Drunk On Nectar - A Nature Sandbox Game

Thanks as always !

It is the same dream! I totally get your idea and would love to see your concept go further. Some day I want to have mammals in DoN too but realistically unless I first complete a simple prototype with a few invertebrates, plants and ecological phenomena, I can’t even dream of moving on to even reptiles, amphibians or birds. However, if the gaming community likes this game I am willing to go all the way.

My game is already magnified 5x though! Would it still benefit from tilt shift? The thing is, the mountain isn’t setup properly (it needs to be scaled up 5x) and that may be messing things up.

Here are some in-game screenshots to give an idea of DoN’s world scale. The Jumping Spider in my GIFs is actually a tiny dot on the foot of the Unreal mannequin (the mannequin itself is scaled up 5x) or just a tiny speck on the underside of a leaf (second screenshot).

BTW - I thank you for the detailed feedback, it is deeply appreciated. I’ll do some more reading on tilt shift later today. If it’s anything like the “perspective illusion” trick that some movies use, it may end up making my creatures look even smaller than they already are though! :slight_smile:

Thanks for considering doing a video on IKinema!

Please feel free to pass questions to the team here, we would like to help by answering specific questions!

I honestly know very little about both the tilt shift effect and game development/UE4, so I wouln’t want to guess how the 5X scale would affect things. I am an artist mostly, learning now to develop games. (Im 30) Tilt Shift seems to encompass more than what I was thinking:

*edit5 (holy ****) I keep reading comments and people seem to be confused. Tilt-shift is NOT a miniturizing effect. You can achieve this affect with tilt-shift, but that’s not it’s true/main purpose. People do not buy large format cameras for miniturizing. The tilt-shift action has many more applications than that. Some actions of a tilt-shift camera can be emulated on a computer, but those types of adjustments are very hard to achieve correctly (you can easily tell the difference between a photograph shot with a real tilt-shift vs one edited in post process)
*
https://www.reddit.com/r/pics/comments/1caw85/incredible_tilt_shift_photo/c9evvea

Maybe something similar but simpler to do…
Here is what I was thinking basically: I took your image and blurred the stuff that was the farthest distance from the camera. Keep in mind this doesn’t do what I’m talking about justice really. I just quickly blurred the background in one blurring. You would want closer things like some of the grass to be somewhat blurred. And you would want to tweak stuff to see what looks best.
Imgur

&d=1454156791

Basically the farther things are from the camera, (in terms of distance in the world) the more they are blurred. You would have to set a distance (I am guessing how it would be done here) from the camera where things would START to be barely blurred, and from there things would be gradually more blurred. You wouldn’t want anything too close blurred, because it would effect game play.

You mentioned that you have concerns about your creatures looking smaller. I’m not sure about this, but it doesn’t SEEM like it would be a problem, in fact I am pretty sure it would HELP. The creatures will take up the same exact amount of screen space, so THAT at least wont be an issue. And the reason it may help is because the blur (even if you do less blur than I did) automatically draws the eye to your insect character, informing the person it’s the most important thing.

A good example:


The blur starts WAY too early here…

Another addition to the blur thing could be that where ever the camera is pointed, even if its at the mountains, would be slightly less blurred. This would somewhat mimic how our eyes focus on various things. It also seems like it might be harder to implement, but again I don’t know.

Ok I have to go, and I’ve been typing this so long I don’t even remember what I wrote! Ill check back later when my brain is fresh. Hopefully it made sense.

^ I’m guessing Unreal’s Depth of Field settings should get the job done with the Near Region / Far Region parameters, if I understood this effect correctly. I tried something similar in this earlier video of mine; not tilt shift or anything (I didn’t even know such a thing existed till today!), just some DOFs which I used to draw the eye to the subject matter just like you suggested. I did DOF manually for that video though, a real-time system might need to figure out things like what the player is looking at (object of interest on the screen maybe) and deduce the DOF region parameters based on context.

Whether I want the player to feel like they’re looking at nature through a camera lens like a photographer or as if they’re actually there is also something to think about I guess.

