Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Telgin]

Finally getting to the point that I can start to call my project a game.  I have an enemy that can detect the player, chase, shoot at them, take damage and die.  I think I made the deathbot a little too aggressive though.

Spoiler: Image (click to show/hide)

I wanted to get a slightly better screenshot, but if I stood still for long enough for the bot to get closer, it would toast me.  3-4 hits is all it takes, and it has 2 guns.  I guess now I can at least start to do a little balance work.  The bot probably needs to be slower and do less damage.  The player also probably needs more stamina, since I was tiring out after running from it in pretty short order.

The AI also needs lots of work still.  The detection right now is a simple distance check followed by line of sight check in any direction.  It really needs to be a lot more intricate, with hearing the player, only performing line of sight checks in front of the bot, remembering the last place the player was seen, forgetting the player after a while, and so on.  Working with AI in Unreal Engine 4 is probably the most complicated thing I've done yet, surprisingly.  The behavior trees and blackboards used for AI are a little complicated and interact in even more complicated ways.  At least the blueprint debugger is nice for visualizing what is happening.

And, hey, I also figured out how to do ragdolling on death and how to add particle effects as the robot takes damage.

Spoiler: Another Image (click to show/hide)

That said, in the process I learned of yet more problems with Blender's exported FBXs and the way Unreal handles them.  Unreal can automatically generate physics bodies for imported skeletal meshes, and that's great.  However, you can't manually edit them if they're exported from Blender, or else terrible things happen.  I tried splitting the robot's tail up into multiple smaller bodies so it would ragdoll properly, and the model just started exploding every time the physics simulation ran.

Apparently it's caused by the bone scale being off by a factor of 100 when exported from Blender and is a known issue that nobody knows how to fix because the FBX standard and import libraries are proprietary to Autodesk.  Pretty strange that Unreal can automatically generate the bodies fine, but only freaks out when you edit them.  Oh well, at least the automatic generation feature worked surprisingly well once I tweaked the minimum bone size used to generate the physics bodies.  And editing the joint constraints works without hosing the entire ragdoll, so that's something too.

I'm scared that trying to set up the player's ragdoll is going to cause issues though.  Unreal complained that it couldn't automatically create the physics bodies when I first imported the mesh, and I really don't want to have to throw away the animations so I can rescale the skeleton in Blender, which is the only known workaround to the above issue.

[Solifuge]

Working on a simple Renpy-based Visual Novel for YuriJam 2017!

Don't know how much we'll be able to get done by the end of this month, since we're both pretty busy, but learning Renpy and making something cute and simple has been fun! Even if we only manage a demo, I've learned that I'd definitely like to make more of these some time! It's a fun medium, and I definitely want to experiment with it some.

[Egan_BW]

You're right, that is cute.

[Parsely]

Working on a simple Renpy-based Visual Novel for YuriJam 2017!

-snip-

For anyone struggling to read all the lines:

Spoiler (click to show/hide)

[Mesa]

At long last, a semblance of a Gun Francisco proto-proto-prototype! Featuring no good graphics (all of them are just the same sprite with some color tint thrown on top), shoddy collisions (which is mostly on me not bothering to fine-tune them) and not a whole lot to do!

There are four distinct movement modes (for now), changed with the 1-4 number keys - one controls the blue character only, one controls the red character only, one controls both and one controls both but the red one moves in opposite directions. Currently they're indicated by the text in tne top left corner, but eventually I'd prefer to switch to them being indicated by nice graphics instead, for a less text-heavy UI. But that's far from a priority.

Still, it's better than anything I've managed to accomplish by myself before, and I'm happy to have ANYTHING workable whatsoever.

For the more technically curious, this is done in Godot 3.0 alpha (living on the edge 'n all) and features glorious 144 lines of GDScript code. (effectively Python tailored to the Godot engine), plus some nodes (think GameObjects in Unity) like physics bodies, their collision shapes, and sprites.

My next goal is to have a semblance of a shooting mechanic (currently all there really is console output corresponding to the given character shooting, and at least that works), but I'm struggling with that one, trying to get Godot's scene instancing to work through code (basically something akin to Unity prefabs - I have a "bullet" scene that has the sprite and physics body of a bullet, now I want to make it so that when you press the shooting button, it's instanced at the location of the character you are shooting with, rotated in the direction of the cursor, with appropriate velocities).

It might actually be very simple and I'm just being the world's dumbest game dev, but so far no luck.


Also note that I never actually used Unity "for real" so I don't know if my comparisons between the Godot and Unity terminology are actually accurate (indeed I decided to peek at the Unity docs just to figure out what a GameObject or prefab is).
I did have it installed for a while but the Unity editor is steaming hot garbage on Linux usability/stability-wise (and really damn heavy on top of that - 8GB?!), so I ended up axing it...

[Telgin]

Bump with some more random images.  Maybe later this year when I have more of the game done I can start a dedicated thread about it.

