A Breath of Fresh Air - a Grimrock 2 Update

December 13, 2013|

Hello everyone, it’s time for another status update! Features and content keeps on pouring on in the game at a steady pace and the point where the entire game would be playable from the beginning to the end is sooo close: there’s just two or three levels (that I actually started working on today) that need to be done to make the game “whole”. Of course, the game is not by any means finished at that point but reaching alpha is a mighty fine milestone nonetheless! Other gameplay things we have also tackled recently were prototyping some new spells and adding a completely new character class. We’ll keep the new class under covers for a while but once we get it to a more complete state, we’ll tell you more about it!

The intro cinematic to the game is now pretty much complete, barring a few finishing touches on the soundtrack, and Juho has already gotten a good headstart on the outro cinematic as well! The full motion cinematics are a huuuge step up from the still images we had in Grimrock 1 and I hope the players appreciate the sweat, tears and love (well okay, mostly the sweat & tears) Juho has poured into making them.

Talking of eyecandy, we finally tackled the issue of how we make the skies in Grimrock. The traditional approach of using a static skydome texture doesn’t really cut it in a game like Grimrock 2 so we had to figure out something that wouldn’t take forever to create and that would still look good, both in a still image and in motion with a dynamically changing time of day. Petri came up with a hybrid approach (as seen on the pair of screenshots above) of using a code generated sky and sun together with a few layers of relit clouds. The relit clouds mean that the cloud textures are not just plain cutaways from photos, although that was our starting point, since we have separate hand painted data for highlights so that the clouds react convincingly to different lighting scenarios when they drift across the sky and the position of the sun changes. While this approach we used doesn’t provide hardcore super photorealistic results, the looks of the sky are actually somewhere between painterly and realistic: a perfect match for the looks of our game.

Oh, and our buddies (and almost next door neighbors from our Matinkylä-days) from Theory Interactive popped in to say hello the other day and to give us a hand with the leftover beers from our christmas party ;). They just passed the 20,000 Eur milestone of their indiegogo campaign for Reset and we snapped a celebratory photo with members from both of our teams. Their time traveling robot FPS puzzle game looks glorious and the crowdfunding campaign definitely deserves to gather so much more money than it currently has! They should also have a new video coming up by the end of the week too and I can’t wait to see what they have in store!

Alright, I suppose this is it for this time around! If there’s any particular “making of” -style article you would like to see, let us know (via blog comments/forum/facebook/twitter) what interests you and we’ll see if we can do something about it! Thanks. 🙂

14 responses to “A Breath of Fresh Air – a Grimrock 2 Update”

  1. Marinus says:

    Dear Grimrock,

    Look very good. I really cannot wait to get my heands on this game. I have a question: is there a Special Edition in the making ?? I love to get my hands on an artbook :).

    Kind regards,


  2. Darklord says:

    Awesome, really glad to see we’ll have outdoor areas this time to! 🙂


  3. jlcarneiro says:

    Great news! Thanks! Can’t wait to put my hands on Grimrock 2!

    I have little question, though: does one of you speak Portuguese? On the second paragraph you used “outro” instead of “other” just like a Portguese speaker would do…

    The intro cinematic to the game is now pretty much complete, barring a few finishing touches on the soundtrack, and Juho has already gotten a good headstart on the outro cinematic as well!

  4. Raggie says:

    Thanks for the update! I’d love to get more insights into the world and lore you must be fleshing out since the game has been moved outside of one dungeon.

    • takis76 says:

      Yes outro is the other meaning of the finale. Or the video you see when some video game ends or you kill the final boss, or you just complete the game.
      Intro is from introduction and outro could be from outroduction but there is no outroduction word in the English language , but is a geek term for programmers or game developers. 😀

  5. Bek says:

    I have some workflow questions if you don’t mind 🙂 Mostly modo related.

    In the older blog post ‘Building the Dungeon’ you talk about how you export the bricks from Zbrush to modo to arrange them; then you build the lowpoly around this. Have you considered using Zbrush’s 2.5d canvas for this task? In some cases it might be quicker (probably not for bricks, but for shapes that don’t neatly tessellate, like cobblestones). From the 2.5d canvas you can capture a normal map using a matcap (Although the default zbrush one is incorrect; google should lead you to a better one—unsure if this comment lets me post links). It’s also handy to save out some other materials to be layered when working on the diffuse. You can also output the heightmap from Zbrush. This is also where the recently released Knald (knaldtech.com) comes in handy—feeding it a normal/height/photograph generates all the maps you could ever need in a few seconds, like AO, cavity/convexity, height… Much much quicker than baking individually in xNormal.

    The only advantage I can think of for arranging bricks in modo is that when you make your lowpoly you can conform the mesh around that, and you can build the UV’s without much distortion. If you were extruding geo based on a tileable texture that wouldn’t go so well.

