Procedurally Animated Monsters in my Monster-Hunting RPG Roguelite | Vicious Legacy Devlog #2 

Believe me, I understand the irony of posting a Unity video right now, but the show must go on 🤷‍♂ For everyone asking my thoughts, see 11:45 in video

The slice through the blocks in the video instantly made me think of MGRR's free form cutting. It would be soo cool to get some clean slices through the random monsters and have the polygons get sliced in a similar way

Funny time to release a unity devlog 😅

yes! I love piles of vague constraints and maths! I did IK once, in scratch. it took me 5 hours, but now I have a stick figure with legs. totally worth it. also, look at rain world. first of all, it's a great game with what I think is easily the best procedural animation ever made. Also, when you said the "world will go on whether you're there or not", I remembered Rain World, because I think the developer of that game said something almost identical to that. also I'm the 69th view

It might just be me but i feel like the water would look better by keeping with artsyle of everything else rather then trying to be realistic. Maybe something like Windwakers water would look cool. Just if you want though.

This looks great! It's heading in such a good direction, honestly, I can't wait to see where this goes

Some advice, it would be cool to see monsters having dismemberment and stuff, so like if you chop of a leg it ragdolls and wildly kicks its legs, also adding weak spots would be great (e.g. a head, lung or heart). Also it would be so awesome if you gave some creatures heard movement using a 'BOID' algorithm and also behavior patterns like eating stuff, and/or drinking or bolting away from each other in a playful way. Also I think you should try adding stealth, so for example if a creature begins to drink its cone of perception would be greatly decreased, eating would slightly decrease it, and while in a herd it could increase. you could make it so the perception and degree that an enemy may predict your movements and search for you is dictated by an intelligence value that would vary in-between different species or even individual enemies. Add these stats and unique elements between enemies like diet, hunting patterns, aggression, sociability and intelligence would make the game so much more unique and immersive, while encouraging the play to connect and observe the environment carefully.

Looks great! I'm just starting out in game development (luckily I chose Unreal haha) and videos like these are really inspiring to me.

I love that the monsters feet don't rotate. Very funny to see

12:13 wow this was my favourite moment of video, looking really good man

This project looks and sounds super cool so far!! It's definitely the sort of thing that I would fantasize about doing in my head, but never actually think of bringing to life. It seems incredibly, brain-blastingly complicated so kudos to you for even finding it within yourself to give it the college try. Will definitely be keeping up with these logs going forward. Excited to see where this heads

Dang this is so sick! I'm excited to see where this goes. I can see this really being immersive.

That practice segment of the spider creature chasing after the player and following looking at it was very eye-catching, good job! Moved a lot like a real spider, and looked very natural. :)

Most impressive, and always learning something new myself.

I really like the concept. Very original. Best of luck moving forward, and I can't wait to see more!

Just find your channel and your stuff is inspiring, thank you for putting yourself out there and sharing! Proud of you.

I would definitely love to hear more about inverse kinematics and/or fsm

It would be incredible for monsters to have specific weaknesses, abilities, and behavior patterns. Like in the game you could happen upon monsters that won't move if you're looking at them or ones that require different weapons to fight effectively, ones that can burrow in the ground or climb and hide in trees. Making it mod-able would add a huge community aspect to the game and increase its lifespan as well.

Nice video mate! I like how it's just a bit of a ramble for 12 minutes. Good luck with your game

Great work, dude! I think you ought to put the forward step goal of the feet considerable further forward; as it is, the monster looks like it's falling forward and catching itself, not *ACTUALLY* stepping forward and pulling itself. The reactive nature of IK is an unfortunate pitfall that can reduce the "proactive" look of movement, but you can kinda cheese it by having it *REALLY* thrust its foot forward to compensate. :)

I love seeing procedural animation it always looks so cool when done right!

I absolutely adore this implementation and your break down.

This is giving me vibes of how the enemies work in rain world and i love that cant wait to see

This looks incredbile! Good job :)

Oh hey! Thank you for the shoutout! Subbed and looking forward to seeing more of your game! Cheers! 🍻 -Nicky

I love this take on the genre (if I can even say that being it is so unique). I think the procedural generation already brongs the world to life in a way that many developers dream of. Cant wait to see what the future holds for your project!