BTW - thanks for the effort you spent on the example pic, it does look better with the spider brought into focus. And that photograph of the two bees around yellow flowers you linked - absolutely gorgeous! That should keep my inspiration tank at healthy levels for a while!

If you like it a little more then I definitely think you should mess around with the DOF. I might be repeating myself, but notice on that spider of yours that I blurred, that none of the nearby grass is blurred for quite a ways. I think that might turn out to be kind of key. If you just START the blur about four feet from the spider (I’m using the scale of the human you showed) and VERY intensified the blur with distance, then REALLY (A LOT) ramp up the level of blur when you get really far away like those mountains, then you wouldn’t have to worry (maybe?) about :

…Because if the mountains are blurry,the player will just accept that and go back to the game. And as for things that are closer and somewhat blurry, if they want to see it completely clearly, they can just fly closer. I feel it might even add to the exploration of a level, like instead of clearly seeing the whole area you can go to, and thinking of it as a tiny meadow or something, not being able to see what’s coming would force the player into the mindset of the insect so to speak.

https://www.youtube.com/watch?v=CxvTzkeMGY0 - from about 1:00 to 1:30 they talk about how the murkiness of the ocean instills a sense of exploration, similar to what I am thinking with the blur.
(btw, that game looks pretty cool. I wish you could just play as any of the sea creatures!)

Hearing that actually makes me feel really good!

(WARNING TO VIEWERS OF THIS THREAD: This post is mostly for the OP, so it may very possibly be boring for anyone but OP.)

Hey thanks for your post in my thread! it was thought provoking.

I figured you might be interested in some of the cool stuff I have come across when searching for inspiration for a game where you play an animal in a natural environment. For insight into why I have some of the links below: My main idea is a possibly procedurally generated simulation of creatures (not necessarily real animals like a deer) that have different abilities. They compete in the environment for survival, they have offspring which would spawn new animals in the world, and then at some point they die. When you have offspring, they have characteristics very similar to your own, but slightly varied.

So its a simulation of evolution. That player of the game would play as an animal. If they die, they are dead, game over, unless they have any living descendants in the game world. If you have multiple offspring, you have the choice of which one you want to take over. You can view its stats and decide.

So that’s the gist of my idea. My idea might be a lot more based around evolution and AI than yours, so many of these links are about those subjects too.

3D game lets you see through an animal’s eyeshttps://youtube.com/watch?v=rT3Iz5VA2eQ
29+ Evidences for Macroevolution: The Scientific Case for Common Descent website devoted to explaining evolution
https://www.youtube.com/watch?v=SeTssvexa9s&src_vid=mcAq9bmCeR0&feature=iv&annotation_id=annotation_467006

Evolution Simulators:
http://www.speciesgame.com/
http://www.freeworldgroup.com/games9/gameindex/bacteriasimulator2.htm
http://biologyinmotion.com/evol/
http://www.framsticks.com/
Redirect a whole bunch of good simulations
Get A-Life: Core Wars / Tierra
BoxCar2D

Some Games:
https://www.youtube.com/watch?v=zauu3qBm-iA Wolf Quest
https://www.youtube.com/watch?v=jlDDOdhMhks Fly Like A Bird 3
https://www.youtube.com/watch?v=7phu7m3zGAw Animal Planet Wild Lands
https://www.youtube.com/watch?v=Ftju_mJwWaU the paper fox
https://www.youtube.com/watch?v=szbFRTgNoEs#t=67 “GRAV”
https://www.youtube.com/watch?v=c1_mHsX0SjQ&feature=youtu.be “shelter” badger game really cool
https://www.youtube.com/watch?v=TCiJqfhgnlA shelter 2
http://store.steampowered.com/app/264710/ subnautica
https://www.youtube.com/watch?v=UksxcTQaBEE creatures in subnautica
https://www.youtube.com/watch?v=liXlZrtRuPI - TerraTech game - robot evolution…
https://www.youtube.com/watch?v=8Szr2NwA-cM Captain Forever - interesting
https://www.youtube.com/watch?v=qYDmdd0lv4o Captain Forever Remix

