Developer Log

DL28. I love blocking in levels with basic tilesets. Environment Art pre-production is over, and I'm going into full real-time development with Unity 3d. I have a fairly decent grasp on Blender as a 3d generalist, which has been a challenge after decades with 3dsmax and Maya. From hard-surface modeling, to characters, to lighting and FX, Blender has a decent toolset. I've been getting some light scripting experience in Unity 3d and I believe I have the art pipeline down. . I'm keeping level size reasonable. Unity lightmap baking times for test levels is killing my old i7-7700K. There are five citadels in Eterna, each with three levels. This one will feature commoner architecture.

DL27: Blender modeling+ spline IK rig + rigid body dynamics = ancient Zarrix Behemoth skeleton screengrab. Modern Zarrix are much smaller, and are known to be the storytellers of the realm. Prized for their intelligence, sly wit, and humor, they are a regular fixture in taverns across the realm.

DL26: Temporary walk for the archer. Just enough for me to get him moving in Unity. Next up is a quick and dirty talk emote animation. I'd like to get a dialogue system up and running for testing.

DL25: Getting an archer character going. This model will be the basis for standard set of over 100 human animations. Multiple characters in the game demo will use this animation set. There is something about these old-school animated sprites that make development a blast.

DL24. Add-Ons. A recent Asset Store purchase, I can highly recommend a little gem called Assets Library Builder. Effective asset management is key if you are running a small project (or hell, a big one). Grab it here: https://blendermarket.com/products/assets-library-builder

DL23. An example of an Aegis Mage dwelling. This is definitely a WIP. This is a prime example of how you have to balance a level big enough to accommodate walkable areas for the player and NPC's with larger levels that require a ton of assets to populate....a tall order for a solo art team.

DL21. The Azhem are a species of metallic guardians. They are ancient beings often kept as familiars by the Aegis Mages. They can be found in the wild, but often guard Aegis residences along the coasts. Fiercely loyal and extremely tough, they have made meals of many an intruder.

DL20. It's just a generic chest. In working out the pipeline for this demo, I'm taking a look at what should and should not go to Mudbox. This chest is a good example. From the camera perspective I could probably get away with an albedo and normal map, but I wanted to add a few rough details to the metal frame...like edge highlights, as well as some tactile texture to the metal. Aegis Mages have really nice stuff, I wasn't concerned with adding rough detail to the wood, so only the frame went off to Mudbox for painting. I got my highlights and spent my time wisely. Another quick background prop to sell the environment. A staff holder and potion rack is next.

DL19. Going Too Far. Detail overkill is a thing. One of a few mage table variants I'll be coming up with to populate a simple dwelling. Sometimes you just want to create something because 3d art is fun, camera detail limitations be damned. Next up is a large, solid, treasure chest and a magic staff or dos. Or tres. Lighting not indicative of the final.

DL18. A LithGuard dwelling (concept) at night, cut from the inside of solid rock. Three central pedestals are their resting/recovery platforms. A central pool contains large gilt discs known to be the currency of the ancient Maul. A large stone bed with fur pelt (right) is for visitors, typically commoners or Aegis Mages. Large cylindrical stone containers hold blood gems harvested from the western mountains.

DL17. Interior rendered in dimetric projection, including color tweaks and cleanup. All that is left for this environment is re-rendering and breaking it up for sorting layers. Post that, I'll drop it into Unity for QA. Next up? Four sets of generic furniture for the commoners. After that we start looking at modeling an interior for the next group on our list. Each group will have it's own culture and architecture. Homes will also vary a bit in different biomes.

DL16. 2D Blender render. Old school, BG1 style. Looks a bit too render-y for me, but there is not much in the way of 2d touchup on this one. Still trying to decide between 3d and 2d for this demo. I think prerendered is the way to go, but I'll be doing UV unwraps for all assets in case I ever want to switch to 3d. There is something about 2d that takes me back to the Ultima days. I have also pondered a pixel art game, which would be a blast.

DL15. Screenshot of a small Unity RT demo level. Character is controllable and has walk and idle animations. Current lightmaps baked at 2048 with default settings. Torch is a simple sprite animation with a "look-at" script attached, so it always faces the camera.

DL14. To sculpt or not to sculpt? How much mileage do we get out of sculpting props that appear small within an isometric environment? I don't know, that's why we run tests. A quick and dirty ~2M triangle sculpt and paint of an oxidized copper burial vault in Mudbox vs a non-optimized 270 triangle game model shown in-game (Unity). Could we go higher poly on the game asset? Sure. Is it worth it? Well,,,how many fish do we have to fry again?

DL13. Getting re-acquainted with old-school UV unwrapping in RizomUV. Five different low-polygon vaults atlas unwrapped in one shot. Hand-tweaked stacked draw UVs on the left, with lightmap UVs on the right. Varied pixel space on the render UVs to get the details where necessary while maintaining more than enough space to tourniquet the bleeding. A more sterile, evenly scaled lightmap setup on the right. RizomUV just rocks.

DL12, Still on the fence with making this a pre-rendered game as opposed to a more traditional real-time game. Created a 4500-polygon real-time character and dropped him into Unity with a basic idle animation. Going to add a walk animation and a click-to-move script to test out some basic functionality. Also going to test some animation retargeting.

DL11. Taking a break from character animations to have a go at some UI popup mockups (gee, that was fun to say). Torn between maximum readability (TOP) and more traditional fantasy (BOTTOM). Either way, the prerendered-style backdrops on the windows will be standard.

DL10. Change of plans, combat idle 1 is designated for unarmed melee attacks. Another Blender test to see how animations come together for this pre-rendered project. Given the distance from the camera, it can be difficult to exaggerate secondary motion without entering the realm of say, Street Fighter 2. This is still a little much at 20 FPS. The fewer frames you have to work with, the more subtle the movement for these idle animations.

DL9. Idle animation test @ 24 frames. I had to rework the rig a bit yesterday due to some IK issues, but I think I have something good enough to go forward. There will be a few different idle animations per character, dependent on their current state. Planned states include: combat unarmed, combat melee, combat missile, injured (standing), injured (seated), critical (prone), critical (back), cover (behind, well..cover).

DL8. I gotta admit, Blender's built-in node-based compositor is kinda disco. The Alpha-Over node gives me the ability to adjust my shadow opacity (only) without rendering it separately. This saves me some action/batch creation time in Photoshop, and gets me closer to having a pipeline that goes straight from render to sprite sheet.