Glad to see you're back. I was worried by the prolonged absence.

Keep it up!! It's looking awesome, can't wait to see how you progress further :)

its looking amazing and this journey is going to be epic, definitely be interested in watching some explanation/break down kind of things or topics or theories of game design or something like that :)

It is wonderful to follow your progress! Thank you for another update :)

Great to see you back!

i absolutly love the game's concept, and also the videos aren't too long, 10-12 mins is perfect

Quick tip, I saw your animator states, I've already been there bu I've now a scalable solution wich could interest you Basically making a state machine of possible action of your player is the best things in the world Make a class called "CharacterAction" wich is always linked to an animation This class have base functions like "StartAction" (who also play the right animation), have interruption parameters etc Then in your Character class, make a field "currentAction" and Start / Stop functions etc You will see that adding new functionnalities in a specific script representing the action, then adding actions to character is a life saver because every logic responsible for doing one thing is in one location ;) And then I suggest you to not use animator transition but instead script based transition, it will keep your animator states clean and scalable as well as having more flexibility to design transitions trough code ;)

Ambitious! Wonderful!

the animation looks great!!

Can’t wait to see how the monsters will look in the future. They already look cool even though they’re just white cubes. Great vid as always, already exited for the next one.

The idea of you being subject to the world around you and monsters existing beyond the player reminds me of rainworld

a game like this with the ecosystem at the level of rainworld would be insanity

amazing devlog, keep it up

0:42 ty for the lotr reference

man these videos are soo good!

Hell yeah. Hope leg length will factor into step height for what they can traverse

I've noticed that many people say the water looks weird. But I think it's good, altough not side by side with the style of the terrain, either he reworks/enhances his models and textures for the latter, or he changes the water shader. Both are time consuming but in my opinion, considering the direction he wants to take, the first option is better, i actually like the water and could fit better with the core idea if applied to everything. But anyway great work with the animations, it gives a weird nice vibe 🎉

Wow this game sounds so cool!

Very cool! Those movements reminds me of Miyazaki's Spirited Away monsters. I think it's because the target location for the legs moves "late" compared to the body's movement; or at least it depends on it. The result is a "dragged" walk. Which however is very impactful for some types of monster!

Love it!

good job!

Game looks pretty cool.

Looks good

Awesome

Damn I am still struggling with making a ant like creature move and you are here making new species, impressive! I hope to learn a lot from your video's!

Great Video. I recommend looking into Were Beast Where Born as it has many similar concepts as your game and might help you in creating the enemy's.

Oh boy, just the words "Finite State Machine" triggered some university-related PTSD I didn't know I had.

Very cool 👍

Oh my, what a timing for a dev log 😂. On the other hand, why not? If not now, there could be a chance that later will be not relevant any more. I'm on a little bit earlier stage making personal game then you are, but still I believe I know how you feel...

nice job!👏😄

The legs seem to lag behind the monster a bit. Perhaps shifting the target positions forward some would fix this.

Ah, refactoring. Right behind bugtesting for favorite part of game deving. It's always nice looking at your old code and half-cringing at the load ahead, but being proud of how far you've come in the meanwhile. I'm really happy to hear the Vicious Legacy is continuing though, in light of... absolutely mind-boggling decisions that I can only imagine were discussed as "$$$ ?" "$$$ !!!". Especially since the aspirations you mentioned in this video are something I, personally, would *love* to see come to life. I'm weak for games that live beyond the player, what can I say? And your use of 'ecosystem' in reference to that has me all kinds of jazzed up. I can't wait to watch this evolve and grow.

I love your videos, man. There's not much else to say at this stage other than that. The game is early in its making. I suppose it's like a first draft. I'm a writer. I haven't done much with developing games past concepts or dreaming mechanics. Messed around with tabletop games since they require construction paper and glue instead of coding. Hit me up, if you want to talk worldbuilding, lore, or mechanics.