TWO SPECIES COMPETING AGAINST EACH OTHER coevolutionary genetic algorithm - Google Search coevolutionary genetic algorithm - Google Search

REDDIT THREADS (about AI mostly)
http://www.reddit.com/r/MachineLearning/comments/1tttu7/a_neural_network_learns_to_play_video_games_by/ http://www.reddit.com/r/MachineLearning/ http://www.reddit.com/r/videos/comments/2sxbas/some_students_gave_super_mario_an_ai_with/ http://www.reddit.com/r/gamedev/comments/2dxp3p/example_games_of_players_teaching_ai_agents/ http://www.reddit.com/r/leagueoflegends/comments/1tb1fm/create_vs_ai_hard_difficulty_via_machine_learning/ http://www.reddit.com/r/programming/comments/1y0s03/hacking_flappy_bird_with_machine_learning_hn_1/ http://www.reddit.com/r/singularity/comments/2x5h50/google_develops_ai_that_is_entirely_self_learning/ https://www.reddit.com/r/MachineLearning/comments/2xcyrl/i_am_jürgen_schmidhuber_ama http://www.reddit.com/r/MachineLearning/comments/2u0unv/askml_where_do_you_get_your_daily_machine/ http://www.reddit.com/r/gameai/comments/2iwvyh/how_plausible_are_anns_for_game_ai_theoretically/ http://www.reddit.com/r/gamedev/comments/12omad/would_getting_a_phd_in_the_artificial/ http://www.reddit.com/r/Games/comments/15ccfi/i_am_a_msc_ai_and_i_am_currently_thinking_about/ http://www.reddit.com/r/artificial/comments/2u73yv/whats_next_after_ga/ http://www.reddit.com/r/MachineLearning/comments/2x4yy1/google_deepmind_nature_paper_humanlevel_control http://www.reddit.com/r/elixir/comments/2ondcl/what_are_people_on_relixir_using_elixir_for/ http://www.reddit.com/r/artificial/comments/2xnc73/eli5_how_exactly_is_computable_aixi_modelling_the/ http://www.reddit.com/r/artificial/comments/y5mce/introduction_to_neuroevolution_an_alternative/ http://www.reddit.com/r/singularity/comments/1zwivq/psychologist_suffers_from_linear_thinking_and/ http://www.reddit.com/r/artificial/comments/27fg2v/a_simple_idea_artificial_intelligence/ http://www.reddit.com/r/videos/comments/1qm0ue/guy_explains_the_ai_in_skyrim_and_makes_it/ https://www.reddit.com/r/SSBM/comments/2yznt1/can_melee_be_solved_with_ai/ http://www.reddit.com/r/technology/comments/12nvbq/noam_chomsky_on_where_artificial_intelligence/

YOUTUBE VIDEOS (about AI mostly I think)
https://www.youtube.com/watch?v=1esNfaTRFuk
https://www.youtube.com/watch?v=HgWQ-gPIvt4
https://www.youtube.com/watch?v=M028vafB0l8
https://www.youtube.com/watch?v=0Str0Rdkxxo&list=PLEpwzRUUrDwhkSlQW2Yrlf4YTMYeN64t8
playlist https://www.youtube.com/watch?v=WXd6CQRTNek - really good (last 3/4ths)
https://www.youtube.com/channel/UCP-Ne1YeH6YK6Jvink87mNQ/videos geeuwd
https://www.youtube.com/watch?v=_m97_kL4ox0
https://www.youtube.com/user/PeterLNewton/videos super good
https://www.youtube.com/user/aicourses
https://www.youtube.com/channel/UC9pXDvrYYsHuDkauM2fLllQ/playlists

NEWS STORIES http://www.proton4.com/artificial-intelligence/google-ai-now-self-learning/
AI WEBSITES
http://ai-jam.com/
http://rednuht.org/genetic_cars_2/

