3d action game development

Alex_J

I'm still not finished with my current game project, but I took a break from it for nearly two months to prototype some ideas for next project. The main problem I want to get around is that there is too much work for a solo developer. How can I make smaller games?

But it still has to be something I want to play and am excited about. So I tried a few ideas. In the end, this game idea both excites me the most and seems the most doable workload wise. It is not necessarily less work than others, but it is heavier on art rather than programming, which suits my interest more.

In particular I really enjoy animation, even if I'm not that great at it. This project will have a lot of animations that I'll have to make.

Briefly, it is a third person roguelite action game in which you use a solely a bow and arrow. You are an Amazonian warrior on a quest to defeat evil monsters. In theme I want it to feel like those old Ray Harryhausen adventure flicks. I've enjoyed the hell out of Souls games but they get old with the overly grim and often grotesque themes. I wanted that sense of adventure without having to be so grim and dark.

Anyway, what I have done so far was mostly R&D to figure out general development pipelines, but there is some actual visual progress to show:

The main thing I wanted to get handle on first was animation pipeline. Because I haven't really done much animation at all in unreal. So I first used the paragon characters to setup a basic controller where you can run around and shoot some targets with the bow.

I don't want to be locked into trying to frankenstein together other peoples animations, so I started getting some practice making my own animations. So far I've got down the basics for locomotion plus a shooting montage. There will be a lot more to do for complete animation blueprint, however this is enough to satisfy basic gameplay and I feel that I'll be able to create all the animations to a satisfactory quality for the many monsters.

I've used Advanced Skeleton to create a rig on the Daz skeleton. THe character is from Daz and I made the toga in Marvelous designer. I just slapdashed things together for now. I will continue to iterate on the character design here and there as I go. There are a number of accessories also blocked in but I'll work more on the character later.

The hair is stolen from paragon sparrow. I'll probably just keep it though, making hair is too tedious and this one fits perfectly.

Code-wise I have the basics of the character controller ready to go. You can run around, shoot arrows, interact with the stuff, there is some simple flight physics enough to make the shooting fun... so next we need some enemies to fight... But that's a big ordeal and I wanted to do something a little easier. So I started with blocking out the first level, which will also serve as the vertical slice level.

I had a hard time getting started because I've never done a blockout for a game like this before. I always try to use the best tools for the job so I don't waste precious energy, and I had a lot of indecision over whether I ought to do the blockout in maya or unreal. In the end, I decided to stick with doing it in maya so that I have zero restrictions in control, ergonomics, and speed. But it does mean that I have to export to unreal and double check often to ensure scale is appropriate and so forth.

I tried to only focus on the playable game path and not get bogged down with background stuff. It's hard to come up with something from nothing, so I just took a map of archaic Athens and used that as a loose guide. Then I define a checklist for all the gameplay beats that I want and just start looking at the map and see how they might stitch together.

I just use some locators to indicate where each beat could happen. This is about as far as I can usefully go with the blockout, so next step is that I'll have to get the enemy AI going.

My own game means I get to make everything how I want, :). One thing thats always bugged me about these types of games is that the environment is always scaled up a bit. I expect its for gameplay reasons, but it just always bugged me that like hallways are large enough for a giant. I'm going to try and keep it such that we keep that old world, quaint charm. The setting is early bronze age, so I think the architecture might be more like you find in afghanistan, rather than huge marble temples you might think of when ancient greece is mentioned.

Offer a sacrifice at the shrine of asclepius for healing

I've tried to do the blockout with gameplay and pacing in mind, and set it up so that extending or shrinking the playable space will be easy later on. This was one of the big reasons I wanted the blockout done in maya rather than unreal. It is much easier to make fast selections and wholesale changes in maya (at least for me).

An example here is "the harpy gauntlet". Here I intend to introduce the player to an airborne, ranged enemy that will harass you throughout a mazelike corridor. My initial idea is that a single harpy should be a big problem, however it could be that so much trouble from a single enemy without scoring a kill could be too much stress and not enough fun. So maybe it will end up being a handful of enemies spread over a larger space, so that you can get some kills to keep up your spirits. Won't know until gameplay, so that's why I prefer a workflow that is fast and easy to make changes.

I am also the environment artist though. It's not practical for me to do a complete job of all the different discplines. So if I can get a little art prototyping done at the same time, I try to do that. This is another benefit to doing blockout from my DCC. I can quickly whip up blockout+ models in order to start developing a sense for general architecture designs:

For instance, setting up some slightly more sophisticated collumns and inlaying some materials to represent designs on this religious building was a lot easier for me to do in maya. It wouldn't be appropriate thing to do at this stage in a team environment, however since I don't really do any 2d concepting, I just mix in a bit of "playing around" like this which serves a similar purpose.

One thing I wont be able to do is have bazillions of unique animations with super tight hitboxes like you need for a good souls like game. Thats major reason I chose to keep it ranged combat only. But I dont want it to be a game where you just backpedal and funnel dumb enemies, or just snipe from a distance. I want to keep player constantly moving - it should feel like cat and mouse. So I am trying to make it so that once you get into an arena, there are no viable camping spots, and if the enemies can generally move as fast you can, then it should be such that you have to do a lot of weaving around the environment in order to create space enough to be able to get a shot off.

Well anyway, that's some of the ideas I'll work towards. Next up will be some enemy designs and animations and after that should start to see some basic gameplay.

I'm happy to receive any critiques or suggestions along the way about art designs, animations, anything really.

Alex_J

Okay, got some actual gameplay going on. Although I am just sticking to typical mythological creatures for the enemies, of which there is no shortage of concepts I could pull from, I wanted to wait until I've actually had a chance to fight them quite a bit before I consider how they might look. Now I have played quite a few hours against these guys so I begin to get a more clear vision how their silhouette should be, how they should move, what colors to be, and so on.

You'll have to use a bit of imagination, but the standard unreal mannequins here will become greek hoplites-turned-zombies. Just cannon fodder beginner enemies. They can kill you fast though if you let them get close.

The big red will be a satyr, which actually wont be a big heavy enemy whatsoever, but rather agile. It makes pretty wide flanks before charging, so I think some animations that make it look like its springing on its little goat legs during those flank movements would be nice.

The blue guy with a gun will also be satyr, but a skirmisher variety that flings javelins.

https://youtu.be/3Jemjfsw5Rw

This previous week was almost entirely code focused, but next phase will be designing these enemy characters and getting some basic animations blocked in. For now, similar to the character, I will forego polish and just try to get the art pipeline intialized so that I can easily iterate on them throughout the project, and for now the major artistic focus will be on shape language.

The gameplay seems like it is most fun when its focused on crowd control so I think being able to easily identify targets within a chaotic crowd will be important. I think silhouette and color will be the best ways to accomplish that so that will be first priority.

