Feedback Dog Go RUN - a doggy adventure game

Hey guys,

I'm just taking my first steps with this project but I want to go ahead and start keeping a blog of sorts about how it goes.

Right now I am mostly planning so I won't give a full description of what the project will be yet, but I can say that it will be a console game with platform/adventure gameplay where you play a dog traveling across the great national parks of north america to get back home.

Right now I am working alone with no timeline. My first goal is to decide on a general art style and get enough of a demo built in unity to make a short trailer from and upload to kickstarter.

So far I have built a simple dog model. I have some automated rigging and premade animations, however I am going to take a few weeks to learn more about rigging and animation so that I can more effectively troubleshoot problems there. It seems a little weird to jump into something technical like this right at the beginning, but it's just a bit of personal development I need because it will help me make better decisions soon.

I'll be working through this excellent and free tutorial series:
https://www.manoanim.com/autodesk-maya-tutorials/quadruped-rigging-101-legs-pt1/

After that I will do some work on environment art. Then I'll get a simple character controller setup, take a run around in engine, and I think then I'll be able visualize where to go with the aesthetic of the game, rather than doing it all as thought experiment alone.

Work so far: first I made a quick concept sketch of the main character (Sipsey Dog), and then turned it into a 3d model. I am not sure how toony I want to go with the style yet. My initial hunch is towards the style of the older disney movies. That is, kind of realistic in form but with simplified shading.

The concept sketch demonstrates a couple ideas I have about shading: Flat shaded, outline with color and outline with black. Outline shader would be like you see in Okami.





I've marked this as feedback because I would really appreciate any feedback you got.

I'll also link any threads I make with questions related to this project. So far those are:

Agile Development in Plain English?
https://forum.unity.com/threads/agi...nals-gimme-a-rundown-in-plain-english.736538/

Ways to use unlit shaders and make big color changes easily?
https://forum.unity.com/threads/any...nd-painted-textures-only.737648/#post-4919780

 

A better doggo. I think I like this style. A touch realistic, but mostly it's left flat shaded so I don't gotta go too far with things. I feel ok about this so will go forward with it and see how it works with the enviro I have in mind.

This uses a single 1k color map and unlit shader.

Next step I rig this guy up and do a simple walk/run/idle animation.

 

A friend sent me a fun bison sculpt, so I started the day by retopo and making a game model from him. Can't wait to get this big bruiser animated, with big plumes of stink-breath steaming out his nostrils.

Alright, enough fun. Time to get back to rigging...

 

Got the pupper rigged! I used Advanced Skeleton. Took me two days to sort of figure it out, but given the power of the toolset it's well worth the time. I dunno why, but I get stupid happy making poses after getting the model rigged.

At some point I gotta do face rig and blendshapes and all, but for now this is enough to build an idle, walk, and run animation so I can get things started in-engine and "feel" my game.

The is Sipsey when he sees a squirrel in a tree.

Oh Sipsey, you shameless beggar!

 

Wow, they look great dude!

 

Thanks. All just quick stuff for prototype right now, will continue to refine the art style as project goes on. This is my first foray into hand painted and stylized modeling so likely the end result may be very different from what I am doing now. Very much a learning process.

 

My first animation! It's not so great but I'll get better the more I do. I dropped it in Mikael Gustaffsons environment pack as it looks similar to my intended style.

Woohoo!

 

Improved the rig and animation a little bit. Made the worlds worst character controller just to get something going on in engine.

 

At that speed he'll make it home in time for supper.
What does Sipsey mean? Is he named after a small town in Alabama?

 

Sipsey was a stray dog I found while camping in the Sipsey Wilderness (park in Alabama). Then we moved to Alaska where he spent most his life living in the woods, chasing squirrels and other adventures.

We live in the city now, sometimes he escapes and goes out for adventures. We live near a park and he'll just disappear for a few days at a time, then come back home very tired. He's a very wild doggo. So this game will answer the question I always wanna know: "just what happens when he goes out for adventures?"

In this fiction, he's going on a tour of the great parks across north america, so he'll be gone for more than a couple days I expect.

 

Alrighty guys, just gonna do a little bit of rubber ducky here. Past couple hours I just been tweaking and messing around with gaia, playing with colors and stuff... seems like a waste of time but it just sitting here, halfway zoning out and just looking at the game helps me visualize what it wants to be. Also helps me start figuring out, "where to begin?"