makes me think so much about rain world

My solution to IK's chaotic maths problem is to define extreme possibilities with various animation clips of body parts, then read the bones' position over the timeline of those clips and then apply them to the actual character's bones in the LateUpdate() phase (I'm pointing out "phase" as I'm not doing it in the actual LateUpdate function, but in a Coroutine that act variously through the various updates phases. Such as the AI-beharior being managed during the normal Update phase, but the bones being managed during the LateUpdate phase). The hard part of this method is that there's no inherent way in Unity to get the bones transforms' values via an animation clip or, at least, on a build/standalone. (As soon as the build is done, all animation clip are converted into something else.) So what I did is create a mini-game with an UI button that allowed me to record and save the bones positions and rotations per frame for various body parts as the mini-game would automatically plays various animations with various creatures. Things like landing soft or landing hard, short or extended paw strikes, close or large bites, etc. As I save those data onto a data file, this mean that I can add more animations by updating only that specific file in my game so, for example, I could add a new set of animation for a new breed of wings for some flying creature. Then, I simply load those animation values into the game's RAM and read & apply whatever animation I want to use on a per-body-part basic. For example, if a monster has back legs like a horse and from paws like a lion, I can use the animation of the walking horse back legs for the back legs bones and use the lion's from paws animations for the from legs. I can mix thing up as I need based on the "nature" of the monster's body parts.

The game you want to achieve hype me a lot as Monster Hunter fan. But with all the procedural animation and dynamic stuff I also hope that your game will not be too CPU heavy.

I really like the headtracking effect

This is getting insane. Here's hoping you don't get burnt out or anything causes this game to not get finished. Really looking forward to it!! Not sure if you had covered it or not (bad memory). Do you need to make these skeletal frames for the procedural generation to then build from? Or does the software handle all of the work? It's unbelievable to me as someone who doesn't know anything about coding and such. My mind is thinking of it like Spore almost, you make a bunch of frames/skeletons then the software picks and chooses legs, torso, arms etc. until it has a full monster with special rules depending on the limbs it choose. Be awesome if there is a "ecosystem"/evolution line that you can see as you travel. Monsters with increasing similar traits you find in a circle, you progress inwards until you find the apex predator that started the line without their offspring's situational adaptions. Maybe as mini bosses with islands of great important if you go that route like Raft. If you go that route. Absolutely confident this is gonna be an amazing game no matter what direction you take it!!! Keep it up!!

In one fkin line you just made me understand what IK is!!!!!! Or im just dumb lol

looks very slay bestie

That's definitely a good idea to make refactoring when it starts bothering than making video in half a hear "creating my game from the scratch :')"

Some short videos of concepts, even almost tutorial-like, would be great.

If you're using navmeshes to control the agents, how do you intend to bake them at runtime when the islands get generated?

Is there a way to make models with free software and send them to you (I'm thinking of something like Spore Editor). I know it's still a bit early for creatures, weapons, armor, etc, but I'm just asking in advance. I normally only provide drawings, but that has always interested me

The problem with animating movement with inverse kinematics is that it always ends up looking like the body is flying forward and dragging the legs behind it. Probably happens because in normal animation charcters that are running put their legs forward ahead of them, where they plan to go. But in procedural animation done through inverse kinematics the legs don't "know" where the creature is actually going and can only react to the movement of the body, which leads to them dragging behind it. But I saw a pretty solid gallop done through procedural animation, so it's not impossible to fix.

nice

What if you added a variety of personalities to the monsters both to make the islands more alive and also to provide diversity or other types of difficulties? (Examples: Territorial monsters that make gestures to intimidate the player and only attack when the player ignores to their warnings or attack them. Monsters that outright ignore the player until the player attacks them or monsters that, either in response to damage or the player's presence, simply run away).

will there be some potions or equipment system? i think it could be really intresting

heart for the algorithm: ♥

Why do all IK walkers look like they’re weightless? It’s like the legs are being dragged by the body, not the body propelled by the legs.

I am loving this series! Good luck with unity

Finite state machine is probably the better solution, but a switch/case could also flatten out your enum checks.

Can you tell me where the clip at 2:48 came from? That is extremely impressive.

you can make a system in which the character will get some raw material in ocean just like raft game but in some unqiue way

Will you add flying monsters? I feel that it would add so much variety (but it would also be very difficult, probably)

This is awesome stuff, procedural animation is so cool! Quick question, how did you get your channel off the ground? Did you blast out announcements into discord servers?

Where did you find the clip at 2:50?