First up, the data center interior that the game's intro will be in.  Lots of experimenting and important things about UE4 were learned with making this, such as how to build an actor that would automatically generate the random server racks using instanced meshes, how to create a material with random color flashing for each server, and so on.  Lots of work left to do on creating content to strew about the area.  It's pretty hard to see in the screenshot since it's so dark, but I also need to do a lot of work on the cabling for the server racks.



A shot from the connecting hallway, showing the buggy interaction highlights that don't center on the object being interacted with.  I think the problem in this case is that the interactions happen above the object's origin, which is the corner of the door.  I'll have to see if UE4 has a way to get the object's center of mass at run time, which I suspect it does.

You can also see some lighting problems here.  The game is entirely dynamically lit since static lighting was giving me fits and also unbelievably slow to boot, but that comes at the cost of me creating lots of "fill" lights that don't cast shadows.  The light bleeds through walls as a result, and it also doesn't bounce like static lighting, giving the very harsh cutoff at the tops of walls where the spotlight starts.  I'm going to ask on the Unreal forums if there is a better way to handle this.



And a shot outside just to show the sidewalk meshes I made yesterday, and to show another lighting problem.  The lamps are having their shadow casting turned off dynamically at run time by the engine for some reason, causing the generic splotches of brightness on the walls for that desk lamp.  No idea why the standing lamp in the other corner isn't apparently casting light at all.  Again, something to ask about in the Unreal forums I guess.



Overall, I guess it's not bad progress for creating most of the assets needed here in a couple of weeks.  I'm hoping that level building will go a lot faster once I've got a decent library of assets built up, but I'm not convinced my methods are the most efficient anyway.  The building was built up manually by me plopping down wall sections and stitching them together with corner and t-section wall pieces for the joins.  I also had to make the external walls a new set of meshes that are snapped into place atop the interior walls, which lets me mix and match materials and wall models if needed, but means that I have to do quite a bit of work to set up an interior building like this.

I could probably build an actor with adjustable extents that automatically set up rooms for me, but I'm not sure how well that would work in practice and so far haven't tried it out.  I'm not sure how to get different rooms to link up together if I did that, since right now corners and wall intersections need to be special models in order to fit together properly.  I might have overcomplicated that, but can't think of a better way to get them to snap together without artifacts or obvious joints.

I get the feeling that most dev studios probably use external tools for building up building layouts and then write C++ actors that can set up the needed meshes from the plan, but that's a time investment I'm not willing to make right now.

[delphonso]

I'm rezzing this thread instead of making a new one. Also, so cool to read through.

[King Zultan]

Why necro this tread when you have your own game making thread?

[delphonso]

I'm hoping to have a wider discussion here, and more specific stuff in our personal dev threads.

Such as:

What if I remade River King? I love fishing RPGs.

What games have good fishing mechanics and what, aside from 'feels like a peaceful place to hang out in', makes a good fishing minigame?

[Iris]

A lot of games I know do timing puzzles for fishing minigames, though I'm not sure that's a good idea (though that might just be because I was always terrible at timing puzzles)

[Mephansteras]

Cat Goes Fishing has some rather good mechanics in it.

[King Zultan]

I'm hoping to have a wider discussion here, and more specific stuff in our personal dev threads.

That makes sense.

[Telgin]

Ah, this is a memory jog.  I did a lot more work on that game above but sadly 2 years of working an hour or two a day didn't get me to the point of having a single complete level.  Boy, making content for 3D games is just so time consuming.  I underestimated how much work goes into level design and building too.

I tried remaking the game as a PS1-style demake too, which did take less time and effort but I was so burnt out on it by that point that I kind of gave up.  Then demakes became trendy...

I've considered trying to remake it in Godot as a 2D overhead game, which I do think would be a more manageable amount of work for a solo developer, but I'm a bit too aware of the level of effort to take the plunge.  Static asset creation would be so much faster if it's just pixel art, but animating the sprites for the characters would be a lot more work...

[delphonso]

A lot of games I know do timing puzzles for fishing minigames, though I'm not sure that's a good idea (though that might just be because I was always terrible at timing puzzles)

I always liked Animal Crossing's fishing - the timing there is only to press a button in the brief window you have when a fish bites. I've never been a fan of the timing puzzle to cast that you see in some games - mostly because it's contrasted almost always by about 10-30 seconds of doing absolutely nothing.

Cat Goes Fishing has some rather good mechanics in it.

I watched a video of this just now, and yeah, I am a fan of this style of fishing mechanic - this is what Legend of the River King has as well, although there's only one fish at a time, and the fish more obviously switches from "time to reel in" to "time to let the fish pull".

I'll keep looking around and see what issues I can see in other systems and if there's anything I can replace/fix.

[EuchreJack]

Just putting this out there: Two people working together on one or more game projects can divide tasks so that they don't all have to know everything.
Ideally, an artist joins as a third member, saving the major cost of paying for art assets.

Also, nobody is tapping into my favorite game type: Ultimate Spreadsheet Simulators. Boxes of Numbers, Hopefully Increasing with Player Help!
Map painters, ala strategy games, are ok too.