So our first environment will be a Yukon River crossing. I try to think of some kind of action sequence that would be a good chance to show something unique and test out some of my ideas. So I imagine our doggy hero needs to head towards the mountain, and the big Yukon is in his way. Player might climb up on this tall hill to get a view and make a plan, and then they could see a couple ways to get across the river. One way might look the easiest, but there's a bear milling around near it. Another way has some platforming, but if you fall in the river you will get swept away down pretty far before you can get back on shore. And the last way -- the smartest way -- is to identify a landing on the far side, go way up stream, and let the current carry you to it. So this way it's kind of realistic to give you that cool survivor man feeling.

So things we need to make this happen:

• finish out proto character controller, which requires trot animation, jump anim, turn anims, crash anim.
  
• basic AI logic for chase.
• simple anims for bear
• unique enviro content (at least I should retexture the stuff I am using and make a skybox to fit the rest of the aesthetic.)
• possibly a more appropriate stylized water shader

Eventually I want to move all shaders to unlit and work out some kind of global vertex color tweaker so I can create colors pallettes for time of day changes, but for now I just use standard lighting to experiment with.

A couple questions to figure out through playtest is how big should environments be? I want the world to feel massive and I want player to feel crazy fast. We also want player to have to plan out their travels, taking into account energy expenditure and availability of food and shelter. All of that will require significant space. The main idea is to make player really feel the massive scope of the land and get a sense of the effort it takes to survive in it. Shouldn't be extremely difficulty or anything, but you should get that "oh wow, I'm really out in the wilderness!" feeling.

Right now I use an 8km terrain, which seems like it may be adequate and doesn't require any fancy solutions to load and use. But the technicals of large open world is totally foreign to me so it's a big question to figure out.

For simplicity sake, I think iits best to use old school zelda type approach to environments. Basically you go to a funnel and it ask you if oyu want to move into next area, then that other area will load. Same as The Long Dark too. This way each area can be very unique and not have to worry over streaming the whole world endlessly.


Art design-wise I am thinking to stick with albedo textures only for everything, but add in a touch of realistic paint job kind of like I did with the dog, rather than totally flat colors like the ground and plants are right now. Maybe I will bake shadows, but I actually like the way things look right now. Obviously we need a lot more work on composition, background elements, and refinement of our color palette, but I think we can achieve a very nice look without shadows at all.

 

A lot of work past few days but little visual progress. Lots of trial and error to try and figure out smart workflow for terrain generation.

This is result of using gaia for height map and vegetation studio for splatmap (veg studio has much faster and more visual splatmap generator than gaia)


but this is the target look I want to achieve. Not 1:1 obviously but the general color palette and layout.

Right now there is a lot of technical limitation wiht the tools I am using, and also working with many different third party assets is pain the ass. I also want to use 2019 for the faster enter-play-mode and some nice QOL improvements with the interface. So that complicates things further.

So... what to do? Try things a simple way. I think I will make a mesh terrain myself by using Gaia to get things started with heightmap, export that as obj, manually breakup and optimize the obj in zbrush/maya, and then in substance painter I can use the full range of artist tools to paint the terrain precisely how I need it. Having a strong visual style is going to be really important so I need to make a workflow that supports that, rather than compromising.

This should also alleviate complications I had with using my own shaders for the terrain. All the plugins and unity terrain only want PBR stuff, so integrating my own shader was big headache. By working around splatmaps, I can just use the materials I want no fuss. I think... let's hope.

The performance issues can be figured out later. For now I need something that looks good.

 

Finally got the version upgrades taken care of, got most my workflow vindicated, so now I actually made some art today.

A few trees made with speedtree and some plants. I tried something weird with the fireweed and did a border around it. Not sure if I stick with that or not.

The plants and some textures is random stuff from google I don't got rights too. I replace eventually, but my main goal for now is to figure out how toony/stylized I want to go, and figure out how to use my vegetation assets in a way that i can sort of use them to paint the terrain for pleasing compositions when you looking at things from afar. This way you always gonna have nice art to look at. Gaia has a nice feature to bake plant colors into the terrain splatmap so that's a big tool that will help me make a nice color palette.

We've got birch, spruce, labrador tea and fireweed. All things you'll find in boreal forest. Nothing a dog would be interested in though. "wut dis? begtabuls? no tank yoo".

tomorrow we'll make some blueberries and raspberries our bravest hero can have a munch on.

 

Nothing real exciting to show but it's time to slow down and learn a few technicals.

So far I've experimented with gaia, megasplat, vegetation studio, and seen what they all have to offer in terms of splatmap creation and optimization. They all great, but they all geared towards PBR. This makes modifying them to work with unlit scenario a task beyond my understanding, so my reaction was to try and find a simpler way to make my terrain.

