi Idiot (apparently they are both silent with their progress this year)

But what about you? Is PBR important for you and do you have some own workflow to achieve photo realism in AppGameKit?

I'd like to see more PBR in AppGameKit for sure. its where modern shaders, textures, and rendering pipelines are and will continue to go.

I've been working in Substance painter / designer lately and I'd love to be able to be able to run a pbr workflow into agk

Personaly I don't see PBR in a AppGameKit game at the first place. There has to be a lot done before lighting, shadows, reflections and so on are working as they should.
There is also a texture-support problem. No ASTC, and no other then JPEG / PNG for mobile.

Also the draw calls are mostly to expensive.

Where I see AppGameKit is in "low budget", single developer projects, maybe for low-spec hardware. Open GL ES 2.0 Android devices.
Why?

For easy level building there would be a better 3D editor. GameGuru is very very slow. PBR would be more used in a high value production environment and they wouldn't save some money on Unreal or Unity, but invest heavily in developing workarounds for AppGameKit or GameGuru.

I tested PBR out, it was to "expensive" on the hardware. And also the creating cost for thextures and models are much higher, so it would make no sense, to use AppGameKit with the budget for the assets, but not for another more evolved engine.

I am focusing on creating games for Raspberry Pi, the OUYA and also the other microconsoles like NES / SNES / Playstation Classic Mini.

Low-spec ARM single board computers. With not much memory on RAM or the flash rom. So PBR would cost to much in RAM and to much on flash space. I only use the diffuse (stage 0). For PBR we would need mostly tree times the textures, I guess.
Without texture compression, I don't see it.

And to have a realistic environment, you also would need high-res textures. But same problem here, without texture in memory compression ...

Unity


Unreal


Where I would see AppGameKit is more here, it would be great.


So do we need PBR? I don't see the need in the first place, because of the other lacking features in the first place.

Would we need PBR for something like that?


We would need a good editor, or build them ourself, we would need better sound effects support, we would need proper skeleton animation support for Spine / Spriter.

What I am saying is, if you invest in model building tools like ZBrush, 3Dcoat, maybe in texturing like Bitmap2Texture / Allegorithmic Substance Painter / Designer, and or high texture quality, you would not stop at the engine.

High quality audio, DAW, Photoshop, other tools like motion capture and so on, the gear would cost a ton, you would not risk that by using AGK.


... and for easy game maker, I miss LUA support, proper 3D-object loading support. There are problems with some FBX files, with some X-files from other programs, animations are missing, bones could not be too much in count ...

So the question would be more like: What would you do with PBR? Would it be a complete game with more then some minutes of game play?
How do you solve the loading-times or the memory usage?

or is a one-time only test, with the PBR shader from the examples and that is it?

Why PBR? I understand, for the workflow and because of the realistic lighting, but would or could you game benefit from that?
What about physics or character-handling? Automated dynamic animation system? How do you make your character speak?

Or would you code by hand all your assets into your scene? How many things do you get out of GameGuru before you need a massive amount to re-code?

What would your target to achieve with PBR in AppGameKit? It is no AppGameKit bashing, but if you really sure about PBR in your project, you can't really use it in AppGameKit, I guess at this stage. For some testing, maybe a very small world, maybe in some kind of a tool, but you could not be sure, if it's AppGameKit or your model / texture or other things, that do not work how they should.

I guess, you could create a game with only diffuse textures and later change the textures in a batch or something like that.

With some further integration between AppGameKit and GG I think there is a bright future for Indie developers working on their 3D projects. Right now I am using GG as a map editor, I don't find it useful for anything else, and then importing everything in AppGameKit with GG loader. Preben made something already with texture compression, from mobile to desktop, so I don't see some problem related to that. However GG needs a lot more job, I am using iClone as my asset creation tool, where I am exporting fbx geometry and PBR textures in layers just the way GG use them. However GG is struggling with some geometry importing and placing them on the map is most of the time a mess: sometimes geometry is imported, sometimes coordinates are very awkward, and sometimes objects are stuck underground - I am talking here about same geometry that I am trying to import again and again until finally being placed properly on the map.

I want to believe that AppGameKit will be more useful for 3D projects next year, there is already a lot of useful functions in the package, and although it's not ready for AAA games, projects like The Stanley Parable, Amnisia, etc. should not be a problem.

Integrated 3D editor should be really nice to have, AGE3D editor is now on GitHub but unfortunately no one is working on it except creator blendman which is sad. What else is needed is native pbr support like in GG and some visual object/material editor that comes with it. With that few things AppGameKit could be attractive to a lot more creators out there looking for an alternative indie game engine.

So what will be the future?

Windows Only with no native, but user created texture compression?

I understand, that it is complicated, to support ETC, ETC2, PVRTC, DXT and ASTC and that AppGameKit supports uncompressed BITMAP in a memblock only for that reason.

https://developer.nvidia.com/astc-texture-compression-for-game-assets

So why is it a "problem" for me so?
Let's say you really want to use PBR for photorealism, you want to be able to use 4K x 4K textures or higher, and without memory compression, one texture would need 16 Megapixels by 32 Bit colors, so that would be 64 MB. Now we use PBR, we would need the Base color, Metalic, Specular, Roughness, Normal and Ambient Occlusion maybe.
That would be about 6 textures for a nice rendering. So one texture would need ~ 384 MB in memory. How many texture do we need for a beautiful scenery? 10, 30, 60?