More coming…

OTHER
http://stackoverflow.com/questions/970060/machine-learning-in-game-ai
http://stackoverflow.com/questions/7468131/games-containing-machine-learning-ai
https://www.google.com/search?rlz=1C1ASUC_enUS612US612&espv=2&biw=1506&bih=721&q=swarm+and+evolutionary+computation&revid=997161774&sa=X&ei=bxMmVZCZHo7hoAS9iICYCg&ved=0CFoQ1QIoAA#tbm=vid&q=swarm+AI+games+-starcraft
http://cotdp.com/wp-content/uploads/2011/06/RedcastDemo_Northshire_Abbey1.pngRedditAIQuotes
http://red3d.com/cwr/steer/
Pong-in-HTML5/pongScript.js at gh-pages · alanwright/Pong-in-HTML5 · GitHub Pong Script in JS
http://en.wikipedia.org/wiki/Machine_learning
http://www.reddit.com/r/artificial/comments/1axvps/looking_for_to_links_to_artificial_intelligence/?utm_campaign=AI,+reddit,+articial,+artificial+intelligence&utm_term=AI,+reddit,+articial,+artificial+intelligence&utm_content=AI,+reddit,+articial,+artificial+intelligence&utm_source=twitterfeed&utm_medium=twitter
http://www.reddit.com/r/gamedev/comments/2r25m0/genetic_algorithms_in_games/
http://lifecompetes.com/
HTML5 Genetic Algorithm 2D Car Thingy - Chrome recommended
https://www.reddit.com/r/cellular_automata/top/ Stephen Hawkings The Meaning of Life (John Conway's Game of Life segment) - YouTube
Computer program that learns to play classic NES games - YouTube

http://www.reddit.com/r/gamedev/comments/31idqb/im_building_evolution_based_ai_toolkit_to/ Evolution simulation: avoiding obstacles - YouTube http://en.wikipedia.org/wiki/Neuroevolution http://en.wikipedia.org/wiki/Evolutionary_computation
AI War: Fleet Command - Wikipedia

Google http://www.cs.rochester.edu/~brown/242/assts/termprojs/games.pdf Black and White’s AI implementation is a giant spreadsheet with varying probabilities assigned to different actions. If you punish the pet the probability of those actions decreases in the spreadsheet. It was very well done and balances somewhat the control needed to design a proper game and the ability to have emergent behavior.

MOAR games with playable animals…
Legend of the guardians - The video game part 9 (Wii) - YouTube Legend of the Guardians
Let's Play Tokyo Jungle [CO-OP] (Survival Mode) Part 1 - Lycaon and Wolf Take Over the World! - YouTube Tokyo Jungle
https://www.youtube.com/watch?v=zauu3qBm-iA Wolf Quest
Kinectimals - XBOX360 with KINECT Part 1 TRUE HD - QUALITY 1080 - YouTube Kinectimals
Hungry Shark Evolution Megalodon Jetpack Gameplay HD 1080p - YouTube Hungry Shark Evolution
School of Dragons - gameplay 1 - YouTube School of Dragons
Life Of Wolf 2014 | Android | Gameplay - YouTube Life of Wolf
https://www.youtube.com/user/MountainwheelGames/videos Stone Rage (playable animals)
Dog's Life - Gameplay PS2 (PS2 Games on PS3) - YouTube Dog’s Life STRAY Dog Simulator - By Gluten Free Games -Compatible with iPhone, iPad, and iPod touch. Android - YouTube
http://store.steampowered.com/app/264710/

Hey OldAccount29, I didn’t see your replies until now. Somehow the thread subscription thingee didn’t send me an email like it usually does!

I can’t promise I’ll read every single link in that list I’m afraid! This week has been especially hard on me with 16+ hour days trying to put together the first full-length gameplay video for DoN. It has so many features I’ve never shown before but it also revealed so many critical technical and performance issues that I’ve been racing non-stop around the clock to plug, fix and re-test, a very stressful time indeed.

I’ll try to check out the videos to see what the competition in this genre is like though :wink: If someone had given me these links 1.5 years ago, I could have taken the game in a different direction, but it’s probably too late now. Machine learning sounds interesting for sure, but I’m already sitting atop a frightening pile of unfinished features, bugs, and performance issues so I simply can’t afford to take on new features!