    Oh and since you mentioned xNormal; are you aware of the PipelineIO plugin for modo? It is a huge timesaver for exporting meshes. In one click I’ll export my lowpoly with automatic hard edges by UV seams, and then the exploded mesh if necessary for baking + the smoothed cage to match. Much easier than copying your model to a new scene each time. You can google this to find the latest version on the lux/foundry forums, it’s made by Ariel Chai (svartberg.com). I made a google doc for modo related workflow stuff where I list some other handy scripts: https://docs.google.com/document/d/1H4Mv-9Zryg-Vn-5cA2g7eqIwLbiJD5ELHcjgGcPN1so (assuming links work)

    What are your general thoughts on modo for game art? There are definitely a lot of features lacking/missing compared to other apps, but scripts solve a lot of these problems and with the modo ‘steam edition’ I can only hope modo’s next revision addresses a lot of this game art stuff. What I’m personally missing most is decent tools to straighten UV islands. By edges just doesn’t cut it.

    Oh, and since I’m being greedy with questions… any tips for doing the diffuse of your bricks? I assume you use a mix of combining hand painted layers with photosourced layers and blending until it looks sexy? When Substance painter is released texturing will hopefully become a lot easier—and since it runs on linux, and might do away with photoshop, I could possibly move completely to linux. Interesting times indeed.

    • Jyri says:

      Hello Bek!

      We do use zbrush 2.5d canvas sometimes like you described here but not that often. This is because of how zbrush handles the canvas export. It’s really difficult to avoid getting a seam on the top&bottom edges. This is a bummer since the best use for the 2.5d canvas would be creating tileable maps. The other problem with 2.5d canvas is that it’s a destructive workflow in the sense that once you drop to 2.5 you can’t go back. (without getting seams that is, there is a zplug for converting 2.5d back to geo)

      What we do instead is use the wrap options for brushes and export highpoly to xnormal where the lowpoly geometry is fitted to capture exactly the right area so that the maps tile automatically.

      We don’t use svartbergs PipelineIO altough it’s great. We have our own script for exporting.

      Modo is pretty great for game art. It’s got probably the best selection tools out there and it’s easy to setup normalmaps for the viewport etc. Also Modo’s Pixar Subds can make hardsurface highpoly modelling a lot easier. But like every software, Modo’s got its share of nasty problems too. Creating alpha test/blend assets can be difficult at times because of how bad the alpha sorting works. Modo’s got some nice scripts sure, you should check out farfarer’s scripts at luxology forums. The vertex normal toolkit is a must have for game artists!

      About diffuse maps: you guessed right.

      • Bek says:

        Yeah Farfarer’s vertex normal toolkit is great (edge-based smoothing gives you much better control and really, makes much more sense), as is his Texel Density kit. That plus Etera’s UV tools speeds things up quite a bit. Of course Seneca’s scripts should get a mention too (although there’s so, so many to sort through).

        I’ve had problems with seams from Zbrush’s 2.5d canvas before too; but they are avoidable. Seams can occur if you ‘bake’ in the layers palette (Which you usually discover too late—as you mentioned, destructive workflows can be problematic). There’s also a plugin to export the cavas free from other odd seams (http://www.zbrushcentral.com/showthread.php?94752-Useful-small-ZScripts-and-Macros-for-ZBrush-4-amp-4R3&p=848052&viewfull=1#post848052 ). Wrap options is something I’ve seen used on polycount recently and it looks interesting, but I haven’t tried it myself yet. Although they still do a grabdoc and crop rather than exporting and baking. Or wrapmode 1 and no cropping.. probably easier if I just link the thread: http://www.polycount.com/forum/showthread.php?t=126224 Cool to try other workflows though.

        Yeah the modo viewport is something that I see mentioned a lot (and having to set normal map opacity to 50% is just weird); I usually use marmoset toolbag for previewing assets with textures though so I don’t run into alpha sorting problems much. Do you use edge weighting with p-subs often? Or do you just mean p-subs are better in general? P-subs definitely seem to handle triangles better. I found edge weighting to be unreliable so I stick to edge loops, which are package agnostic too which can be handy.

        Maybe when Grimrock 2 releases you guys could consider doing an art dump on polycount, would be great to see 🙂 And thanks for the speedy and awesome reply!

  6. BouH says:

    I’s always a plesure to hear good news, and all news about sinishing of Grimrock 2 are good. I have one question about storyline. In Legend of Grimrock we learn the plot through dreams and notes and i really love that idea, it gives deeper dive in game for player. But in open world game it may be impossible to push story that way. So my question how game would lead us throught it’s plot?

    • petri says:

      Log2 is not exactly an open world game where you explore large towns and vast areas of wilderness. The main focus are still puzzles, exploration, creepy crawlers and a story which is not forced upon the player with NPCs and cutscenes. We think that the tile-based dungeon crawling genre works better this way.

  7. willow says:

    Legend of Grimrock 2 will be available only in englisk like the 1 or will be translating in other languages with french and german?

Leave a Reply