But for a large world, there is no simple way. At least, using standard modeling/texturing techniques I cannot find simple way to do so. Reason is simply because of vertex count. Big terrain with high amount of resolution all being displayed at once? It's too much to ask. Need to have automatic lodding happening on it, or at minimum split it into tiny chunks and LOD or cull the others. All of this requires some technical know-how.

So far I accomplish a minor goal in outputting a terrain as a mesh so I could easily tweak it in zbrush, then reimporting and converting back to a terrain to take advantage of units auto-lodding for performance. Also broke the large terrain into small chunks. Tool I used to do this was from InfinityCode called Mesh To Terrain. Very simple and useful.


Next task then is to learn enough about shaders that I can either accomplish a simple version of my task on my own or be able to modify megasplat. Megasplat does provide nodes to use with amplify shader editor, so that may be good approach. It also supplies a ramp toon lighting workflow which maybe i could use, but i do wnat to test out my initial idea of completely unlit first.

I will start by doing all the amplify shader video tutorials, then there is some fantastic shader instruction from a guy called Roystan I'll do as well. From there I think I should have enough foundation to start tackling my specific needs piece by piece. I expect within two weeks I should be able to showcase a 16km terrain full of life and everything completely using unlit shaders.

I also gonna do a little experiment with simple toon shading, like you've seen in Zelda BOTW. It's a nicie look, definitely more performance intensive, but I just wanna see how that might play out.

 

I feel like the blur camera effect is such a good idea. The walk could be just a tad slower, but the run being as speedy as it is makes it feel nice and swift. I often see people have their run animations too fast, I feel like it's better to give an extra kick to the speed, you really want the user to feel that rush that a dog would feel.

To be honest with you, I'd love a first person view for this as well. Add little idling dog sounds here and there like dog grunts and it'd actually be funny but in a good way.

What I really want to see is almost like a forward dash move and a nice leap, the rush and freedom of being a dog seems exciting.

Are you going for mobile platforms? For the optimization techniques I'd highly recommend occlusion culling and using per layer camera culling at the Unity asset store (free), that one has helped me so much. You probably already know this, but also make sure everything is GPU Instanced, and keep in mind that combining game objects will add many MB to your scenes and could slow down your game, which can be combated by using an asset that streams game objects in.

I would definitely look for posting on twitter consistenly with the #dog and #funnydogs hashtags for promotion, you could really draw in the dog people audience with this, you've got a niche.

 

Hey @astracat111 , great tips. I about to leave back to alaska with the big pooch for a new job and will be busy for a bit, but once I get back and kick this back into gear I think getting a proper devlog going on twitter will be good thing to do. Because even if you aint into games but you a big doggo lubber you might find it interesting.

I was hoping to have more to showcase by the end of next week but it's taken me a good amount of time to figure out workflow and tools for the world building. Now I just got properly started and I gotta leave!

First person view is something I plan. Or, at minimum, a very close right over the head option.

Hilarious dog noises and "goob boy" dog language is like half the reason I wanna make this game. Happy boy panting, random belches, big yawns, "aw hecks! you dun me a frighten", etc.

Not going for mobile simply because I don't play mobile games myself, and also I want the game to be pretty robust in terms of world size. Graphics are simple enough but I trying to have pretty large worlds with high speed movement, so unless I get big funding and can hire real experts I don't think I'd have the skill to put a game like that into mobile hardware. But I'll look into those techniques you mentioned anyway because if possible I think this would make a good switch game, which I've been told is basically a powerful mobile device.

You've nailed it with the idea's about movement. Capturing the wild adrenaline rush and reckless abandon a dog feels when tearing through the woods is exactly what I want. Right now all the animation is just placeholder first pass stuff. So far I only made a crude walk and gallop. But there will be walk, trot, gallop, and sprint, along with transitions, jump, and of course sit and otehr contextual poses.

A large part of my workload is simply going to be working on the character controller to death until it feels so good that it's fun just to run around in the world with no goals at all. Apart from theme I think that will be the most important aspect of the game. It must be absolutely fun just to run.

The motion blur is something else i wanna work on a lot. I need a very specific look. I want to capture the kind of tunnel vision i imagine dog has when he's chasing squirrel. THe entire world around him should be intensely blurred like you going warp speed. The only thing you focused on is the thing you're chasing. Just 1000% crazy doggo focus. Only thing that matters is getting the squirrel!
I think this along with very fast movement, responsive turns, and plenty of natural obstacles to dodge and leap over will really make for a powerful, immersive dog experience.