Of course - you have a fresh start ahead of you so make the most of all this research you’ve unearthed! In my experience it is best to define functional requirements clearly before the project begins and stick to it (while being nimble where necessary) instead of making giant functional leaps from time to time like I have done; I’ve thrown away much code and assets from time to time it makes my head spin when I think about it lol.

BTW I appreciate your unbridled enthusiasm, this is a very interesting genre for sure and there’s so many different approaches and sub-genres within it, some day when my game is complete I just want to sit back, relax and play all these lovely games of the natural world one by one :slight_smile:

Yeah, I DEFINITELY wasn’t meaning for you to read each link (there is many I haven’t read). Nor would I expect it to change anything anything in your game! I just have all those links already grouped together in a couple of pages, and I figured it wouldn’t take much to copy and paste them here. I guessed most likely you would watch some of the videos, and maybe save a couple of links to look at sometime. Honestly I just copied and pasted them over here! (partially b/c it sounded like you had a possible interest in making a game with mammals at some point)

Anyways, good luck with the debugging and optimizing! Hope everything turns out great!

^ Understood, thanks so much for your regular feedback and ideas!

drumroll… Finally, here’s DoN’s first gameplay video that covers all three pillars of the game - flora, fauna and ecology:

It’s 30 minutes long! I’ve added timestamps to the video in the comments so people can quickly jump to any section they find interesting.

My next focus will be improving the stability of the game and adding a basic hunger/heallth system. Then - preparation for the free alpha/demo: main menus and HUDS for procedural ecosystem generation followed by complete lifecycle and metamorphosis setup for the existing creatures in the game.

Of late all my time has been dedicated to releasing DoN’s voxel/3D navigation as a free plugin for everyone to use. It’s taking so much longer than I thought, but we’re getting there.

Some pics for dynamic collision detection in pathfinding and voxel approximations of collision geometry:

I’m not very happy with how this GIF game out lol, but there’s better stuff coming up including some sample content I’m preparing to go with the plugin which should do a better job of illustrating the different usecases covered by the system.

So how does your procedural fauna work in a technical way? Are you manually calculating and creating procedural meshes over time using something like Procedural Mesh | Unreal Engine Documentation or do you have individual stem, leaf and fern meshes that you procedurally place on a ‘base’ stem and then you rotate + scale this network of small meshes over time while simultaneously correct the position of the meshes such that they appear to ‘grow’ longer in length?

Very inspiring progress and great idea. Hope it all works out for you!

@VSZ
Awesome work. Can you change the flying camera and make “follow smooth” instead of being attached directly to the insect because it looks too rigid ?

Hey Galeon, thanks for the feedback! Are you talking about the camera for flying creatures (bees/butterflies) or the spider? I have a simple “follow smooth” setup but the interp speeds are very high at the moment.

Btw I’ve seen you contribute to so many WIP threads, thanks a lot for your involvement and work in encouraging so many devs in this forum, it is deeply appreciated.

Thanks for the wishes! Just remembered I haven’t checked out your new videos of Wormhole City; will take a look when I get some time.

@Capsup - thank you for asking this! Your second guess is right - each flora is a collection of many tiny meshes attached to each other at cardinal points to make a full plant. Some use morph targets, others animate their scale/rotation/position/material WPO/ over lifetime. A visual state machine on top of this allows transition to growth stages, seasonal variations, etc. Normally the performance of something this would be unacceptable but thanks to the Unreal’s HISM (hierarchical instanced static meshes), aggressive visibility culling and a system under the hood that constantly switches meshes between instanced and un-instanced versions on demand (ever morning when a flower blooms it needs to switch back to an un-instanced version so it can run morph targets and instances back when it is done), performance is good. Last year I managed to simulate around tens of thousands of such plants at 60 FPS. The game has grown much more complex since then and I frankly doubt I can reach that level again but I’ll be happy with even 1000 plants at 60 FPS at this point - even a tiny leaf has complex collision enabled so spiders can walk on it with feet properly grounded.

