sdunny

These are awesome.

Fish is correct. To be more specific, the maps are render passes that Mental Ray (our renderer) spits out. They require using the mia_material_x_passes shader. Anything else won't give the proper render passes. You need to have Albedo, Specular, WorldPos, and ViewNormal passes enabled. Our tool does the work of getting these passes into Unity.

We do tune every map. It would not work well to have a universal solution. Most of those screenshots are from a couple months before a pass was done. Dyrford is using newer look up tables that I made.

To assuage everyone's OSHA concerns, that section actually houses a spiral staircase, one with a lip to prevent careless guardsmen from falling to their doom. I'm sure Lord Raedric will keep the health of his employees and proper workplace safety procedures in mind when engaging in future construction projects.

No harm, no foul. Neither you nor anyone else should be afraid of posting critiques about something you feel passionate about.

There is actually very little to be gained from uneven surfaces. Because of the isometric camera, changes in height don't read well. Subtle changes - like rolling plains - are nearly imperceptible. You do get shadows cast, but if it's something like a grassy surface, those shadows would be diffused and wouldn't show the height change very well. If we were to ignore that and still go for some sort of undulating height, you get issues like the one Sensuki mentioned. The navmesh now must meet up with the height changes perfectly, otherwise you get characters clipping through the ground or floating in places. Generally both. This also makes it take far longer to create as we can no longer just work with flat planes to control nav. We also have the issue of making placement of props more difficult. Everything now has to conform to the slope it's placed on and we can't just snap things around to the ground plane anymore. We have to move everything to fit and rotate it to make sure it's at the correct angle. Now we have added time and difficulty to scenes to get results that may not even look particularly good. That's not even accounting for problems like effects that are placed on the ground and would now have to account for situations where you'd cast consecration or something on non-flat terrain. The net result is that you have a very time-consuming process that doesn't add much visually and causes a host of other issues. This is something that we did experiment with early on and the idea was quickly scrapped. The best way to go about it is a sort of terraced system where you have flat areas with ramps or stairs connecting them. This solves a lot of the problems by minimizing the time you spend on an uneven surface. EDIT: I guess I could've just said, "what Adam said."

Small, one room areas can take as little as 30 minutes or so to render. Most of these areas are somewhere in the 8-10 hour range for a render. I think Dyrford is right around that range. Some areas that are particularly dense (lots of reflective materials, etc.) clock in at 15+ hours. We're using advanced materials and lighting calculations and that stuff takes a lot of time. All of the machines we're working on have 32 gigs of ram. Any less and Maya will run out of memory trying to render these scenes. We don't have anything super crazy in terms of processor or graphics card, but there are also a number of tricks we do to optimize our viewports in Maya to avoid chugging. When everything is unhidden, framerate gets a bit rough. I'd say an average scene is somewhere between 50 and 100 million triangles. It varies a lot depending on area size and what's in the scene. The inside of a cave is going to be less intensive than an exterior area with a lot of foliage.

I don't mind answering questions or clearing up misconceptions or any of that. On a project like this - backed by the fans - I think it's important that you guys have your concerns addressed. I generally avoid posting on forums or really anything else online, especially in a professional capacity, but it's not a bad thing to keep you all informed and maybe I can give some reasons that various decisions were made. As for why we would make the game using 2d backgrounds instead of fully rendered 3d, there are a number of reasons. The project leads could better illustrate some of them, but the gist of it is that we're doing it this way because it captures the classic IE aesthetic. That's sort of the base reason behind it, but there are a number of corollaries to this. For one thing, this game is really, really big. To give an idea (although I know it's been discussed by Bobby and others), when we did our team-wide play week, I didn't make it past act II of the game. I wasn't rushing through or anything, but I also didn't do everything, I completely skipped the Stronghold and Od Nua, and I didn't pick up every quest. In addition to that, I already know the layout of most of the areas and where everything is located in relation to everything else. Even with all of that, 40 hours was not nearly enough time for me to finish the game. As has been mentioned, we have a very small art team. There is no way we could have made environments of this size and scope with a team this small in this period of time in a traditional 3d workflow. Maybe if we heavily re-used assets and had a very limited, modular building style. But it certainly would not have had the same scope and variety that we have in the game. There are going to be pros and cons to any development choice that is made, but I think for the results that we've got the benefits have far outweighed the negatives.