 

@astracat111 , not a ton to see here but I'm using some of those Flatkit shaders on the terrain, the dog, the skybox, and also the fog is a post effect.

I'm really happy with these shaders because they give you a ton of options to experiment with yet the are not complicated or obtuse to understand. Basically everything I imagined in my perfect shader for this game they got and also have really straightforward documentation and plenty of demos.




____________________________________________________________________________
Before I head out in a couple days I try to get some basic vegetation rules setup to populate this terrain. I won't go too far with that, just get the place filled in with the basics.

Then I may be gone a few weeks or a month.

After that I get my "obstacle" objects set up and then get jump animation in there along with better controls for lateral movement. Then I can start to imagine what the infinite runner gameplay might feel like.

From there I may refine the art further, or I may use playmaker to get some simple AI in there and make a quick goal-oriented gameplay scenario. I think at that point, with a bit of polish whereever I need it that should be enough to make a nice demo trailer from to throw up on kickstarter.

 

Sounds like you know what you're trying to capture and it sounds like the kind of thing I'd be into. Okami was great, but it wasn't about roaming free as a dog rather than solving puzzles and platforming. When you get a bit farther along and want some extra feedback on the feel of it I'll give in what I think, e-mail me at astralojia [ @] gmail.com

 

Will do!

Yeah I loved Okami but was always a little disappointed because it didn't quite give me what I was looking for. I don't particularly like puzzle games and I found the drawing to be less artistic and more frustrating.

So yeah, my goal here is to really capture the experience of being a dog and immersing you into that. If I do it correctly, player will naturally be making the most doggy decisions and feeling smart about that.

 

A new style of tree. No wind in the shaders yet but that will make a big difference.

Right now all my vegetation is just perlin noised across the terrain. I have a handful more species to make today then I will spend time making specific biome rule sets so that we 1. look more realistic, and 2. have distinct areas of gameplay.

 

Had too many errands today, but with a few hours I improved the shader a bit for the trees, got a lot of stuff organized, and got my workflow all finalized and written down.

Since I managing a lot more than just art, I find it necessary to write very clear checklist on how to do each thing, that way when I swap hats all I gotta do is skim read and then can jump right into work without lengthy warm up. Taking the time to write seems like it adds lots of time, but in the long run it saves a lot and prevents fatigue.

I've added a bit of outline on the normals, which gives nice interest to the fluffy trees. There is a small issue of see the outlines through certain objects like grass -- but I think this is a simple matter I can fix later. For now I am liking the way things look.

I wonder if the color palette is a bit in your face right now, but that's easy to tweak. If you following along do feel free to drop any suggestions you feel. Keep in mind target audience is pretty broad, but would focus on high school age kids primarily.

 

Hello,

First, all the vieeos I had up previously I am going to migrate over to an official channel for the game. just haven't done that yet. Plan is to get twitter, facebook, all that stupid stuff set up and linked together in an interconnected web of evil!!! err, i mean marketing. Whatever.

Now, the good stuff! I've settled on art theme pretty much. Got it all figured out how I can make everything and how it will all work together. I've meddled styles from several awesome artist. It took a lot of tinkering to find something that worked, and I even began to wonder if I was just wasting time at one point, but then like magic things started looking not bad. Still needs a ton of work (mostly in logic of how things are spawned, and of course a custom touch because no procedural gen system can make it quite as nice as going by hand).

So the world building workflow is basically:

• set up a biome, define it's rules
• tweaks its assets until it has a distinct feeling
• then go through manually and tweak things for most appealing look and gameplay.

So far you are only seeing the first two, but I think I trust it enough to give the stamp of approval.

I think it may be good time to switch gears and flesh out the controller a bit more so that we can have a nice trot, turns, and jump. This way we can start interacting more with the obstacles in the environment and start to treat it more like a gameplay world rather than just dog-walk sim. Basically do grayboxing for gameplay now that I have lookdev mostly settled.

So then, next up is a lot more animations and then plugging that all into the controller. Then we'll need plenty of fallen trees to leap and duck under... and then, I want a bear to flee from.


ps. little easter egg, this heightmap comes from a big river up north our real life doggy hero has actually crossed. This time around though he'll have to cross without a hooman to help him.

 

Trying out some different skybox styles. One is just an ordinary skybox with photoshop filters, which I like a lot and is easy to make, but very static. Ideally I want a look with this amount of complexity but I do want the clouds to move around. I wonder if there is a way I could animate just certain parts of an image, or maybe overlay two images and move them separately. That way I can keep the nice painterly look but also have it be dynamic.