For adding new species to the game I use a visual front-end I’ve built to quickly join meshes together using a “Flora Masterplan” which controls both morphology and lifecycle events.

Like this:

I still need to create UMG widgets for making this player friendly - a daunting task because when the system was first written I never had player customization in mind. Obviously for players to use all this it needs to be very simple, very clean and easily understandable to anyone with a non-technical background.

I might investigate actual procedural meshes for players to customize basic shapes in the designer but that is a very very distant dream - Fauna also needs organic animations of legs and other body parts so it is far more challenging.

Yes, i mean a smooth camera mroe fexible follonwg the player and not tied directly to the flying insects. Will we be able to make the camera more distant (zoom in/ou) ? Some option to have a fur shader on some insects like the buzz would be great.
Your game is very unique, i will buy it, keep the great work.

Hey Galeon, here’s a hasty implementation of your suggestion using just the spring arm.

I think the butterfly looks so much better now, thank you for your suggestion! Spider walk looks more or less the same, although there are more serious issues with the spider walk that I need to resolve first (jerkiness, instability). I haven’t worked on my game for almost 3 weeks now due to plugin work, hopefully I can get back soon.

I’m in the final stages of releasing the 3D Navigation plugin now - preparing sample content and maybe a video with an overview or tutorial before I formally release it.

Anyway, here are the GIFs:

Edit: And thanks so much for the kind words!

Edit 2: converted the GIFs into links, they’re pretty large in size and I didn’t want to hog the bandwidth of everyone who takes the trouble of visiting this page :slight_smile:

Finally… my 3D pathfinding plugin has been released!

Grab the plugin, source and a fun sample project with flying bots from this thread below:

Here’s an overview of the system and a tutorial to get you going:

Super cool stuff.I’m glad you went with the simulation route.This is one of those dreamt of playing,and creating but lack the technical knowledge to do the coding stuff etc.Goodluck with the project.Keep at it sub’d.Sent you a pm

Updated OP with some pics. A quick status update:

Spider AI incoming!

Something I’ve been really anxious about pulling off is spider AI. DoN’s tiny AI need to navigate slender plant stalks, crawl through the underside of flower petals and expertly jump from one foliage to another enroute to a promising meal.

To make things hard, navigable surfaces like plants germinate and grow procedurally at run-time; is such AI even possible to implement?

To make things harder, Unreal’s native NavMesh won’t work here:

  1. It doesn’t work past 90 degrees, which is a no-go for spider AI (or indeed, any of DoN’s agile invertebrates!)
  2. Even the smallest cell-size in Nav mesh cannot handle plant stalks, leaves, flowers, etc. They’re simply too small and too tiny for the NavMesh to be anywhere near accurate for my game.

So yet another coding mini-project was underway :slight_smile:

The good news is that it’s almost ready! Video of spider AI hunting bees is coming up soon.

Change in gameplay direction

To improve my chances of releasing the first major version soon, I’ve decided to shake things up vis-a-vis gameplay a bit.

A pure Nature Sim will take time to implement, there’s no way around it. Lifecycle, metamorphosis, mating, etc need robust sub-systems, loads of art work, non-trivial gameplay and AI logic, etc.

So I’m thinking of a simple combat game mode (prey vs/ predator, predator v/s predator, etc) which players can immediately dive into while the simulation mode continues to grow over time. This will allow me to add new species to the game in a fraction of the time that it would take to add a new species to the full-blown simulation.

Multiplayer also becomes a possibility with the combat focused mode, something that I wouldn’t dare to even think about in the full-scale simulation.

To be clear, combat doesn’t mean military spiders training high-tech bazookas on kevlar-wearing flies or anything silly like that :slight_smile: It’s intended to be true to nature - i.e. visceral, limb-to-limb, silk-spinning and reflex heavy combat. There’s a lot of fancy options on the table when it comes to invertebrates (they’ve got so much more style than mammals!) so I’ll see which ones are easy to implement and which ones are best left for financially greener times :wink: