|
Working on a 2.5D Platformer and just started to work on an enemy model and basic animations. Anyone have experience with Blender 2.8 animation to UE4 imports? Or would it just be simpler to import a character rig/mesh to UE4 and animate in there? Managed to import a simple bipedal model but the animations didn't import with it correctly. Tempted to plop down some money on UEfy but I'd like to keep spending minimal. Kortel fucked around with this message at 04:32 on Mar 11, 2020 |
#
?
Mar 11, 2020 04:09
|
|
|
|
| # ? Apr 20, 2024 01:34 |
|
|
Unreal actually just released an add-on I believe, to make bringing stuff from Blender into Unreal easier: https://www.youtube.com/watch?v=c3_xUMQ6hhs https://www.youtube.com/watch?v=UCESgYBphLY Also I have a vague question that's probably better for the CG thread, but reading that thread is like reading a technical manual for some NASA hardware to me, so I figured since I'm trying to make assets for a video game anyways and you guys seem to all use Blender really well too (and I don't think like anyone in that thread uses Blender) that I might as well throw this out there: basically I'm still pretty terrible at Blender, but I still love using it and seeing the slow progress I've made, but I'm trying to simultaneously come up with a 'look' for a game, and one of the first things I thought of was this sort of stringy looking architecture as seen here in the Dark Souls 2 design works book, and I guess I'm asking what the best way to go about achieving his sort of look would be:   I feel like there were more examples in the book, but also paintings by zdzisław beksiński have the very detailed and organic look that I'm thinking of. Would texturing and like PBR be the way to go with this, or would I have to make the mesh in a specific way as well? I haven't really messed with materials and textures, etc. and focused mostly on learning the fundamentals of modeling, so I should probably start to learn how to use Cycles and whatnot.. imhotep fucked around with this message at 06:01 on Mar 11, 2020 |
|
#
?
Mar 11, 2020 05:51
|
|
|
If I was going for that, I'd build a vine and moss generator and let it go to town.
|
|
#
?
Mar 11, 2020 13:39
|
|
|
Former game dev artist person here: For the first example, I would model the overall shape of the architecture, really just the large forms, with no dangly bits, and have a texture that had a lot of vines everywhere. I'd use some alpha cards for the hanging / draping vines, and then some lower poly modeled vines to break up the silhouette of the building and to hide some of the intersections of the alpha cards and the large architecture geometry. It's definitely going to be pretty challenging, and you'll likely have a bunch of alpha overdraw, so performance may be an issue if you have a ton of these hanging vines everywhere. Your second image, with fewer vines, will be easier to model and easier to make performant.
|
|
#
?
Mar 11, 2020 14:35
|
|
|
I'm just gonna recommend Wayne Barlowe as another source of possible inspiration
|
|
#
?
Mar 11, 2020 17:14
|
|
|
 Ok so some work from the last week or so of lunch times on this. After hard coding like 3 cutscenes I decided there had to be a better way. So I set about making this system to direct scenes easily. As you see here there's a heap of arrays sorted by step which triggers an event at the step's time value. It works stupidly well! Atm it doesn't load from an external ini like I want it do because... it refused to? I have no idea why so I just went and got the other poo poo working. Right now I can do stuff like set a sprite, animation speed, movement direction, movement speed, call dialog, and shake the screen. I also have options built in to let it either loop endlessly for wandering npcs who aren't 100% part of a cutscene or to have it just end. Also if you notice by the last few entries I'm only putting in a time entry and an action. That's because today I figured out how to populate the arrays with blank entries so they can be totally skipped over which is gonna save me a bazillion years copying and pasting.  Here's all the features in kind of action. It's not a lot but it's cool to know I've made something that I don't know the limits of at all. Current plans are to add a tag to make the npc jump, to mask the player character and control it's actions, to take control of the camera and to spawn effects Also while I'm here I took the ai from a creature from an older game and improved... it  And also I did a load of stuff with the start menu. There's an intro coming at some point too?  I still have no real idea of what the story of this game is but I'm getting there with the features. There's like... REAL game features in now. FOR A REAL GAME
|
|
#
?
Mar 11, 2020 21:35
|
|
|
This is looking great!
|
|
#
?