So far everything is using custom lighting only. There are no shadows so performance is still very acceptable despite lots of foliage. There is still more to add though but also lots that can be optimized, so for the time being as long as i'm not dipping below 60fps i just don't worry about performance. Will bring in technical specialist to help with that when time comes.


A simpler solution is this particle cloud effect from teh Illustrated Nature pack. It is beautiful but a bit too abstract. I feel like the sky is a major part of the canvas and mood of the scene. I want a lot of color and depth and character in there. One thing I always remember when I visit a place is the sky. I think that's very interesting. No matter where you go, it's the same sky, right? And yet every corner of the world has it's own unique sky that is maybe the biggest part of it's nature.

I think making good skyboxes may be a major thing to experiment with.

 

Couple new species to flesh out the Draw and Riverine biomes. Got some River Willows in bush and tree form. Also started working on deadfall (not seen here) that will be some of the core obstacles to get over while chasing or being chased. My idea is for the world to look completely natural and yet you are jumping and dodging like an infinite runner game to get around. This may make AI a challenge but we just see about how that will go pretty soon here. It may be simple matter of relegating AI predators and prey to less dense areas, and letting the thicker areas act as safe havens. In a survival game there is also travel time and energy usage to consider, so something like a big thick draw could pose a logistical consideration if you are trying to work out some kind of daily hunting/gathering routine. Anyway, these are some of the gameplay questions I working out as I am mostly focusing just on looks at this stage.

~

The draw stands out more nicely with the bright green Willows in it. Once I get a tiny little stream trickling in there with some sound effects and some nice bird family it's going to be really nice.


The big white flowers are Cow Parsnip, which will make you itchy if you bump into it. Don't want the itchies!

Along the banks of the Yukons you got thick stands of River Willow and grasses. It's a place so dense you could easily bump into the backside of a moose by accident.

Dogs don't actually savor sunsets. They just dogs. Here Sipsey is contemplating dinner.

Next up I get some more deadfall in there and maybe some cliff meshes to line the banks of the river.

 

Some cliffs. Quick solution here is just modify some megascan cliffs. I'll need more varieties but this is good for now.

Unity's new terrain sculpt tools are pretty good. The bridge tool is great for leveling off the terrain along the edge of the cliff meshes.

 

Got my billboard shader fixed with some help from a kickass techartist. Also kudos to Lennard who is the maker of VegStudio for prompt support. Great product and really great support.

So, like I said earlier I am pretty cool with this general look. I like the way it feels and it's a workflow I can manage and shouldn't be a bear performance wise to handle. I feel it strikes a good balance between playful cartoonism and realism that shouuld be appropriate for the gameplay. Of course, if you got anything to say about it don't hold back. The sooner I get feedback the better.

So from here my next steps are just continuing to make it look better but also working in gameplay and getting a feel for that. Much of the gameplay will depend on the character controller which I'll only be able to take so far myself, but I will need to have all my animations running smoothly at least for a good first presentation on kickstarter.

The gameplay I will focus on first is infinite runner style movement. The first thing we need is more obstacles to leap over which will mostly consist of deadfall trees. Jump animation to go with that. Then a bit more challenging will be adding in turns and strafes, which will need a little nuance to make them look good and help give a sense of speed and momentum.

The final goal for this demo will be to have an objective based scenario with a couple characters, dialogue, and some contextual stuff. Sipsey dog will need to find enough food to build up his fat stores so that if he can manage to get across the Yukon safely, he won't be so chilled that he can't get warm again. Basically a mini tutorial that will give a sense of the gameplay but also the overall game structure. Move to an area, master survival in that area, face unique final challenge, move to new area. Each area has unique characters who teach you some new skill.

Oh, and here's the big bubba himself

 

An idyllic creek, and a particularly disgusting muskeg. Sipsey loves disgusting muskegs.

 

Definitely making progress with the art style!

 

Thanks @angrypenguin .

still miles to go but got to start by throwing something on the canvas to work with.

 

Have you tried combining the two?

 

@kdgalla

I have thought about it but I am away from home right now and haven't been able to get into any of the art for past month.

I have some devlogs that talk about custom skybox solutions bookmarked I got to go through. I think it is going to have to be a custom solution I devise to fit my own particular workflow style but yeah, likely I use some mix of elements to get a painterly but dynamic skybox. Above all I want to maintain simplicity, so probably I will be looking at hand painted rotating spheres with layers of 3d or particle fx on top. If I can build layers with simple techniques I already know, that is ideal.