With ASTC we could have 8 Bits per Pixel, not 32 Bits per Pixel. Or we could have 2 Bits per Pixel, not 32 Bits per Pixel. So we would have about 4:1 to 16:1 compression.

If someone has a 8 GB memory graphics card, such one would also want to see games like "Shadow of the Tomb Raider", "Assassin's Creed Odysee", "Just Cause 4", "The Witcher"

If someone accepts Skyrim running with 20 FPS on his Net-Book or 2 GB RAM tablet, okay.


Also if it has to be profitable at the end, you want to bring your game on other Platforms like XBox One, PS4, Nintendo Switch, or iOS with PVRTC or Android Oreo with ASTC, nVidia Shield ...

The Nintendo Switch has about 4 GB of RAM. So without texture compression and the need of 6 times the texture, a diffuse map with baked shadows would need, that is about 64 times the need of memory. With the same texture resolution.

And yes. I am working on my own 3D placement editor, because I can't trust GameGuru in any case and sometimes I put a new Prefab somewhere and the complete level is usesless. I also can't invest in LUA in GameGuru, because I don't know, if AppGameKit would support it.
So I have to invent scripting for AppGameKit completly. All that time lost for creating more complex 3D objects.

But if we are not interested in making money with our product, we also could use Unity, Unreal, Godot ...

I understand, that PBR would be great and is something beautiful and I also want to have it in AGK. But to have more stability, reliability in AppGameKit itself or have a 1-to-1 conversion from GameGuro-Projects to AGK-source, would be much more interesting.

I also don't understand, which way TheGameCreators are going. There is MyRPG, I mean, MyWorld, I mean, RPGWorld, there is GameGuru, there is AppGameKit, but you could not use one product and use your creation in another completly. So AppGameKit seems to be the only choice.
But there is also AppGameKit for Android, AppGameKit for Raspberry Pi. But these are not capable for PBR.

So PBR is at this moment a niche in a niche.

Also for Indies ... which Indie games are you knowing, are from a single person or are very popular?

- Minecraft
- Stardew Valley
- Terraria
- Super Meat Boy
- Spelunky
- Dead Cells
- Astrooner
- Owlboy
- Hotline Miami
- Messenger
- Hollow Knight

A lot of them are 2D, okay. So Astrooner is made with Unreal Engine.

And these 2D games are not all made with Game Maker Studio or Unity. They often used there own Engine.

AGK has a lot of potential. But it is not the lead in 2D, it is not the lead in 3D. And it is not the lead in console-export functions.

So is it PC only? Or could it be the best choice for a new game shop on a Raspberry Pi?


@psychoanima
"The Stanley Parable, Amnisia etc. should not be a problem ..."

They are from 2011 and 2010. Half-Life 2 is from 2004, Doom 3 is from 2004.

I guess, it could be done somehow in AppGameKit, but it is not only the editor, it also lacking the tools. We have to write all our tools ourself.
And these games are not using PBR

I would like to see a Doom 3 clone made in AGK2 ready for OpenGL ES 2.0 that was my dream in 2014 to use AppGameKit for the OUYA. Oh, it was a mess. object-loading didn't work very well and bugs everywhere.
So 2018 it would be possible, I guess. But we would need, to create our own "Level-Editor"

And we could use the FPSC Classic OpenSource 3D models and textures for that. We also could use the normalmaps, bumpmaps and so on from these assets. But what would we get? A game in 2019 looking like it is from 2000+
Or we do a 2D game, and showcase, that AppGameKit can do proper "Rayman Legends" like games or games like "Ori and the blind forest". Oh, wait, also for that, we would need a good 2D/2.5D editor. And for a lot of 2D textures also the memory-compression thingly or we do Pixel-Art again.

Windows Only Game Maker Studio 2 is about the same price like AppGameKit 2. But with an visual editor for setting up sprites and animations. Better Spine support.
5% of your income for Unreal Engine after your first 3.000,- $ So is AppGameKit able, to make us that money? Or does it justify the time, we would need for the workarounds without the placing editors?

What I like for AppGameKit is the lightweight. But that also means, no visual editor, no tools like asset compression, optimisation, generation, and so on.

PBR for production? Not in 2019, I guess. You could build your assets, plan your levels, but beginning to integrate your dreams not before 2020 into something from TheGameCreators.

2D, maybe. 3D without PBR like Half-Life 1 or so, okay. It is more for smaller projects, and less assets, I would say.

@Xaby


I totally agree, I was referring here actually to game mechanics (simple, with captivating story) not graphic. As you said later in your text, you can make something interesting in 3D and to look Like Half-Life, but is that really something that players want to try in 2018/2019, I doubt about it. And that's how GameGuru got PBR I believe.

GameGuru is currently "more popular" but only because of twisted assumption from a lot of users, thinking that they can make an actual game (that doesn't look like a mod) without coding. It's sharing same strategy as FPS Creator, but that's the market and I think that's the reason why they are pushing more on model dlc's than on a new features, to satisfy that group of people.

AGK has strong foundation and it needs to go forward to satisfy today's standard in game industry, engine like godot is a perfect example to that.