Mar 11, 2020 21:42
|
|
|
edit: whoops wrong thread. But thanks for the advice, and Wayne Barlow is pretty wild, he's got a ton of work and even though I'm not as into like super sci fi or 'horror' looking stuff, he's got a lot of very unique stuff with the sort of weird textural stuff that I like a lot. Would you recommend Blender's Ivy-gen add on, or something else. I'm unsure of how I'm going to attempt to go about actually making the game, but for some reason I've made a lot more progress watching some Unreal Engine tutorials and basically copying them and then attempting to tweak it to taste. And I've just had an easier time or I've just found the level design stuff more intuitive to actually do more complex stuff with and overall is just easier to navigate, imo. Plus there's all the quixel texture stuff freely available for UE4 as well, and like all the other free stuff they give away. I think for whatever I do now, I can just get away with hacking together blueprints while attempting to learn C++. That said, regardless of Unity or UE4 it doesn't seem like people ever really import meshes into the engine and apply texture and basically finish them that way, or I guess I'm not entirely sure if there's any hard and fast rules with modeling for games and how to texture them properly and in what programs, etc. imhotep fucked around with this message at 00:21 on Mar 12, 2020 |
|
#
?
Mar 11, 2020 23:50
|
|
|
Haven't posted in a minute; recently became re-infatuated with board game design thanks to some consulting I did recently on one, and managed to distract myself for days at a time. Someone stop me from starting all these projects plz. Anyway! Work still getting done nonetheless. Namely canoes, docks, and inventory.      
|
|
#
?
Mar 12, 2020 03:48
|
|
|
Imhotep posted:Unreal actually just released an add-on I believe, to make bringing stuff from Blender into Unreal easier: Oh man this is great, thanks for the link. I'll give it a solid watch tomorrow.
|
|
#
?
Mar 12, 2020 03:49
|
|
|
Shoehead posted:I still have no real idea of what the story of this game is but I'm getting there with the features. There's like... REAL game features in now. FOR A REAL GAME  I can't help but comment on it every time. Not sure if I already posted about this but my very first gamedev project I ever did was actually an attempt at doing something similar, but uh... that was before I could code or knew anything about gamedev so it failed spectacularly. It just makes me so excited to see updates even if I'm not involved in the project.Bert of the Forest posted:Haven't posted in a minute; 
|
|
#
?
Mar 12, 2020 04:40
|
|
|
Yeah, likewise, y'all have been very helpful when I have silly, vague quesstions, and also even though I mostly lurk in here, you guys are still such a great source of information and I've learned so much from everyone. Plus, seeing the updates for all these games and seeing them go from the 'I just got this rough prototype working' to almost being finished with a game is really amazing and inspiring, and I'll definitely be buying eveyone's games when they come out. Even BYOBattleship, which will be my first naval/battleship/boat combat game ever, but it looks fun and while I just posted about considering art styles with lots of weird organic looking materials I'm also heavily leaning towards the no texture look, but I guess a little more styized, as the genre I'm going for is sort of more reliant on visuals.
|
|
#
?
Mar 12, 2020 08:32
|
|
|
Blender to Unreal add-on is still TBA. Guess I'll have to learn correct import procedures anyway.
|
|
#
?
Mar 12, 2020 12:34
|
|
|
Aww, thanks! I'm just glad I have something substantial to post about finally. I started posting in this thread in...2013? Criminy. And this is the first project I've really been able to dedicate the time to for it to get anywhere. I guess it helps that this one isn't a procedurally-generated Metroidvania.  It's work but it's not trying to solve problems that nobody has solved before (to my satisfaction).
|
|
#
?
Mar 12, 2020 13:01
|
|
|
I'm pretty sure procedurally-generated Metroidvania maps are insoluble by definition. But I appreciate the effort and am secretly jealous anytime I see someone working on them.
|
|
#
?
Mar 12, 2020 18:16
|
|
|
Nah, they're functionally the same thing as procedural dungeons that include locks and keys, and we have plenty of robust algorithms for those.
|
|
#
?