To be fair, writing off the system used as 'impractical' does it a bit of injustice. The height map also controls occlusion in Unity (and things like water rendering). If we didn't have this tool, we would have to build out the occlusion by hand using maps or geometry (or something along those lines). That would not be a time-saving option. Building those occluders back in the IE days was no easy task (thank the gods for Scott Everts) and trying to do it using the high resolution images that we have with high-poly objects would be a nightmare. It would easily be a full time job just to make those occlusion masks. When you say that you don't care about the real time lighting or feel that it doesn't add much to the game, that's a fair opinion. It is incorrect, however to assume that by getting rid of the heightmap system that we have that it would be any kind of a time saver. If anything, real-time lighting affecting the background render is a byproduct of the occlusion system.

I'm from Minnesota, so I just appreciate anything that's covered in snow and ice. It is too bad that we were somewhat hamstrung in terms of paintover stuff, but this was a very small project (relative to game development) and we had to be very careful in choosing where we spent our time. I'm confident in the techniques that we've learned and refined and, like you said, there will hopefully be more time for this stuff in a future update. Overall though, I'm very satisfied with the work we've done and I look forward to people getting to see the game as a whole. I know we're not going to please everybody (the nature of life, especially game development) but I think there's a lot to like here.

Thanks for the comments. It can be a little tricky unifying work that comes in from multiple sources (especially outsourced work as they may have different workflows and different ways of doing things than we're used to). The areas around Dyrford were actually the first ones made for the project. The Temple of Skaen was literally the first area that I made that was not just a test scene. As the project went along, we got faster and more experienced at making these areas and that allowed the designers more leeway when it came to how they laid out the areas and what exactly they could expect of the final results. Not every area has had a paintover pass. Because of the sheer number of areas in the game and the very, very small art team, we were not able to paint over every area. Our goal was to make areas look as good as they could in the base render under the assumption that we would not be able to paint over them and then we made sort of executive decisions in terms of which areas we thought would benefit most from a paintover. Most of the areas that have been seen have either not had a pass or only had very minor tweaks. Because of the real-time lighting situation, we actually can be fairly limited in what we can do with a paintover. My favorite IE games, visually, are absolutely IWD and IWD2. Those games still look great and I reference their artwork frequently. A lot of that work, however, was done in a paintover and they were capable of painting things that we cannot. We need to be aware of height and normal information in addition to just the render itself. This can make editing these scenes after the fact even more difficult and time-consuming than it would be otherwise. All of us enjoy working in 2d and it's fun to get a chance to do a pass on areas, but we just don't have the time and manpower to devote to giving every area a paintover. -Sean.

Hey, thanks for your feedback. We've referenced the old IE games a lot during the area creation process. This includes going through and replaying them and having static screenshots up on our servers to look at. We did make some different stylistic choices than were made with those games - a bit more realism in colors, textures, and design is one of those choices. It may be that these choices are not as aesthetically appealing to you (or they may be more so, beauty is subjective like that). Anyways, we've put a lot of consideration into how we wanted these areas to look. There is generally a reason that an area will look a certain way. We also have gone through a lot of places multiple times to improve upon the look and try to better make it fit the area's design needs. In some cases screenshots or video that you're seeing is out of date - either new, improved renders were made, additional effects added, or the scene was given a paintover (or all of the above). We also have been keeping some of the most impressive screens out of the public eye - a lot of the really cool set pieces play heavily into the story or are just something we don't want to spoil before it's encountered in game. That being said, we do take concerns with the game's art seriously. I appreciate any feedback that people have to give and we are working on improving more areas as much as possible before release (for example, working on character/3d object lighting has been a big part of my life the last couple of months). Thanks, -Sean.