I'll also start a second pass on the level blockout - the first thing I've realized is that most of the arenas will need to be larger and more open. The game gets pretty stressful in tight spaces so I don't want to do too much of that.

Bolovorix

Seems like you're off you a good start, looking forward to seeing more!

Fabi_G

Looks like a fun project! Cool setting!

Played Resi4 recently and really liked the way crowd control worked there: slowing down a crowd of enemies by targeting certain areas, when staggered open for a follow up attack dealing area dmg.

Ability to place traps would be interesting too. And friendly fire (Medusa creating lots of new statues :P).

Regarding the visual design, maybe you can find solutions that allow for iteration and avoid uncanny, something like mannequins or statues. Or more abstract like Ashen.

Surely you planned many things already, looking at the thread just made me very excited, hope you don't mind :D

Keep it up!

Alex_J

@Bolovorix thanks!

@Fabi_G , thanks! I hadn't considered player placed traps, but I have sketched out some ideas for environmental traps. I was thinking that some enemies might not be possible to defeat directly with your firepower, but have to be lured.

Regarding visual design, I took a look at Ashen and thats definitely a neat look. I feel that I lack the imagination for much stylization though. Even as a kid I pretty much only enjoyed "realistic" styles. I don't doubt that realism is not the best choice of styles for a developer like myself, but without a concept artist to guide me, I don't think I could do very well at it.

I have a Gorgon boss planned and there is friendly fire with enemy archers, but now that you've mentioned it I am thinking about the glowing eyes attack and how it might turn the cascading horde into stone at the very last second as you cower behind a pillar, lol.

Few notes about development:

Took a left turn before getting back into level blockout. I was feeling pretty certain that the gameplay lacked a needed mechanic, and if I added the mechanic it could change things up enough that it might effect level design needs.

When playing, there was too much back peddling - felt too reactionary and not predatory enough. I dont want to feel like the doom slayer necessarily wiht 100% push-forward type of combat, but also dont want the other extreme where it is just running backwards to lead dumb zombies into a funnel. That's not fun.

So I added a javelin which is more powerful than bow, and you can throw it without having to stop moving. But less range and you can only carry one at a time. Charge the bow will drop the javelin.

Some messing around with it:

https://youtu.be/vfc8bsWl-LE

I feel that it plays nicely with the way bow combat works - they complement each other pretty well. It gives more opportunity to be bold on the offensive and gets you into situations where you have a little dance with the enemies which is when game is most fun. To hell with caution! lol

As you can see, animations are very rough. This is because 1, I am not very good at animations yet, and 2, I am trying to just do bare minimum needed to service the gameplay at this point. At first I wasn't even going to make any animations and I just had the javelin fly out from in front of the character, but I felt that it was difficult to get a guage on when you had javelin.

Alex_J

Here's some play testing first level after blocking it out in unreal

https://youtu.be/nduJOeh21uk

The basic synopsis is:

• you are an amazon warrior, it is early bronze age
• on your way to athens, something terrible has happened, the city is sacked, the soldiers are like zombies, and monsters are doing monster stuff
• fight through the streets, at shrine of asclepius an NPC will provide motivation for why you need to ascend the acropolis and kill the biggest bastard up there desecrating shit

I have two cinematics planned for this level, one at the shrine and one prior to the boss fight, but I will likely save that until the very end (of production).

There is a lot of "final" art even though its really just a blockout. That's just because I've used some of the many free assets from unreal marketplace. I figure it there is some stuff ready-to-go that it close to what I'm going for, getting that much closer to final look sooner is good.

I'm pretty happy with the pacing, difficulty, and fun factor and feel like it's ready for other people to play. Before I bring to public though I will add a few more behaviors for the enemies, and get them to have at least a blockout mesh so that you can understand what they are.

The biggest task will be animations for each enemy type - though for now they only need basic locomotion, a single attack, and an additive to represent when they are wounded. Since it is likely that much changes this early on, I will only bother with bare minimum animations - just enough to convey the gameplay.

There is also some unique game mechanics that I doubt anybody will pickup on without some visual feedback. Namely the bleed-out mechanic. Once you hit an enemy they bleed and will eventually die (even if you shoot their pinky toe). How long it takes depends on where you hit them, and what type of ammo. But without a particle effect to show the bleeding, plus animations to indicate that enemy is wounded, I think people will think that enemies randomly dying is a bug.

I have not blocked out the last part of the level, at the foot of the acropolis and then on top of it. The remaining gameplay in those places will have unique enemies requiring unique code, animations, and models, so I think I'll forego that for initial playtesting.

sacboi

The main problem I want to get around is that there is too much work for a solo developer. How can I make smaller games?

(…I'd assumed was rhetorical?!)

However TBH, I think it's all relative.

Basically in my opnion probably determined by ones character traits/workethic i.e. what might be a personally reasonable investment of time and energy on this piece of software may not otherwise suite someone else, in terms of scope or even whether paid or free-too-play passion project because I think working towards eventual finalization, is reward enough although not forgetting someone actually having fun playing what you've made, in the first place.

That said, solid start - but lol looks like a ton of work lucky me, I just like modeling stuff with hard edges :)

Edit:

You'd mentioned coding, just curious if referring too Maya scripts or something else?

teodar23

If the idea is to use this exercise as a way to learn and practice something you enjoy (eg animation) then keep at it. As they say, you do you.

If this is meant to become a game at some point i suggest you reconsider before you comit to such idea.

Reasons being (brace yourself!):

- it doesnt look fun to me, maybe because its not quite my type of game, nevertheless it should be fun and this doesnt look like fun. What it looks like is a generic third person from 2005. And its not just my impression, there are multiple wip games here on pc and those get more attention than this.

- its a huge task, so yea, going solo you either have to pay to outsource some of the work (which kinda puts under question the "solo" aspect) or scale down. Scaling down depends alot on how good the idea is, ie. if the core gameplay loop is very good then the rest of the game could be somewhat barebones.

Of course you can automate some tasks to make your life easier and take advantage of the huge library of free assets but still, this looks kinda ambitious and before comiting make sure you have a good base idea.

Its easy to fall into the trap that if you can quickly build up a prototype, you can also make a full fledged game. The reason its easy nowadays to build a "game" is because someone else already did most of the work for you (eg the game engine). And you are doing things at speed not taking into consideration best practices and optimization. There's a saying that basically states when you're 90% done you have 90% more to go. When youre doing a prototype you do 10% thinking you did 70% or more because everything is "almost final".

The difference between a triple A and an indie/solodev game is that of quality, most of the time. So knowing your game will lack polish and the art quality is so&so, what you are left with is the gameplay. That should be above average at least, to get any attention.

To summarize, im not saying you shouldnt do it im only giving a heads up as someone that has multiple prototypes done but no game finished, reason being, its a different type of beast.

