Has anything good been done with Atlas yet?

#21

Quote:What are the pros/cons of L3DT? Limitations?

L3DT is great for quickly creating "I don't really care" maps for TGEA. I find it especially useful, since I'm not an artist, for creating maps with a good-looking texture. The free version only allows you to create a heightmap 2k x 2k in size. Considering the Atlas documentation recommends an import setting of around 2 (game units / pixel) that gives you a game map 4k x 4k. Also, the free version limits your texture size to 2k x 2k. So, if you're making an "on the ground" game (like an FPS) with a 4k map, covering it with a 2k texture will give pretty poor results. The Pro version (only $25 US) of L3DT removes all these limitations, allowing you to create much large maps (up to 8k), and massive textures if you use tiling (up to 4M x 4M).

The downside with using L3DT and Atlas is your inability to change the texture and geometry after its in the engine. You can go back and edit the heightmap and texture image, but then you have to go through the import process all over again.

We like the size and performance that Atlas gives us, but we are still depending on that editor to be there before we call our product finished. We are using real-world map data, and are trying to make our game as true to reality as possible, so we don't have the option to say "Let's move our city over here." So, we're going to need that editor eventually.

#22 @Stefan - I too am very disappointed with the lack of editors in TGEA. It is sad that when someone has an opinion other than, "TORQUE IS AWESOME, NO MATTER WHAT!", you are labeled a troll.

I like Torque. Do I think it is perfect, VERY far from it. It lacks BIG TIME in the art pipeline. It lacks editors that are now considered standard in many engines (Constructor should change that). The DTS format needs to be replaced. TGEA feels like it was released just to silence people who said it will never be released. TGEA feels VERY incomplete to me.

I'm sure I will be attacked for writing all this. I just want to make it clear that I DO like TGE/TGEA, I just feel like a fast one was pulled on the early adopters.

#23 There was no attacks either way, and I was the one labelled a "troll" (ironic, since I'm actually a moderator).

Stefan made this statement:

Quote:
I really didn't feel like purchasing another 3rd party application to be able to use the engine which I purchased, (ie. Atlas),

And I responded to that statement. It was a blatent/blanket statement, and received a blatent/blanket answer.

You do not have to purchase any 3rd party tools to use Atlas.

You may wish to purchase third party tools, and/or used the built in flexible import/export system to enhance your Atlas terrains via third party products, and/or simply write a decimator routine and implement real time editing yourself, or simply wait until the editor package is added to the base release.

#24 Ok, no one was calling anyone a troll as far as I saw, so lets calm down.

I don't think anyone at GG can deny that TGEA has less features (as of its initial release) than originally planned. We (the users) are disappointed about the features that were cut, that's all. You had to know there would be some upset people when that decision was made.

At the same time, I work as a professional software developer and I understand how it goes. We released our big new product with a bunch of features missing. I understand that eventually, releasing the software has to be a feature too. We will all be waiting for future updates to the product.

Until then, I must say that Procedural Shader Generation is one of the coolest things I've seen. I don't know a lick of HLSL, but I can write a game that uses its capabilities thanks you guys. Congrats.

#25

Quote:
There was no attacks either way, and I was the one labelled a "troll"

Where?

Quote:
(ironic, since I'm actually a moderator).

Very. Moderators can't troll.

Quote:
And I responded to that statement. It was a blatent/blanket statement, and received a blatent/blanket answer.

Perhaps you should quote the context you are replying to next time so it's obvious what you're discussing.

Quote:
You do not have to purchase any 3rd party tools to use Atlas.

Right, and no examples. You can create applications without a compiler and linker as well, it just takes too long to be practical.

#26 @Mark

thanks for the info on 3DLT :D

#27

Quote:It is 6.5KM island, has 130K x 130K texture, we use a 8KM view distance, no visible repetition, no detail map (at least not in the traditional sense), and is 11MB zipped on disk. It currently runs at about 100fps.

I'd be very interested to hear about the art pipeline you used to create this. Post it here or send me an email, if you feel comfortable. I'm particularly interesting in knowing how you created a texture that large, and how you stream the Atlas file from a ZIP.

#28 @Mark - We're using a combination of Terragen, Genetica, and Photoshop. People tend to underestimate Terragen... it can do a lot and generate great base terrains which you can then modify and preview. Russell has been using Terragen for about a week now and he's generating really great stuff from it.

We're not streaming the .atlas from a zip. It is 51MB unzipped on disk, but when bundled in an installer it is ~10MB 7zip compressed (we use InnoSetup). To us what matters is that you can distribute the game in a reasonably sized download... not that it takes up less space on the users hard drive.

#29 A couple of ideas I will be persuing soon.
1. Editting atlas. For now, the legacy editor works pretty well. It can't do overhangs et al, but it still does 90% of what a terrain needs. So... if current atlas import code is well able to handle 'tiled' input (ala L3DT), and also can convert geoterrain with ease, why can't I just edit a few geos, tell the atlas importer each one is one of the tile segments to read in, and thus have a pretty-darn-close final result? A variable here, a define there, and a coupla mathutils::xx calls and it should work fine.

2. Multiple details. This has to be easy, we're just not thinking it out right. At least for blended terrains, I have one idea I'm sure will work. Atlas2 blended uses one tex with 4 channels to determine the amount of which hi-res texture to apply to that chunk. You can easily see them in L3DT's output window. 4 sucks, but the idea is pretty cool. Okay, so if we already have an index to look at that's going to tell us which texture to use, we should be able to say whichever of the 4 has the highest alpha value (thus being the predominate paint at that spot), is also the ID num for the detail tex we will pass to the shader. Perhaps I'm missing something deeper in the code, because this just seems too easy to do. But then.... most Torque needs are much easier than they seem when you actually digest the code properly :)

#30 Check out EarthSculptor. It can export both heightmap and alpha blend textures that could be used as a basis for Altas terrain. Ideally if there was a shader set up that let you directly use what EarthSculptor creates, that would be excellent.

http://www.earthsculptor.com

#31 Legacy ftw! I've usually avoided Atlas until I could actually do some good heightmap work =P