Zwiebelchen

That is ... encouraging, I guess? At least that means there won't be any licensing issues when releasing those tools to the community. Just like Blizzard did with their ArtTools plugin for 3ds max 2011.

Update: Since Obsidian uses Maya to render their scenes and I want to keep compatibility towards possible releases of plugins, I will redo the whole process in Maya.

Hmm... so in case we get the export tools (btw, what are the chances for that? Are they licensed or did you guys code them?), we just click a button in Maya and get all the maps? If so, awesome! I can probably get my hand on a free students license of Maya... I had a lot of trouble tinkering in Blender to get the normal map as they were in the tech demo. I couldn't figure out how to do the depth map yet. As you guys have probably seen, I'm currently writing a tutorial about the rendering process with Blender: http://forums.obsidian.net/topic/70902-tutorial-how-to-render-landscapes-made-in-blender-w-normals-depth-and-albedo/ I still couldn't really figure out how to shift the result into the desired color range. I photoshopped the result to look like in the tech demo, but I'm still not 100% there. Also, the depth map is still confusing the **** out of me; I don't know how to generate that or even why it's like that.

Yeah no problem I plan to edit the stuff in tomorrow. If I come too late I can always ask for deletion and post them again. The paintover job is part of the final render. I didn't do that for the sake of this tutorial (after all, it's just about demonstrating a workflow), but it's possible with tools like GIMP or Photoshop or Adobe After Effects. The real problem is dealing with the enormous file sizes and resolutions. However, this can be fixed by splitting the image up into several sub-images. Due to the orthogonal camera, images can be connected with each other without creating artifacts or seams.

And another placeholder.

Placeholder.

This thread is currently a placeholder for my in-depth tutorial about how to render landscapes built in Blender into the required data maps. Table of contents: UNDER CONSTRUCTION As an appetizer for you guys, here's a quick landscape I built from scratch in Blender, with rendered and corrected normal maps. Depth map and Albedo map will come later.

As people like to post resumes about the art style here, here's mine so far after watching lots of videos and screenshots: Area/landscape design Bipolar. Several landscapes look stunning and crisp, incredibly detailed and overall well-made. Some, however, are bland and boring (especially the main town lacks a lot of love and detail). That flat ground looks weird at times; some landscapes don't feel natural at all because of this. I know that obviously, since the game is 2D, there can't be any real z-information on your heroes position; however, it is still perfectly possible to create the illusion of height via the design of the landscape background texture. Remember Umar hills in BG2, which had an awesome feeling of depth and 3D-ness to it. Interiors are stunning and perfect. Overall, the terrain artists imho did a great job and there is not much to complain about. Some areas lack detail, but that is imho nitpicking. Animation and Spell effects I got to say that spell effects have gone a long way since the earliest backer betas. They have improved a lot and it's great to see flashy light and fire effects going off everywhere. Not much to complain about in terms of spell effects. However, the character animations are still lacking. Especially the missing combat idle animation is a big turn-off. Characters don't even look like they are participating in a battle when not attacking. The complete lack of believable combat stances is a real immersion breaker. A character using sword and shield doesn't even raise his shield inbetween attacks. As those flaws are still there in the latest released trailers, I have my serious doubts we will get good combat idle animations in the official release build. I hope Obsidian will ship them in a content patch post-release. Character models and portraits The portrait style is a matter of taste, really. I can understand some people not liking it, but so far, they seem fine to me. I probably use one for my first playthrough, but then go for custom portraits. After all, we are not forced to use them so not much to complain about here. The character models got a huge revamp since early backer beta releases. Especially in close-up scenes, they have improved a lot. They could use some more polys in the face, but overall, they look acceptable now for a modern non-AAA release. I think we can easily count on the community here to get HD remakes of some textures or models or even completely new designs for those that like more variety. UI and image-events The UI looks functional. Nothing fancy or heavily stylized. It doesn't distract from the game and as long as some elements are movable (like the combat text... I prefer to have it in the center, not on the right) via mods, there's not much to criticize here. The image events, however... I gotta say that I wasn't immediately convinced when the idea turned up first. Especially as it "rips you out of the action", giving you just a heavily stylized painting as visual representation, I feared it would break the immersion of the game. However, it turned out my worries were unfounded. The intermissions look great, the paintings are well done and I like the freedom it allows (especially as you don't need to support any special action with animations or videos). In fact, I could imagine this feature being very dominant in a hypothetical modding environment. Modders could use sketches from deviantArt, using filters to match them with the PoE style and thus create loads of custom events, supporting their mods. Or they could completely take them out of context and use them for completely new mechanics, like text-adventure-style minigames or management HUDs (or options menus for mods). I could imagine using them for strategical maps in tactic games or simply for new party interactions like scripted campfire-scenes. Summary Overall, the art style is fine. Backgrounds are stunning, the unit models are decent. UI is functional, scripted intermissions yield a lot of opportunity for modders. Combat animations are the weakest point of the game. This, imho, should be a main priority issue for a possible expansion.

From what I see now, it looks like there is just a striped overlay over every vertical wall. This means replicating this would be as simple as projecting the striped texture onto all objects in X/Y directions. Am I on the right track here? Still, I don't see the purpose of this yet.

Yeah I agree with Sensuki here; a better Idle/recovering animation is the main issue. Watching the newest videos I just had to shake my head about those... The characters were often just standing there with weapons and shields down, waiting patiently for the next attack. Seriously, did nobody notice that during the production? I can't imagine it being hard to add some "combat-esque" standing animations that override the normal standing animation when in battle. Let the units take cover, shields and weapons up; let dual-wielders duck with their swords in duelling position, etc. ...

So here is a quick demo scene I made from using some free models from the web, which I will use to render the normal and depth map for testing purposes:

Hello everyone! I'm currently writing a tutorial on how to create and render PoE-esque landscapes with Blender (to allow future community projects that add custom user content). However, I'm confused about how the depth map of PoE works, which is vital to create the occlusion effects ingame. This picture is from the tech demo video: http://media.obsidian.net/eternity/media/updates/0079/pe-passes.jpg So, here are my problems: Why does the building on the screenshot have a layered black-and-white striped pattern instead of a steady bottom-to-top gradient from white (ground plane) to black (maximum possible occlusion height)? Also, why do the wall columns on the left side of the screen have a bright white top, brighter than the ground plane, which is clearly behind the wall columns, not in front of them? This is very confusing, imho and makes it hard to find out the correct parameters for the rendering process. Can you please share us some details about how the depth rendering is done (or at least, how the depth map works)? Final, Normal and Albedo map are no problem so far.

I've begun writing a tutorial on how to create and render maps on a 3D landscape built using existing free assets and the free blender software. The workflow is almost finished, I just have to find out the correct values to generate the PoE-style maps. Especially the viewing angle and distance is still hard to reproduce as I have to take a guess here... Also, I'm not sure how the depth map works in PoE; in the rendering tech update, the depth information on the building are vertical lines? This is confusing to me, as I thought the depth map is greyscale and represents the height of an object. I don't see why there should be vertical lines on the building seen in the tech demo. Can a dev maybe chime in and help me on this please? I'm taking my info from this: http://media.obsidian.net/eternity/media/updates/0079/pe-passes.jpg From the Tech demo, everything else is pretty much clear to me: 1) the "final" map is the textured and lit overlay; pretty much the final render of the scene 2) the "normal" map is an ordinary normal map 3) the "depth" map contains the z-height of an object for occlusion... however, I don't get the issue with the building: Why are there vertical lines there? Makes no sense to me; Also, why is the upper part of the wall on the left side of the image white, not grey? If darker means higher, why is it brighter than the walkable ground? 4) the "albedo" map contains the surface textures without shading; it's pretty much what happens if you have a global 100% diffuse light source enabled.

K then ... I'll try to make a blender tutorial about how to create a simple PoE-style map this week.

but no 3d exists of the incredibly unique style of Planescape Torment, I have so much love for it I thought this was about developing a workflow for writing a tutorial... If not, then sure, go ahead. I'm just saying you are making this way more complicated than it has to be You can easily replicate the 2D image with any 3D modeller and get the benefits of being able to render those maps directly. It's just flat ground with rocks after all... Here's a tutorial on generating normal and displacement maps with Blender, a free 3D modelling tool: http://wiki.blender.org/index.php/Doc:2.4/Manual/Textures/Influence/Material/Bump_and_Normal When I get the time, I will make a tutorial on how to build 3D landscapes from existing assets and rendering the maps with blender.

I don't think that starting with a 2D image to generate the maps is a good idea; try to use a 3D modeller instead and use the render tools to create the maps. This is the workflow that Obsidian uses and should bring you much closer to usable results. As a basis, you could use this to extract high definition 3D assets like trees or cobblestones to create a small 3D landscape: http://www.turbosquid.com/3d-models/free-max-model-villa-environments-house/648735 Also, you could extract the trees and buildings from Neverwinter Nights 2 if you need more 3D assets to work with.

Definitely not manually; but there might be a good method to automate this; after all, the game also has to know where to look when it load it's files... and this is rarely done by using O(n) complexity searches. So either there is a defined nomenclature or there's a file management logic based on filepaths. We just have to find out the internal logic. Which I agree might be difficult.

Is there a chance that they just moved them elsewhere (accessable) for performance reasons [and just removed the duplicate]? If so, there might still be access to the missing textures and shaders, we just have to look in other bundles.

MMORPGs usually circumvent this problem by having two variations of each hair model: The normal model and a "knotted" model which is used as soon as there's a model in place that the hair might clip through (like helmets or giant shoulder pads). http://i113.photobucket.com/albums/n201/hk_2/GW2/collar_updo.jpg Quite effective idea imho. It doubles the amount of required models but that's about it.

11 companions on the game files? Makes sense, considering there are temporary companions at certain points. In fact, 2 of them will be part of the prologue already (there's been plenty of vids on that already that state that both companions you have at the prologue will leave the party later on). That leaves us with 9 companions ... of which 1 more might be temporary aswell. Maybe a companion only available for a certain quest? The ominous "Mother" npc might be this. But yeah, I agree that it was a bad decision to take Cadegund out of the game. I liked the concept of her. Wonder why they thought it wasn't a good addition to the game? Maybe because they needed another "evil" spellcaster to balance things out?

This discussion is going in circles. Crowd A: "No, it's complicated, noone will ever be able to do it!!" Crowd B: "But it's a common technique that can be done with any commercial or free modeller. I've done it already and there's plenty of tutorials for this; also, here's examples of community projects in other games that achieved about the same graphical fidelity than what we have in PoE. See, it can be done, it's just a matter of externalization." Crowd A: "No, it's complicated, noone will ever be able to do it!!" Crowd B: "... wait, didn't I just explain how-..." Crowd A: "NO, IT CANT BE DONE!!" Crowd B: "..." Crowd A: "It requires to design the landscapes in 3D and custom maps." Crowd B: "I know that ... I explained that 2 posts earlier and why it's not as hard as it looks. You just LEGO-build a landscape out of public art assets you can find on the interwebs. I've done it already." Crowd A: "But it can't be done!!" Crowd B: "..." So ... before this discussion moves on and we are repeating the same already disproven mantras over and over again; I'm out.

I don't think it will be possible to get access to the 3D models used in the creation of the landscapes (simply because it would be a waste to include them in the game files when they are baked into the texture already)... we will probably only get access to the images themselves. However, there's no reason why we need those assets anyway. There's plenty of free high quality props available on the web. Also, there are "unofficial" packages of models from various other games we could use... if you catch my drift. And if you are willing to drop some money into the creation of new areas, there's always turbosquid: http://www.turbosquid.com/Search/3D-Models/environment/max You can even buy entire landscapes there that easily provide enough area for 10-20 maps, like this: http://www.turbosquid.com/3d-models/3d-mountains-landscape-terrain-model/569910. EDIT: Just found this scene which is completely free and contains lots of environmental elements in HD quality that can be used (trees, stones, bushes): http://www.turbosquid.com/3d-models/free-max-model-villa-environments-house/648735 EDIT2: It's not hard to create your own assets btw... I've done plenty already with varying degrees of detail. Also, it was never hard to create animated assets for WC3, dunno why you say that. There's a giant database available oder user-made wc3 models: http://www.hiveworkshop.com/forums/models.php

Yeah no offense or anything man but adding some UI mods, or creating some basic game functions that already exist in a slightly more efficient way is not a huge deal. DOTA was a mod, but it was a game that already existed and they put 100% of their game out there to mod as people wanted and even had support for it. Blizzard even went on to create it's own code language for modding pretty much all their games. Then you have Eternity. Made on a customized version of Unity which is not the free version which you will never get access to. Possibly including unique code or functions written just for this game that will never be explained, pointed out, or a wiki on what they do and how to modify them. Maps that are designed in a unique way very different from anything existing in any current game on the market, maps that have to be made by hand from the ground up and manually meshed out, not in a "map editor". I have talked to a few people about creating custom UI bars for this game already and possibly including scaling of the UI components already. When asked about the work they made it clear doing UI replacement was not easy.... at all. Actual Scaling of the UI? Basically they would have to code their own scaler from the ground up and build it themselves because the one that exists in engine doesn't actually work and makes everything blurry post scale. That's funny cause every game you talked about modding included functions that allowed the things I asked for easily from day one. Don't assume every game was made to support mods, or that just because it was made using some form of Unity everything will be easy. This game is not "mod friendly" and will never "support mods". It just isn't outright impossible to mod it, big difference. Actually, from my modding experience so far, mods such as changing or scaling the UI or changing core game mechanics are more complicated, not less complicated than creating actual new content from the ground up. I don't thionk we really *need* an editor to code new events into the game. After all, we did that with BG2 already and there never was an editor to begin with. As long as the files are externalized properly, everything is cool. I don't need mods that change the core game mechanics, tbh. I just want the content mods. Those that extend the game beyond what the devs had in mind. Those that add stuff the devs didn't want to add due to priorities. NPC mods are probably the most popular mods for RPGs. Depending on how the files get externalized, this might range from easy to super-hard to make. I'm not showing the WC3 example as a rolemodel for how a "non supported game" got modded. That would make no sense, considering WC3 always had a supported modding community. My example was just to show that modders know their stuff. And sometimes even surpass the devs in certain elements. It was to show that you shouldn't underestimate the technical or artistic skills of a community. So, basicly, there are artists out there that have the skill to design landscapes at the obsidian quality level [and to be perfectly honest, some landscapes from the 1 hour gameplay vid on youtube didn't even look that impressive to me]. The real question is, if it will be possible to use those landscapes and fill them with events and npcs. This is up to obsidian alone and has nothing to do with unity or legality at all. It's just about externalizing files.

Am I the only one who is perfectly fine with 40€ as long as the game provides me with more than 40 hours of gametime? I usually apply the magic formula of 1 hour per 1€ for games I buy. Most AAA games don't even come close to that magic number. PoE, however, has lots of potential to give more. I don't care if the game was funded or not. Backers probably won't retail buy this, as they already have a copy as a backer reward, so why do you guys complain? 40€ is reasonable for a long singleplayer game with replay value, even if it isn't a multi-million dollar production. If the revenue goes straight into the expansions, I'm more than happy to give you my money, now that I know for sure that I get a product in return (backers didn't have this security). ... I just hope we get improved animations in the first expansion. Damn do those anims look retarded. :>

Hah, so here's our ability to create and add new models. And the game isn't even released yet. Awesome!