Alex_J

@teodar23

all valid concerns, but this is my third game. This was winner from several prototypes I did because this project is more aligned with my favorite disciplines in game dev. Also bear in mind, I've put about 3 weeks of work into this so far.

@sacboi

It is a lot of work but I enjoy it.

Fabi_G

Did you think about first person camera? Painkiller, Dark Messiah and Thief come to mind as FPS with projectile based weapons. Any gameplay aspects that make it neccessary to be a third person camera? I imagine switching has potential to reduce the amount of work: less player anims, room for fakery, less camera struggles in narrow environments, etc. Of course depends on execution and will likely introduce other challenges :D

Alex_J

To be honest @Fabi_G , I haven't really given that a proper consideration.

The inspiration for the game started probably waaay back when I first played Morrowind. I always played archer builds (same in subsequent elder scrolls games) but I wasn't completely happy with how it worked in those games.

Later there was Demon Souls and the rest of the Souls series and again I always enjoyed playing archer builds, but it was barely feasible in those games. I wanted to be very fast and nimble, but have that challenge of aiming a slow firing, physics based projectile without any sort of aiming aid.

Mount and Blade offered the most compelling archery to me - the only problem I had with that game is that I really like the structure and focus of smaller, story/action focused games more than sandbox.

So this game is the answer to that game I've been wanting for so many years. The common theme across all these games is that its always third person. Hard to nail down why but I have just always enjoyed that the most.

So far the major downside to third person is like you mentioned, it doesn't play great in tight spaces. Of course level design can be built around that and so far I think I've been able to position things such that there isn't too much fighting with the camera, though it definitely does introduce some design restrictions.

My major beef with first person I think it that it seems very impersonal. Like, you lose touch with the character and then I feel like there is a subtle shift in way you experience the game. Its more about you versus the situation, whereas if we see a character dodging and sprinting and mightily pulling the bow, it seems to make me think more about the character versus the situation. More like when you read a book and them empathize with the characters, versus reading some dry history that just says what things happened. If that makes any sense.

Also, to be blunt, I feel that a strong female character adds sex appeal, which is surface level but probably the biggest selling point can be added to a game for the most likely audience. Aside from that, female heros seem to be popular right now, however I feel they are usually done poorly. "Mary Janes" and such. So I wanted to make one that was what I consider a proper hero.

I've emplaced some code to detect directional changes. From that the next step is to have some animations which trigger so that the character will make foot plants and such to help make it feel more realistic and modern. And even though this will slow down turning, I think it will make controller feel more agile. I'm pretty sure that should work but if its a huge hassle I'll look into using Motion Symphony which will simplify things a lot - only reason not to is because I'm trying not to spend any money until at least vertical slice is done.

I'm also going to experiment with some root motion dodges. I don't want to have a dedicated dodge button (like souls games) if I can avoid it, but I have an idea where if we are moving at top speed and have a sudden direction change, it could automatically perform dodge in direction before changing orientation. Sort of like a football (american) player doing juke manuevers but traveling in same direction. I am not sure if this will work or not so its an experiment.

Reason I want to avoid extra buttons is because I want to keep the challenge pure and simple - like old school mario, anybody should be able to pick it up and understand how to play in 3 minutes and from there it's just seeing how far you can go using those same core skills against ever increasing challenges. That's not to say I'm against adding a dedicated dodge, but it does open can of worms so I avoid it if possible. Anyway, that's a tangent.

Fabi_G

Ah yes, good if such a project aligns with ones interests. Forgot you also wrote that you're interested in animations. Based on my own (small) experiments, I just worry when much weight is on visual/non-gameplay features - will the whole thing die if they can't be realised within budget? But I suppose that's also game development in general: hard to execute a plan start to finish, at some points re-evaluation and as a result from that redirection is necessary.

On the animations: I think the running animations and/or transition to aiming in the last video looked a bit awkward, wip status surely plays into that a fair amount too. Could be in part due the sudden turning of the character between walk/run direction to look direction when aiming. Maybe the option to walk/run (slower) while aiming would make it feel different. Image player going backwards while aiming forward at the enemies towards them. Watching some videos of people doing running/waking archery might give some ideas on possible state.

Got curious about other third-person games with archery combat 🤓 AAA examples I could think of are Horizon games and newer Tomb Raider games. An UE3 game that had some archery-third-person-combat elements was "Hunted: the Demon Forge": B-movie vibes, bit repetive, but interesting for a short time. Results of a steam search with the tags archery + indie + third person, I thought looked interesting, although quite different to your project, were "The Pathless", stylized, reminded me of Shadows of the Colossus and "Project Sparrow", a competive PvP arena shooter built around the paragon character. Overall very few games using those tags.

Looking forward to see how it continues. Keep it up!

Alex_J

@Fabi_G Yeah for now I only use minimum time to block in the animations and not doing any polish. There are no transitions at all so its a bit janky. But playing feels smooth and responsive so getting animations to fit will just be a process of iteration. But its only a single main character and I don't have any special abilities or anything beyond basic locomotion and the the two weapon types, so I'll have plenty of ability to just keep refining it over the course of development. For now, I mainly focus on getting everything (both code and art) to have some basic placeholder in, before I go into detail at any one area.

As for project budget, I just try to get it finished in ~1 years time, and I do all the work myself. I don't have to sell the game to pay mortgage so there is no stress about that - but of course the goal is to make a game that people are willing to buy and tell their friends about. There is too much involved with making the entire game that I pretty much don't think too much beyond the design of it and technical problems. I just trust that if its a game that I thoroughly enjoy playing, enough others will so long as they can find it. That's reason why I want to get a vertical slice done first - not because I seek publisher help, but just so that I can start showing off something interesting as soon as possible.

I would like to hire a musician at some point though I wont think about money until I have something proper to show off. Then it will be easier to get help if necessary.

—

Today I've done a little work towards getting my own model in for one of the enemies. This is to be the Melee Satyr (in the videos, its the huge one that rushes you):

First time I've opened zbrush in well over a year. Thankfully not much has changed so I was able to get around.

I am only worried about the basic shape at this point - the sculpt is a bit blobby and unrefined but its a good chance that I play quite a bit with this guys shape. One thing I've learned is that its pretty much pointless to evaluate teh model in zbrush or maya or marmoset - anywhere other than animating in the game engine and seen from players perspective. It can just be so much different from that unique perspective. So I try to just whiz through a basic sculpt, I won't even retopo it for now - just get a quick rig and some basic animations going and start fighting the enemy in the game. Then it will be easier to figure out what it really needs to look like.

Aside from his shape though is how he moved. I have an idea that he runs kind of chimp like, and then makes a loooong leap for his attack, so the stubby goat legs and long ape arms I think should play into that. But we'll see!