How did you control the character? I need tutorial for this view :D

5:51 @cog ❤

feels weird watching unity content with all thats going on.

rember it is not about making a game, it is about makeing a peice of art. if you can't make a long video thats fine, if you have to leave the project for months due to burnout that is fine, just remeber to make the game into somthing you are proud of. no matter how short the videos become or how many weeks you take off for any reason, it will in the end make your game better than if you give yourself impossable deadlines to complete. pls take care, and I wish you luck on this project.

You need to change the colour palette, the models are not bad, but the overall scene doesn't look great because the colours do not go together. I'd recommend you find a reference and try to copy the colours. I myself suck at picking colours, so that's what I tend to do.

"Like I said at the start, I have not made as much progreses as I would have liked" Do not undervalue the experience you picked up on the way. Also nice game idea, I wish you the best of luck to get through with it. I also wish you all the strength to cope with the gamer-community, it can be such a terrible crowd. And don't listen to the "I have the best idea since sliced bread, you need to put XY into your game"-backseat gamedevs ... at least not for now ... they usually have no clue whatsoever ;)

Just a quick heads up... Two segment ik is stupidly simple to roll yourself, once you break it down into a few small mathematical problems. If you know how to use vector functions and trig functions in unity, you're there.

I suggest switching engines. From the videos so far it doesn't seem like it would take much to port your work.

If you don't think you have the time to change engines now, you REALLY won't have the time to change engines later.

The new unity fees are dumb and should be changed/reverted, but I increasingly get the feeling that bunch of people who have never actually made a credible attempted at making a game are riding it as the new "current thing". Almost nobody who's in the middle of a project, and who has a realistic outlook on the financial side of indie gamedev, is going to erase months / years of work.

hes so cool i wan pet him..

The problem with navmesh is that you can't use it for procedural generated maps...

Is there some clever trick to make IK look better, or is it just hard work? In this example, and lots of others I've seen, the IK is pretty nakedly just dragging the legs around under a smoothly floating torso. I've seen some very impressive IK implementation in things like Shadow of the Colossus, but even if each of those animations isn't exactly bespoke, the parameters were all carefully tuned for the specific creatures in mind. Without some special technique, how are you going to generalize those principles across creatures with variable body plans? Additionally, if every randomly generated monster chooses its behaviors from a small enough pool of options, players will feel it. Last I played No Man's Sky, the randomly generated animals weren't meaningfully different, so their random generation didn't matter at all. Do you have a system in mind to link form and function? I worry that keeping them separate will make the random generation meaningless, but so would randomly handing out features that tie directly to specific abilities. Whatever identifiable ability-enabling part a monster gets would just become a category the monster falls into, and the rest would be meaningless. Spore let you make crazy weird looking animals, but their verbs were still incredibly limited, so the difference was mainly aesthetic and didn't particularly matter beyond humor and a small list of abilities that required a specific part. Basically, if the shape the creature takes doesn't meaningfully affect its behavior, random body generation will just act to obscure which type of thing the player has to deal with, instead of hinting at its capabilities, but granting pre-made abilities based on specific parts would override most of whatever unique identity a monster could have had. I also wonder about the implementation of ecosystems. IRL, they mostly appear pretty stable after aeons of evolution working it out, but if you just randomly generate several species on an island, I'd expect it to crash pretty quickly. Either a predator eats everything else and then starves, or it can't eat enough and starves, leaving the non-competing herbivores to wander freely, not interacting with much of anything. Or are you intending to implement plants that herbivores have to find to eat? The problem with simulations like these are that you can always make it more complex, but the effort required is exponential, and the value it adds to the game gets more obscure. In any case, it would be a problem if the ecosystem starts to collapse as soon as the player arrives because it's not created in a stable state, but finding that stable state could require a huge amount of computation. Either that or you fake an ecosystem well enough that it looks like its there without having to keep track of the details, but then the player can't really meaningfully change it, and any attempt to interact with a system that isn't there threatens to reveal the the facade.

Correction: Refactoring is taking code that is shit and trying to make it less shit. (When you come back and look at it later it will always be shit.) 😄

do you think characters should poop and the poop makes more things grow and that's the circle of life?

This is coming along swimmingly....Pity about the Unity thing.....Hopefully, you can switch to Godot or Unreal without losing any of the steam you built here....Idea is sick

The “Raw and Wriggly” devlog. Uhhhhh