Mar 12, 2020 22:29
|
|
|
Yeah, there are existing games that do an adequate job of procgen Metroidvanias. Probably the most detailed one I know of is A Robot Named Fight. That's why I included the "to my standards" qualifier there, because they all take some design shortcuts that mean that pretty quickly (in a relatively small number of playthroughs) the player will start to recognize structural patterns. For example, in ARNF you always start on the surface, dive down through a series of zones to get to the bottom zone, get an ability that lets you fly, climb back up through the zones, and go to the final zone in the sky. Plenty of Metroidvanias also don't even bother to have each zone on the same map -- you take a teleporter of some kind (maybe glossed as an elevator) so the different zones don't need to worry about overlapping each other. It's totally understandable that these kinds of decisions are made. Map generation is hard and usually you look for any shortcut you can to let you make your game, especially if the shortcut doesn't compromise your vision for that game. Personally though, I feel something is lost if your exploration doesn't involve periodically retreading short sections of places you've previously been, ideally from a connection point that you didn't even notice the first time you were there. Like, you're exploring a new section of map, you drop through a hole in the floor, and come out through a secret passage in the ceiling of an old room. Huh! I recognize this place! I love that kind of thing and so I'd want my map generator to support it, but it makes map generation a lot harder. I do think this is a solvable problem. I've made decent-ish passes at the high-level structure in the past. Where I've gotten stymied is in populating that structure with content. Last time I worked on this I tried to use a variant of the Spelunky approach, filling in room terrain by placing down small blocks of hand-made content that could mesh with each other cleanly. But I was never satisfied with the results, and calculating traversibility (can you get from point A to point B?) turned out to be very hard, especially when one-way routes (like collapsing bridges) were added to the mix. Basically, this is the kind of problem that I could happily work on for years if I had nothing more pressing occupying my time. It might make a fun retirement project if I'm not able to work on it before then.
|
|
#
?
Mar 12, 2020 23:03
|
|
|
TooMuchAbstraction posted:Plenty of Metroidvanias also don't even bother to have each zone on the same map -- you take a teleporter of some kind (maybe glossed as an elevator) so the different zones don't need to worry about overlapping each other.
|
|
#
?
Mar 13, 2020 00:06
|
|
|
A truly procedural metroidvania with a cohesive world that doesn't feel procedural would an insanely difficult challenge. The more you rely on purely procedural content, the less the world feels cohesive and the more you hand-craft/hand-tune the procedural content the less meaningful variation it has. In my mind, there are two reasonable approches to a metroidvania with procedural content: The run-based roguelike approach that Dead Cells uses, or the approach we're taking with Signs of Life where we have hand crafted set pieces layered on top of a procedural world.
Lucid Dream fucked around with this message at 01:09 on Mar 13, 2020 |
|
#
?
Mar 13, 2020 00:32
|
|
|
I mean, all procedural content eventually feels procedural if you have a functionally infinite supply of it because you'll eventually learn to see the patterns. That's why players have a lot of trouble telling when a game that was largely procedurally built but has a fixed number of curated worlds is actually procedural. So possibly the key to making a metroidvania feel not-procedural is to just ship it with the one "castle" and keep the next iterations of the generator for the sequels. Although it begs the question of why you'd need a procedural generator in the first place (to be fair, if it takes less time to built the generator than it'd take to hand-build the map, it's still a good use case. The comedy option is to use the generator to make a second hidden map, inverted castle style, and then a hidden map behind that one, etc. Turtles all the way. Also called the Noita option).
Chev fucked around with this message at 00:56 on Mar 13, 2020 |
|
#
?
Mar 13, 2020 00:52
|
|
|
Lucid Dream posted:A truly procedural metroidvania with a cohesive world that doesn't feel procedural would an insanely difficult challenge. The more you rely on purely procedural content, the less the world feels cohesive and the more you hand-craft/hand-tune the procedural content the less meaningful variation it has. Yeah, I'm not talking "reasonable" though. My ideal for this project was to have a tightly-interconnected world that still gently guides the player along a specific "tour" of that world (while still allowing sequence breaks, of course), that minimizes the impact of "oh I've seen this room before" by varying the contents and significance of the room, and that has a sense of pacing so you get buildups to boss fights / major upgrades, and denouements afterwards.It would still require a ton of handmade content, don't get me wrong. But individual components of all that have been done procedurally. I believe it's possible to tie them all together into a coherent game. It's just a fuckton of work and would not remotely be worth the effort from a commercial standpoint. Chev posted:The comedy option is to use the generator to make a second hidden map, inverted castle style, and then a hidden map behind that one, etc. Turtles all the way. Also called the Noita option). I like your moxie.
|
|
#
?
Mar 13, 2020 01:12
|
|
|
KillHour posted:https://blogs.unity3d.com/2018/04/09/new-plans-for-unity-releases-introducing-the-tech-and-long-term-support-lts-streams/ well, it wasn't and my project is busted  it's going to be fun trying to figure out everything that went wrong. At a first glance I can tell that somehow all the animations got messed up (how??) so the idle animation is now the rolling animation, the jumping animation is now the t-pose and so on. Oh the joys of updating software....e: I guess updating has broken importing from Blender in some way? I looking at the animation import and it seems like every animation is using the same take, when previously they were all set up to use different takes (saved in Blender as Actions). There also doesn't even seem to be an option to choose source take anymore in the animation importer. e2: turns out I was right! They changed how Unity imports from Blender and you have to edit the import script in order to get it to work how it did before. Apparently this isn't something they plan on fixing either, so if anyone else runs into it here's the solution: https://issuetracker.unity3d.com/issues/using-multiple-animation-clips-in-blender-not-all-animation-clips-are-imported-using-a-blend-file e3 because i don't want to doublepost with nothing interesting: on a whim I decided to upgrade my old LWRP test project to 2019.3 just in case the weird inconsistent lighting issues I was having were fixed and it turns out they were! Now I'm a bit torn - do I port all of the stuff I've done on the skeleproject over to URP and use that (long-term probably smart, but might run into other weird issues?) or do I stick with the old default 3D project where everything just works? 
Your Computer fucked around with this message at 07:49 on Mar 13, 2020 |
|
#
?
Mar 13, 2020 03:29
|
|
|
I don't know anything about Unity, but either way, I've been wondering if you were still working on Bonejo Kazooie, and I look forward to seeing updates! edit: speaking of Unity's next update LTS, isn't Blender making 2.83 LTS? double edit: Also every procedurally generated Metroidvania-like I've played has been pretty awful, although I ihaven't played Noita but the appeal with that game is a lot different it seems. imhotep fucked around with this message at 11:05 on Mar 13, 2020 |
|
#
?
Mar 13, 2020 10:51
|
|
|
Imhotep posted:double edit: Also every procedurally generated Metroidvania-like I've played has been pretty awful, although I ihaven't played Noita but the appeal with that game is a lot different it seems. The term "Metroidvania" (like "Roguelike") is getting used in a lot of contexts I wouldn't expect, so maybe your definitions are different from mine, but I wouldn't consider Noita to be a metroidvania (it is absolutely a roguelike). The key for metroidvanias, in my mind, is that you have a freely-explorable "open" world where much of the world is locked off by limitations on the character's abilities. It's a lock-and-key system except the locks are things like high cliffs or impenetrable barriers and the keys are things like jetpacks and explosives. They don't have to be side-view platformers; arguably most Zelda games are metroidvanias by this definition. The best procgen metroidvania I've played is A Robot Named Fight, but there really haven't been many entries in that category intersection.
|
|
#
?
Mar 13, 2020 14:54
|
|
|
Yeah, I was only namedropping Noita because it goes for the comedy option mentioned earlier, that is to say, it has an infinite stack of "inverted castles" (secret hard to access trickier variants of the main map or main dungeon sequence, depending) in each cardinal direction, not because of any actual metroidvania similarity.
|
|
#
?
Mar 13, 2020 16:40
|
|
|
I've got a pretty basic question on material workflow- here's a low-poly thing I banged out to practice UV mapping and texture crap for architecture: While modeling it in Blender I used four mats- bricks, plaster, window frame, and window glass. Am I correct that for games, the recommended workflow would be to generate a single atlas of all four textures (or maybe two, one for the glass and one for everything else), then generate my PBR stuff from the atlas instead of each individual material, like this?  It feels like an unintuitive approach, and my instinct was to generate one PBR mat for each surface type (bricks, plaster etc), but my assumption is that having one atlas and shittons fewer draw calls/object beats out having a smaller number of total materials, but tons of material slots on each object?
|
|
#
?
Mar 13, 2020 21:32
|
|
|
Noita is kind of a secret Metroidvania in that it has lots of areas blocked off by abilities, but since those abilities are randomly generated, the areas they block off are all hidden/not required.
|
|
#
?
Mar 13, 2020 23:03
|
|
|
Omi no Kami posted:I've got a pretty basic question on material workflow- here's a low-poly thing I banged out to practice UV mapping and texture crap for architecture: I haven't been paying any kind of attention to doing this, all of my objects have 2-6 materials on them, and I haven't noticed any performance issues.  I mean, you're probably right that doing that would be more efficient, but at the same time for the kind of game you're making I don't think it'll make a noticeable difference. KillHour posted:Noita is kind of a secret Metroidvania in that it has lots of areas blocked off by abilities, but since those abilities are randomly generated, the areas they block off are all hidden/not required. Yeah, fair enough. Torches/unlimited flight/swimming/digging/ability to turn magma into rock all let you access areas that are otherwise de facto inaccessible, but yeah, there's nothing there beyond obscure bonus stuff and a few small zones to explore. To me one of the key experiences of a Metroidvania is getting a new ability and then thinking about everywhere you've previously been to try to see if any of those areas are accessible now. Which is another challenge for procgen content: making memorable zones to help the player navigate via landmarks!
|
|
#
?
Mar 13, 2020 23:53
|
|
|
TooMuchAbstraction posted:To me one of the key experiences of a Metroidvania is getting a new ability and then thinking about everywhere you've previously been to try to see if any of those areas are accessible now. Which is another challenge for procgen content: making memorable zones to help the player navigate via landmarks! This is one thing Noita DOES nail. The individual levels are all proc gen, but the layout of the world is constant. In fact, that's kind of a standard for roguelikes in general.
|
|
#
?
Mar 13, 2020 23:58
|
|
|
TooMuchAbstraction posted:there's nothing there beyond obscure bonus stuff and a few small zones to explore.
|
|
#
?
Mar 14, 2020 00:00
|
|
|
KillHour posted:This is one thing Noita DOES nail. The individual levels are all proc gen, but the layout of the world is constant. In fact, that's kind of a standard for roguelikes in general. It's definitely common, but is it necessarily a good thing? Like, if you could mix up the zones while still providing a coherent game and appropriate difficulty curve, would that be bad? I get that it gives the player something to expect, but I feel like an important aspect of Metroidvanias is that you don't know what to expect. I guess it'd be weird if, like, Spelunky randomly started you off in Hell instead of the starter caves, but that's more because they baked some of their build-up of tension into the design of the zone, not because it's theoretically impossible that the Hell zone levels could be generated with an appropriate difficulty to start out in.
|
|
#
?
Mar 14, 2020 00:43
|
|
|
Omi no Kami posted:
The other thing is if you're using a shader to change how something appears at runtime, like highlighting an object when it's infront of the player, it's way more preferable to do one 'set the boolean on the shader' call instead of a foreach loop for each material in the renderer.
|
|
#
?
Mar 14, 2020 00:59
|
|
|
So, I've been toying around with making an adventure game of some sort. Lots of choosing where to go and objects to examine and choices to make, that kind of thing. I can code it just fine, the problem is in the script writing. A simple word processor is not quite up to the task. With the amount of individual descriptions and lines of dialogue and branching choices that can pop up in a very variable order, things get messy and impossible to navigate incredibly quickly in something like Word or Google Docs, and figuring out how to organize it all is nigh-impossible. Anyone have any suggestions for programs to use for writing/organizing a script/collection of smaller documents for that kind of game?
|
|
#
?
Mar 14, 2020 02:15
|
|
|
TooMuchAbstraction posted:It's definitely common, but is it necessarily a good thing? Like, if you could mix up the zones while still providing a coherent game and appropriate difficulty curve, would that be bad? I get that it gives the player something to expect, but I feel like an important aspect of Metroidvanias is that you don't know what to expect. I think it is a good thing. A game where literally everything was random wouldn't be any fun - you have to balance uncertainty with an increasing familiarity with the things that are guaranteed. A game where different areas are in a random order is inherantly hard to balance because you don't know how powerful the player will be when they encounter it.
|
|
#
?
Mar 14, 2020 03:12
|
|
|
Metos posted:Can't weigh in on PBR as my game is toon styled, but yeah basically the one material atlassed is the way to go. Won't make a huge difference running on a PC, but optimising for less powerful platforms like the Switch, and especially phones, it will. Is memory a concern? The only potential downside I see to atlasing is that you have a unique material for every prefab, which means you're loading a heck of a lot more materials into memory than if you had like 50-100 mats getting applied to a zillion objects. (Unless materials can't be instanced in memory?)
|
|
#
?
Mar 14, 2020 03:57
|
|
|
KillHour posted:I think it is a good thing. A game where literally everything was random wouldn't be any fun - you have to balance uncertainty with an increasing familiarity with the things that are guaranteed. A game where different areas are in a random order is inherantly hard to balance because you don't know how powerful the player will be when they encounter it. See, the way I imagine tackling this is something like, all of your dungeon components and enemies, for every zone, are scalable. At low levels the layouts are simple, there's few hazards attached to any given level component, the enemies are in their weakest variants, and there's few of them. At high levels the reverse is true. By that method you could generate a smooth continuum of difficulties for any zone. It'd require a ton of work for every single thing that goes into your world, because you can't just say "here's a pit with a platform over it". It's more like code:This lets you do things I think are really cool, like give the player a preview of a later zone, by generating a subsection of the zone with lower difficulty than the rest of the zone does. You could even if you're feeling really fancy let the player see all the really hard stuff through a window or something. It also greatly increases the scope for revisits without it being "well I'm returning to the green hill zone, nothing will be a challenge because I'm midgame-powerful already." Like, Super Metroid had side rooms that you had to come back and visit later and they had way more hazards and difficult enemies than the rooms right next door. There's an awful gauntlet full of acid and spikes right next to the very first room in the game. This kind of thing is IMO crucial for rewarding exploration without it being merely a secret-door hunt. You don't just want to find all the items crammed into the nooks and crannies of the world, you want to be challenged while you do it. If it's just "find secret wall, get treasure, move on to next room" then it becomes rote tedium, and way less interesting.
|
|
#
?
Mar 14, 2020 04:55
|
|
|
Omi no Kami posted:Is memory a concern? The only potential downside I see to atlasing is that you have a unique material for every prefab, which means you're loading a heck of a lot more materials into memory than if you had like 50-100 mats getting applied to a zillion objects. (Unless materials can't be instanced in memory?) I don't know what your specific requirements are, but having a huge bunch of the outdoor environment props in one material is pretty normal.
|
|
#
?
Mar 14, 2020 08:20
|
|
|
Metos posted:You don't need to have one unique material per prefab, they can share them the same way with an atlassed texture. Wouldn't you need a unique mat for each UV map though? Maybe my workflow is totally wonky, but I'm not doing the 2d/sprite thing where the atlas is a neutral sheet you can refer to specific coordinates in- at least for the moment, I'm building them by taking something that has a bunch of materials in blender, and baking everything into a single diffuse texture. It works great on that one object, but it can't be reused- if I want something else to have the same brick and plaster combination, I'd have to apply the materials in blender and bake a new diffuse map. (That's the big reason this feels weird to me, because it feels like no tileable mat like brick or wallpaper should ever exist more than once in my asset folder.)
|
|
#
?
Mar 14, 2020 08:52
|
|
|
Omi no Kami posted:I've got a pretty basic question on material workflow- here's a low-poly thing I banged out to practice UV mapping and texture crap for architecture: You generally want to atlas stuff that isn't supposed to tile and is always going to loaded together, like textures that all go on the same character or prop. Environment textures that tile are probably better off separate, that way the resolution they are displayed at isn't impacted by the size of your atlas (and you can tile them, obviously). Depending on the size of your environments having like 4 separate materials that tile is going to better on your memory use than putting every bit of mapped geometry into several giant atlasses, too. Also keep in mind that generating PBR maps for an atlas that contains different materials is going to give you very inconsistent results (plaster should probably have a different smoothness value compared to brick or concrete for example).
|
|
#
?
Mar 14, 2020 10:23
|
|
|
|
| # ? Apr 20, 2024 01:34 |
|
|
https://twitter.com/byobattleship/status/1238877689261355010 A.k.a. "man I dunno what to do for Saturday, all my work this week was on little stuff that doesn't video well, uhhhh"
|
|
#
?
Mar 14, 2020 18:41
|
|