This enemy will probably get pretty long fur cards around his beard and neck. Sort of like a big cape. I want it to look pretty wild and beastlike since it will be a slightly rare and dangerous enemy.

The concept I'm going pretty much directly off of is this:

Though I feel it's a bit darker than I will likely prefer, so I'll end up tweaking a bit. This guy also looks majorly dangerous, but there will be a more powerful miniboss in the level (minotaur), so I don't want him to look like he's the baddest dude. Probably the size and shape of horns is the biggest thing that makes it look badass, so I may scale those down.

Alex_J

I've got two variants of the satyr blocked in. Keep in mind this is preproduction type of work - in addition to exploring basic theme and style for art, I am also developing the overall assembly line.

I decided to add an additional satyr variant. I had a skirmisher and a charger, but I decided to add an archer as well. Reason is because I was going to have a female variant just for variety, but I figured if I was gonna bother making an additional model it should introduce something unique to the game. The way I've setup the enemy AI code will make it easy to get some archer behavior that is slightly different from skirmisher behavior, without having to do a lot except for change some data entry. So I figure why not give it a go and see if it adds any fun.

As you can see I've done some quick texturing. Pretty much just toss on some smart materials and paint a mask. I just want to start developing a feel for colors. Right off the bat I feel that we may want some hue variations because I think many people will find it hard to pick out enemies that blend in so well with the background.

The darker satyr is the heavier variety that rushes for melee attack. It is about 25% larger than the main character, however I found that in its four legged run position it feels quite small. But I won't change anything yet, because it is meant to have an epic mane and cape of fur, so that could make the difference.

The female is to be an archer, which will move less than the javelin-tossing skirmishers but have more range, plus arrows fly faster than javelins so it's shorter window to dodge. So just a little twist on how you have to engage with them. I decided to model her a little after something like a pronghorn deer - perhaps rather than running, she makes big bounds and jumps up in the air when you shoot her (check out some hunting footage if you don't know what I mean).

The gold disc is just place holder so that its not furry porn. Something like a kardiophylax for now. I am not sure what sort of clothing or armor the satyrs might have, I haven't thought too deeply into enemy lore. I am also just reusing the same hair from paragon that main character has, though I think it will look best to have the hair match the leg fur. Right now it just looks like an indian woman wearing animal print leggings, but it should look like a wild animal that has some human semblance.

I'm jotting down notes as I go for things that will be needed so that each department will have a master checklist. For instance, I'll either use fins and shells technique for the furry legs, or maybe just fur cards. I expect fur cards may be more simple, because a fins and shells or maybe a parallax fur would need to be masked from the skin - that's all material graph stuff which I am not very familiar with. Whereas just placing fur cards and letting them have a separate material is straightforward. But for now I only focus on identifying needs, not digging into the details.

I added a tail to the female - not sure why. I think I'll chop it off. It makes me feel too much like its a furry game. Not that there's anything wrong with that! I just don't want to confuse anybody. The other satyrs won't have tails though so not sure it makes sense for one but not the others. Big decisions!

Next up will be animations for the satyr archer, and then model + animations for the final satyr type - the skirmisher. Once all those are plugged in I will probably take the first pass at level geometry after a few rounds of play testing.

Alex_J

https://youtu.be/mdCZePVvFRk

• Integrated two new satyrs.
• added combat movement (crawl, duck, dive) to help in dealing with enemy archers.

I couldn't find any reference for the dive animation so just made it up from memory of army times. It's fine for now, makes the gameplay work how I want. Naturally, most of these animations need transitions from one state to another, but I'll consider that a polish phase thing to do since its only visual .

I'm not excited enough about the basic zombie enemy to make custom model and animations for that. need change of pace so I'll get started on replacing some of the level geometry. Probably will poke at cleaning up some enemy AI bugs here and there as I go. For the most part they are not bad though, just occassionally getting stuck but that's more of an issue with level geometry than code.

Alex_J

Okay last time I posted my next plan on the agenda was to start refining the level geometry. However, I realized that I didn't feel confident enough about some of the gameplay to be ready to commit to making art. The ranged vs ranged combat I felt was pretty solid, but fighting pure melee enemies left a bit to be desired. Its no fun if you just run backwards and shoot. So what can I do to make that more fun? Well, I've done a few things that I think work pretty well, but I'm still working on it a bit further so will talk about that in future.

In process I ended up reworking all of the enemy AI code over a few times... wasn't just about the gameplay, I'm also trying to settle in on framework for the AI. It doesn't need to be great code, just organized enough that I can easily debug and extend it without a big headache. Anyway, I feel that the gameplay is starting to get to where I imagined it, and I am having an absolute blast with it. The additional dive, crawl, duck animations add a lot.

There is much to do still, as you can see it is quite rough looking, but I feel the fun factor is getting there. Its not going to be a game for all people by any means. It's quite challenging. But I do have some difficulty settings worked into the mix already. I've made it such that difficult effects aggression, speed, and accuracy of enemies, but never healthbars. Basically it plays the same no matter the difficulty, you just have to be nearly perfect to survive on the higher ones.

I've tried to capture the terror and excitement of a close range firefight. A few design ideas I'm kind of forming are like:

- No safe spaces. Player should never be able to hole up in a corner and play peekabo and just exploit typical video game AI behavior

- Every method of movement should be necessary to use, feel good to use it

- within arenas, should not hit any dead ends. in other words, level knowledge shouldnt be necessary to survive. like, should be no hyper-advantageous spots. Should be possible to be surrounded anywhere but also always have some escape. That way you can always move and feel like you have a chance, but also always in danger.

- every enemy kill should feel like big accomplishment

- replays should be fun all of their own, no progression mechanics necessary (of course I intend to have upgrades, secrets, and such, but just trying to master each arena should be a satisfying journey on its own)

So those ideas help me figure out how to tune animations further, and especially will drive the environment designs. That's major reason I didn't wanna get into it too soon - I am not 100% certain how exactly the levels should play out. Arena to arena, narrow corriders, lots of cover, long sight lines, etc etc.

I'm also eager to design the characters a bit further as I am feeling pretty confident about at least of a few of them. Simply by playtesting against them I am getting a strong feel what they should look like, how many variations are needed, and so on.

I think being solo-dev I should focus on just a few enemy types with broad usage. I found that by just tweaking behavior of the Satyr Archer slightly so that when you get close she has a 50/50 chance to either retreat or charge, that makes her way more interesting to fight and also reduces need to have other dedicated charger types. Just by random chance sometimes you get one who gets stuck in a cycle of charging you and it becomes a real problem. So then when you go into the fight you have incentive to really control and dominate the arena. If you let things get out of hand its a quick death. I feel it hits a good spot where replays are welcome and not frustrating, despite the difficulty.

Some playtesting footage:

https://youtu.be/PpWvSpIh6GU

Not sure what you will see next... Getting to where all the things to do don't feel like foundation blocks, so I'm a bit hesitant... don't want to go too fast, and realize I forgot something important, lol. Probably work on characters and animations a bit further if can't identify any other higher priority needs.

Alex_J

Was looking at some ways I might improve the character model and I started looking at metahumans, and then spent some hours figuring out a workflow how I can get the parts I want from it to improve realism, but get rid of the extraneous resources and complexity... what a rabbit hole. I started to think, "this is the opposite of what I need to be doing. Half a day spent with nothing to show. That's not the kind of days I like to have!

So swing in the other direction and I start to ask, okay how can I make it so that authoring art is really simple and requires minimal resources? Cause I am just a lonely developer, I am not super technical, so I don't want to make something that requires deep tech to keep it running well.

Regarding enviro art, I had a notion like what if it's high-ish polygon models that use nanite, and for texture just a single color texture, and it doesn't even need to be high resolution? Cause thats where the project gets bloated, from textures.

First thing to try is grab whatever free or cheap post process things were available on marketplace. I dabbled with a few painterly effects that I think looked beautiful, but it makes the game look too abstract and a bit hard to read, especially when going fast.

I tried a number of outliner shaders, toon styles... wasn't quite convinced though I could tell its a good direction because it accomplishes goal of simplifying the overall rendering pipeline and resources to support it, it takes what looks like C - grade realistic art to something a few degrees closer to looking like a legit game with basically no effort, and it also makes the game more readable - like seeing arrows fly and spotting enemies is just easier on the eyes. So lots of pluses.

But pretty much every shader I found on marketplace doesn't really work for complicated scene, or naturally includes many things I don't want and not everything I do. So I took some time to follow a few tutorials and learn a bit about post process materials, and ended up making a few. Thankfully there is some amazing teachers on youtube because trying to think in math is like pulling teeth for me.

The materials I made from tutorial are:

- outliner - from scene depth and world normals

- hatching effect which replaces shadows/occlusion

To these I added a simple scene depth fog which also slightly desaturates based on distance. I found that the scene looked a bit flat so this very subtle touch of realism seems to help with that.

I added some lerp between tiling amount of the cross-hatch shading based on distance as well. The problem is that larger hatches look nice on walls, but not on the character. This solution works ok, but it needs some more robust blending, or perhaps I should look into being able to mask out characters, and apply a separate solution for them since I want to keep them more legible.

Development speed is priority right now so my initial idea is just a few basic shapes for walls, doors, etc, grab a displacement map from quixel and convert to geo in zbrush. Decimate as far as possible and then nanite can take care of the rest. Since most the things are noisy organics and we don't have a full range of lighting and microsurface to show every flaw, I think this should work out okay.  I've test a few cubes but will build actual stuff in coming days.

post process distance fog and desaturation compared to without:

they both look fine, I do like the comic book feel of without, however when moving throughout the world it feels a bit flat and seems a little of the wrong tone

zetheros

shader looks kinda jank, maybe take a look at Genshin or Honkai Impact. I think it looked better without the shader, and decimated zbrush models are terrible for actual in-game assets.

Art is always never simple, because it doesn't matter if you're doing pixel art or PBR - there are methods of increasing the fidelity of both, and the bar never stops lifting. You'll just have to decide what artstyle you prefer, and when it's 'good enough'.

I think something that would be beneficial at this stage of your game is to write a master document of your game and establish good workflows for assets and programming that can be improved upon over time. Development speed is not necessary since you're indie and not subject to the whims of a publisher or investors; take a breather, step back, and organize your thoughts and priorities. It's very easy to get overwhelmed while working on a project like this solo.

Alex_J

@zetheros

I've done that already - the planning. What you've described is kind of where I am right now - developing the assembly line by conducting research and doing experiments. Along with that I am also iterating on game design, starting from the master plan and trying to verify that all of it makes sense in practice. The first goal is to realize what was originally planned, and from there either change it if needed, or just polish it if its good. So far I've felt that the original vision is holding up to how I thought it would, which is awesome!

I just keep my write ups simple here and just focus the little uncertainties that spring up, but every detail about the game is clear and has been iterated on with paper planning pretty extensively. But there is only so far that can go before its just spinning wheels.

 Why I favor speed at this point is because I aim to test out my ideas and answer the unknowns. Many things change once there is actual gameplay to deal with, compared to paper planning.

I took a look at those games and they indeed look fantastic. That is quite a different style though and not one I think I could do. It is a legit art style requiring people specialized in that style. I am more of a "pretty okay at realistic 3d guy" but seeking ways to simplify things down a bit. I also think that it is important to look at and learn from the best in the biz, but I am a solo developer so you are right, I have to be ready to use the "good enough" stamp much sooner than a large team full of specialist. In general, I try to aim for a workflow and style that allows me to create one rigged character in a day, and basic animations in another. That's just for an initial pass, but ideally with just a few more passes the characters would be considered finished.

I did some testing with the "displacement to polygons" workflow to generate high poly models and it looks pretty good, however flexibility is lost, so I will save it for final art pass. What I mean is, once you convert something like a brick wall to geo, now it can only work with that material. So for I keep just flat shapes so that I can mix and match materials at whim in engine - then finally once I've settled on what looks best, I can convert to high-poly meshes as needed. That would most likely only be close to camera stuff, as once you are like 10m+ away its not even noticeable.

I disagree about speed not being important - there isn't infinite time and as a solo developer the major issue is getting too spread out which causes fatigue. So when I talk about speed it's not about racing to the finish line, but rather being super strict about prioritization so that I mitigate the amount of work that has to be redone or scrapped entirely. So I try to be super clear each day about what is the important question that I am working on answering currently, and is the work that I am doing going to answer it? So even though I might want to go a little further to make some art look nice, I know that with important gameplay questions still unanswered, it wouldn't be good to do that because there is still a high chance that I'll end up with something that won't be used.

About the shader and general art style, do you have any suggestions? By jank, do you mean that it reads too busy, or just looks unrefined? Something else? Something you think might fit better with a game like this? I've felt that it shouldn't be too toony because it's a pretty hardcore game - like right now I'd consider it to be above Souls games in terms of how much attention it requires. I'll likely make it more accessible than that but I do intend for it to be a game that has that sense of severity to it even on easier difficulties.

I do think that the general comic book style could kind of work because it is not unlike super hero comics in terms of story. Basically we have a hero who is fighting evil and that is all the depth to it - no under or overtones - just about the characters facing danger. So something like a little bit more realistic comic book style I think makes sense?

I was happy enough with realistic style as well - especially in cases with fog, smoke, lots of atmosphere, and then sometimes the enemies appear from it, or are obscured. It can feel quite intense and badass that way too.

Would love to hear any additional thoughts about general art direction - now is the time where I am focused on experimenting with that and I am not 100% settled on anything so far.

edit: one thing I noticed in Genshin is that the outline only seems to apply to characters, but not the environment. This seems to help with readability and makes it look cleaner. That is something I could test pretty easily so I'll do that soon.

Update: Here I've masked out just the characters to recieve a thin outline. It improves readability and I think bring focus onto the silhouette a bit more, which I think makes it more attractive. Then for the environment, I enabled a Kuwahara filter which just adds a little bit of painterly feel. I really like that a lot and it would make authoring some assets a bit easier, as you don't have to deal with anything very high resolution - it feels a little more abstract.

I think the greatest benefit of the outline is that it helps make faces more readable from a distance. I have to edit the main characters face and especially the eyes to make it work with the shader but you might already be able to see the benefit there.

Keep in mind nearly all environment art is placeholder, so you have to imagine art that is a little more tailormade to work with the style (you'll see a start on that in next update)cted:

 

Update 2:

Another style to compare with. Here it is character outline + painterly filter + 6 tone cell shader. I kind of like this a lot! It is very easy to read and I think makes the foliage look beautiful in particular.

Fabi_G

I like your latest experiment with the style much better than the one before, less noisy. I believe if the level assets have less high frequency everywhere and instead some resting areas, that might further improve the look. 

Some thoughts on the last video: The player always stopping while aiming and shooting and the slow recovery after a dodge break the pacing a bit. Would look cool if she makes a sidestep while aiming to dodge a projectile. I think having the action look fluent to someone just watching would be great to generate interest.
I also think the enemy archers jump too much (too high?), so it looks a bit comical. I would expect creatures like that to have a bouncy step but use jumping just to get over obstacles. Maybe they shoot, then change position (to get in the players back?), sort of like a sharp shooter.

Keep it up!

Alex_J

@Fabi_G

Thanks for feedback. I think I agree with all of it. Looks like the level geo will have me busy for awhile but eventually I'll be rehashing the characters and animations a bit.

I'm planning on having a reward to improve the dive so that you do an immediate recovery so you maintain momentum the whole time. Hadn't thought about a sidestep while aiming but i can see how that could work. I do want for movement to break aiming because it changes strategy a ton but a snappy recovery would definitely look a lot nicer and make character feel more badass. I just don't want it to feel like a shooter with corner peep shooting as primary tactic.

Alex_J

Okay, a little bit of headway has been made for the level geometry. I have done a bit of enviro art before but not at a scale like this. So I had to do a bit of experimentation to figure out how I'm gonna handle this all.

The first big question is, "what do I actually need?" That stymied me for awhile so I decided to just get some paint on canvas and work things out from there. So I just put aside workflow concerns and just got to modeling. I made a couple of public buildings like we might find in the Agora:

Just model them in the fastest, simplest way that I know. I put into unreal and try to get a handle on scale of things. What are the largest units that could be used to generate a variety of buildings like this? In other words, find the greatest common denominators.

One thing is that these are in-tact buildings, but most of my buildings will be at least partially destroyed. So with that in mind, much of the largest my compositional pieces are actually true to real life - that is, individual stone blocks.

From these I can create larger compositions. Blocks into walls, and then walls into buildings. But then I should also be able to "deconstruct" them pretty easily. In Maya I am constructing those compositions (like the wall) from instances of blocks. This way I can update the geometry or UV's later without needing to rebuild the entire thing. Can just convert from instance to object before export. It also keeps maya running faster.

For now all the individual unit pieces are in-tact but I'll create a few variations of each, some of the variations being damaged.

Much later I could convert things like these walls to a single mesh by baking it down, however I am not sure that would be any benefit or not. Less geometry but it would make for a lot more materials and make workflow more complex. But it's non-destructive thing I can simply try out later.

A few more unique meshes as well. Will need some statues but I'll buy those. Would take ages to sculpt and there is plethora of nice scans for ancient statues already:

I plan on building the buildings in Maya because I dislike the editor selection and snapping tools in Unreal comparatively. I may go as far as actually laying out the level in maya as well. It is possible to export entire scene as a single fbx and then have unreal break it apart into individual static meshes. This way layout can be made in maya much easier, but in unreal we can still tweak final position of individual meshes (and of course for culling and such to work it shouldn't be a single mesh).

Trying to be minimal with materials - for instance each of the blocks faces can be moved around a texture like this:

Or it can just rotated. This way get a lot of variety but its all from same material. The resolution is low like this but it seems like it may be okay given this art style. Of course some custom textures can be made so that resolution can be more optimal and efficient if needed but this works for now. Another thing that could be done is atlas more material types together. For instance, wood and another type of stone could make up the unused portions of this texture seen above.

Well it took some time to get the ball rolling on this but now it looks like everything is squared away so I can dive into actually producing some shit.

one additional thing I learned: You can convert displacement to geo in marmoset toolbag, which is much simpler process compared to either zbrush or maya. I won't be doing this now but if I eventually convert level geometry to high poly that is the best way I've found to do so. It's just a bit hidden in toolbag, there isn't a named tool for it. You just subdivide the mesh, apply displacement in the shader, and then if you export that the displacement is applied to the mesh.

Alex_J

Okay I haven't reached next milestone but learned quite a bit in last week.

Initially I was turned off by Unreals viewport tools so decided to do level blockout in maya instead. It is definitely feels a lot faster because you have many QOL things like isolate select, invert selections, full suite of modeling tools, component selections, layers, a billion ways to snap, MMB movement without needing to select manipulators, etc etc. Just really easy to get around compared to Unreal which has bare minimum comparatively.

But there are a few reasons why blockout in DCC left me feeling uncomfortable. You don't get immediate confirmation what changes look like exactly as the player would see. You can export pretty quick and double check but having immediate feedback is nice. Another problem is that with complete modeling power ready to go, it is very easy to start over-detailing.

So I did a ton of back and forth, but each way has big pros and cons that make me immediately want to change workflows. I need to get a method down and stop wasting time!

I know that pro level artist tend to work within Unreal - I figured maybe they just don't know any better? But more likely it's something that I don't know. So I bought a complete enviro art course from Thiago Klafke which I can highly recommend:

https://www.thiagoklafke.com/

Very nicely breaks the process down into stages with a clear goal for each stage. It is production focused, not portfolio, so I appreciated that. He shows full process of actually doing the work in each stage so plenty of tips for working faster in Unreal. Those are exactly two things I needed: a clear methodology so that I can know what the major steps are and when do I know they are accomplished, and then tips for how to work the Unreal way so workflow isn't too tedious.

Thiago does a great job giving clear instructions what the goal of each stage is, how to avoid common problems, and how to breakdown a large project to make it more manageable. It is really the same principles as if you were drawing a head or sculpting a body - but it helps a lot to see in action within different discipline.

So I had a basic level for Athens blocked out already. From that I determined the enemy encounters that I wanted - roughly how they should play and be metered out. I wanted to mix up the level a bit though. All of the other maps in this game will be pure make-believe, but this one is set in Athens, so I felt like having it at least vaguely match real world layout would be good. Historical accuracy is not important whatsoever here but but at least a few of the major landmarks in relation to each other I think is worthwhile just for a sense of authenticity.

It's hard to know where to start, so I just block in the major borders, major landmarks, and then the biggest buildings. First I am only trying to block in the gameplay area. I drew out quite a few alternate paths - most will be blocked off however I'll keep my options open until I find the best gameplay for each encounter.

I will try to get one zone blocked out per day. Today I managed like 1.5. Some zones are larger than others, but we get more repeating elements to reuse so it should go faster and faster.

The major zones are:

• Entry area - impressive guantlet with larger statues, but they've all been destroyed. For destruction I am using Pulldownit in maya, but for now I only place some placeholder statue models along path.

• Agora - largest area, once a lively market, now desolate, completely destroyed. A big archer battle happens here, so it will be pretty open but the crap all over ground is sometimes piled high enough you can dive behind it for cover.
  
• Fancy area - behind the Agora is a a fancy garden / bath + a destroyed temple where miniboss resides.
• Streets - More grim and bloody, full of dead bodies and zombie-like soldiers. the main path goes through labyrinth like streets
• Boulevard - beneath the acropolis, main road that is littered with smashed monuments - ranged harpies attack here.
  
• Acropolis - boss fight happens here

About 1.5 out of 6 major zones blocked in (not counting non-playable areas as zone)

The main way I am blocking paths is to have a pulled down building. I did some test with Pulldownit plugin in maya to see how I can create the actual models and it is really easy. Just shatter a model, do a quick physics sim for where it should lay - finished. For now I just make a pile in unreal though to indicate the spots.

I'm imagining Athens in earlier days than typically portrayed in media. Many of the famous buildings wont have been made yet - some are under construction. Here some public building in agora was under construction. My plan is that most of these large buildings can be derived as instances from just a few static meshes. A couple column pieces, a few different sized marble blocks... should be very efficient for loading the scene despite it being relatively big and open.

My plan is to get the blockout complete and then refine it for gameplay. To complete that 100% I'll have to build a couple new enemies, so it will actually be awhile before I refine the blockout. I feel confident that it will not majorly change and even if the layout is slightly different, I am still almost certainly using the same elements (like columns, marble blocks, etc), however... since this is my first time building a level like this and it's a large scope, I think it's better to be extra cautious and make sure I get the gameplay nailed down before committing so much time into art.

Alex_J

A few things to share that I've learned about level design:

The first question I've tried to answer is workflow. How will I actually make the levels? As a solo developer I have to find every little trick to cut corners however possible while still meeting my desired quality. So it took some experimentation but I think I've got a well oiled machine ready to go.

I decided to get some expert advice and bought a tutorial (link to that in previous post) which was great and taught me everything I need to know to do a level blockout in Unreal. But after spending a couple days working like that, I still felt unsatisfied with the workflow. Did not feel fast enough or offer enough flexibility.

So I revisited original hunch that I should do the work in Maya. The question is then, how can I solve the problems with the DCC level blockout workflow? Mainly that is not having immediate sense of scale and players perspective.

Maya does have a "walk tool" that allows WASD navigation, and then you can just place many human scale references throughout. That is still a bit abstract but it does at least allow you to know for certain that you aren't totally out of whack.

The other thing is just to be disciplined about sending work over to unreal each time something new is built and doing a quick play test. Honestly this seems completely fine, and the speed boost of using maya for layout compared to unreal I think is well worth it. I did a time comparison of blocking out one section in unreal versus maya and I was nearly five times faster in maya. Plus it is trivial for me to make any type of modifications - large or small - in Maya, whereas in unreal it just isn't possible in a lot of cases, or if it is it requires a ton more clicking.

So I've got about 75% of blockout done but as is natural during development, many things have changed since my original plan for the level. New questions arise and I find it difficult to answer them because it's been too long since I did inventory, so to speak. For instance, I have a checklist for what the main beats are, but a question like, "should the encounter with charger enemy type happen in this street, or that one?" can slow down progress.One way to answer such questions is to just look things over and go with your best hunch, and then later after much play testing we will weed out the bad decisions and then work on replacing them... but I feel like it's probably possible to setup a good template to work from so that these decisions can be made more accurately the first time.

Basically when making a decision, I want to make sure that I always have all relevant data ready for easy review. This way I don't forget something important. So a template to follow which defines each beat of the level should help accomplish that. And then how will I know that the template is working towards the goal of making it faster to make faster and more accurate decisions?  Basically after a decision has been made, there is a feeling a confidence when you know that you considered all angles, whereas if you know something may have possibly been missed, that can erode confidence and cause anxieties. Makes it hard to stay focused if too many decisions are made that way!

When initially laying out the level it's just kind of chaos of juggling a bunch of different considerations and then just slapping some paint onto the canvas. I know that there are some broad goals for the level, key things we want to clue the player onto at certain times, key lessons we want to teach the player, times we want them to have a break or some easy victories, etc etc. But once that initial layer is taken care of and I am looking at a second pass, I take a look at each beat and ask, "okay, given the goals for the level and for this beat, are they satisfied in optimal way with this layout?"

Don't want to spend too much time on these sorts of micro-decisions because there are too many to make. But they are important so don't want to trivialize them either.

One example is, "do i introduce harpy (ranged airborne enemy, very dangerous and difficult to defeat) in arena 3, or 4?

There's a lot of things to consider - pacing encompasses a lot of things like where average players attention span is at this point, are they up for a fresh challenge or they need a break, have they developed prereq skills by now or not, have they faced previous enemies enough to feel some confidence or are they desperate, etc etc.

There are artistic considerations as well. Ideally each beat takes place in an environment that frames it appropriately, and sight lines help to frame the long and short term goals when appropriate.

Well anyway, mostly stuff that takes a billion words to describe but probably most art oriented people already know intuitively. So here is the template I am working on:

Games are different from novels but I still think Dwight Swains Motivation Reaction Units are a helpful way to think about each beat. A nice summary of that here:

https://www.advancedfictionwriting.com/articles/writing-the-perfect-scene/

It doesn't always fit each beat perfectly but I think it still gets you to ask the right questions and ensure that each beat rolls naturally to the next and the question of player engagement should at least always have a theoretical answer. Of course play testing is needed to confirm.

Okay in next update I hope to have level blockout 100% complete and begin testing it thoroughly.

Alex_J

Here is a big workflow decision that is going to take some testing to figure out.

There are two main ways I can import level geometry meshes into the project. I can export an entire building from Maya and then Unreal will consider that a single static mesh. It may share materials with other buildings, but if BuildingA and BuildingB are both largely composed of BlockA and Column C, there is no benefit from instancing.

So the question is, how big would the benefit of instancing be? Would it make any difference, or enough to warrant a more complicated workflow? Because importing models to preserve instances is quite a bit slower... certainly at a level blockout stage it is not worth the hassle, but if the optimization is needed, it will be easier to import models this way from the beginning.

To answer this question I'll just have to compare one scene versus another. I have done a small scale test of one static mesh instanced thousands of times, versus same static mesh batched into larger actors and duplicated - there was no performance difference. However, it seems that the major benefit of instancing would be if there were many types of buildings, and then unreal views them just as instances of the same few components (like buildingA and buildingB are just seen as many instances of ColumnC and BlockB).

It is going to be double work to setup level with both styles, however I think it is worth figuring out for this first level because then for future levels I can just use the best workflow. If the instancing doesn't make enough of a difference to matter, then I can go with the much faster and simpler workflow.

So far the recommendation from forum responders is that instancing is generally better, but most examples pointed to are going to extremes - like unreals matrix city demo, or AAA titles.These are the guys to look to for best practices, however I might not be going anywhere near the extremes they are, so such optimizations might just require more time from me but not have big payoff. So it's worth testing out to find if there is a corner I can reasonably cut or not.

Another important distinction is that all of those examples involve level streaming, which I aim to avoid simply by having smaller levels with a lot of switchbacks to make maximum use of space. In other words, I should only load things once at the beginning of the level and not have to worry about things streaming in during play.

Example here is the Parthenon Building:

In Maya this is mostly composed of many repeating instances. I just export that with Preserve Instances ticked in export settings

In Unreal use Import Into Level:

Set is as a "Level Actor" and all the instances are preserved. Then they all can be Batched into an actor here:

and finally that actor can be turned into a reuseable blueprint.

This is not awful but it requires a lot more thought going into organization and also there is usually some hijinks with the batcher. So you end up spending a couple hours troubleshooting instead of churning out art. If it turns out that the optimization is worth it, then I probably look into writing a tool to handle all of this.

This is a bit thinking ahead - blockout is not 100% finished but it is very monotonous so needed to shift mind elsewhere for a break. Should plan on setting up the level art in the quicker way first - each building as unique static mesh - so that gameplay and art direction can be achieved first. Setting up the more complicated instance pipeline can happen after, but before any work on future levels is done.

Alex_J

Small progress update:

General project management:

Spent a day and a half preparing on boarding material for a music guy. I was lucky enough to snag a guy who is doing a final project for his college and needed a game project to contribute to.

Preparing info forced me to have to think about how can I breakdown project into very broad themes, try to be accurate and descriptive but concise enough to not overwhelm. In my initial planning I do a bit of this kind of thing, like write out one sentence summary - basic marketing shit like that - but I tend to gloss it over because I don't need to sell the game to myself and the big theme is clear in my mind. But it's probably a good thing to actually write it all out and iterate on the language I use because will have to do this later for marketing. It seems that thinking about the game experience from music perspective is also beneficial when shifting angles and thinking about the visual side of things to.

It's also a good chance to zoom out and reevaluate everything at high level. Since initial planning, is what I've developed so far still all pointing towards the target? Sure enough, some little corners of project have become outdated so it's good to clean those up. For instance, it is not necessarily a story-driven game but I always want for there to be a clear macro-objective to provide external motivation. As level layout changes, some of the story beats no longer works in ideal way, so just making a few changes in the running screenplay is necessary. I think it's a good indication if changes in level design can make for story changes that make it simpler - more clear objectives, more easily understood character motivations, and more dramatic confrontations - that things are shaping up into more cohesive experience.

Workflow:

Still working on blockout. Should be ready for testing in a few more days. It was getting tedious so took two days to do a little work confirming workflow for the next step after it - level geometry. Looks that setting up buildings via instancing in maya and then just import into unreal so that instances are preserved is going to be both the fastest workflow and also give best performance / smallest size for the game. Because pretty much everything boils down to just a few marble blocks of various dimensions and a handlful of other items. Tons of reuse is possible because limited materials and methods used to build at the time.

Blockout progress:

Agora area lacks all the destroyed market vendors but the major landmarks and boundaries are in place. I'll need to do cloth simulations so I'm just saving it for a day on it's own.

Acropolis also has the big landmarks and boundaries in place. A giant statue of Athena will be toppled and her sad busted head will greet our hero after she enters through the Propylaea (seen ahead). I believe the historically accurate statue was bronze but I want her head to be broken off, so she will be stone. Bronze, I imagine, upon falling, would bend / smoosh.

I will make a prefab out of generic buildings like this and place some locators along the roofs. The flying harpies will seek those locations as landing points. Missed shots will be painful because no way to retrieve arrow though, so I'll consider these to be high difficulty and have to pace it out carefully.

Going to rework this area a bit but the basic shape is good enough for gameplay testing. Will be much larger temple up there and the stairway will be blocked with destroyed columns / statues. Minotaur tearing place up, go kill him is optional miniboss. The fight takes place in a fancy public bath / garden area so I am working on a few layouts to hopefully make it look awesome and also try to develop some nice rhythm in the way player will have to evade the monster. Also considering adding a jump so that you could hop on broken pillars to get across pool without slowdown.

Within first ten seconds of level you are presented with choice - go left or right. I am using destroyed bridge as a one way only point-of-no-return. It took a surprising amount of thought to come up with some way to gate players movement in a way that doesn't feel contrived or require a bunch of unique animation or code. Basically you can hop down but not climb back the other way. Same deal with some crumbled buildings in other parts of the level.  This should suggest to players that it's expected to go through things a few times.

I am using pulldownit plugin for maya to fracture buildings and do simulation so the bits pile up nicely without me having to place individually. It involves a bit of waiting though and also I want to make sure the destroyed bits are all instanced as much as possible, so I have to save it for once my modular kit is pretty well decided upon.

The original Parthenon. This has little bit more than blockout geometry because i used it to test the instancing workflow. Right now it is composed of like a dozen different static meshes being instanced. The friezes I will see about doing a displacement-to-geo on so they look like actual sculpture. Nanite should help a lot with that but for the most part this entire thing should be pretty low poly since it's just made of basic shapes. Rectangles and columns.

Doesn't seem that hi poly stuff matters too much with nanite but it still slows down work quite a bit when passing between programs so anything that will be higher resolution I'll save for a finishing pass.

to the right of the original parthenon, the newer parthenon (the one that exist today) will be under construction. I will give it a partial roof because that is how you'll kill the boss. (crush him). It will also be the only spot up here where you have any cover to weave through to get distance from the monster. I consider swapping positions just so that it's a little more prominent and obvious, but we'll see.