next »

[GrumpyGiant]

I've been working on a game in my spare time for a while now, and I thought it was time to start posting something here. Right now it's about 1% complete, but I wanted to get it out there.

I've tried making indie games before and so far it's not amounted to much. I usually start with grand ideas, but then I always reach this point where the gulf between what I have and what I intended to make is so incredibly vast that I get depressed and give up. I'm a programmer at heart, and not the best artist in the world, and so without decent art to work with I've not had much luck so far.

People often recommend to start with just an empty box world, nail the gameplay first, and then worry about the art later on. Some people can work just fine like that, but I don't think I can. It seems to me that the art and the game are intertwined -- it's not just about raw gameplay, you need to create feel. You need to make the player feel that this world is real, and that the actions they take in it are tied to that particular world, and wouldn't be the same actions if the world were changed.

So I figured this time I'd work the other way around, thinking I might have more luck. I would build a world first, and place in it whatever felt right for that world. I'd focus on creating a feeling first, and whenever a piece of the world spoke to me I'd listen, and iterate on that part. Paint a picture, then explore the depths of it as I go. I decided to start with a simple goal -- to make a pretty forest -- and work up from there.

It doesn't have a name yet, nor will it do for a long time. I know there's going to be a forest, big monsters, and puzzles to solve. And unlike other indie games that promise Houselegs but fail to deliver, this game will provide 100% GENUINE HOUSELEGS  .

[Floresy]

Lookin good! The distance-fog fits the scene really well, are you sampling a gradient or is it just one color slowly overtaking all others?

[GrumpyGiant]

Thanks! Right now it's basically just a mix between blueish and yellowish depending on the angle from the sun. The HDR rendering makes the blues look whiter as they get more foggy though.

I quite liked how Firewatch did their fog, with a custom palette based on distance, so I might look into that later on.

[nathy after dark]

You're definitely succeeding in making a pretty forest. I hear what you're saying about getting too ambitious with projects and then giving up, and I wonder if you're sure this isn't getting overly ambitious as well. How do you plan on fighting scope creep when starting with so little a plan?

I guess what I'm saying is make you sure you ship something. The first time is the hardest (because we all start out with high expectations and struggle to meet them as beginners) but you've really gotta break the cycle of letting ambitious plans fizzle. If you feel that happening with this, I'd rather you release something much smaller than the goal, cause I want to get to experience it in some form no matter what.

That's just me jumping in and assuming you wanted advice--sorry if that's not the case! Keep it up.

[eigenbom]

Hahah yes houselegs have returned! Also this looks super.

[GrumpyGiant]

Nah that's fine
It would be easy for the scope to creep, certainly. I've been very aggressive over what I'm deliberately not putting in. e.g., there are -

• no characters
• no interiors
• no scripted sequences / dialogue
• no dungeons / levels
• no normal maps

Most of the work so far has been just getting the overall scene done, and the tech to back it up. I think I'm mostly done on that front now. I kinda want the game to be about getting the most out of a set of small ideas, rather than continually introducing more content.

I plan for this to be an ongoing project. If I do wander off, I can come back when I'm ready. The next phase is to start building puzzles, and see where it goes. If it goes nowhere, that's fine. I just want to enjoy the journey a little, and post things about it here too.

[davemakes]

Ooohhhhh, that's gorgeous. The atmosphere is wonderful! A+ houselegs, too!

[toblotron]

I like the feeling of the world, and have a great respect for your approach!

When I'm thinking "what happens here", one of the first things that come to mind are, how is the player represented? - How do we move and interact with things? Is the player a disembodied being floating around, with godlike powers (change terrain, fill valleys with water, etc), or is the player a human/robot with limitations on movement/ vision? (must climb to high factory chimney and use binoculars to see other factory hidden behind impassable peak, and see the secret combination to the door in the mountain pass, scratched somewhere?)

What is your vision on these things?

[GrumpyGiant]

When I'm thinking "what happens here", one of the first things that come to mind are, how is the player represented? - How do we move and interact with things?

The approach I'm taking is that I want to recreate the feeling of exploring the forest. I've been hiking a few times up in the Pacific Northwest, where you'd get these giant redwoods hundreds of feet high. Sometimes when you're alone, you might occasionally get this weird feeling of panic come over you, as though there were something looking for you (the word 'panic' actually comes from the god Pan who was said to manifest in lonely places). So to try and give the same experience, I'm making it a first-person game, an exploring adventure. And there's going to be big monsters houselegs looking around for you.

And as a kid growing up in England, if you walk around forest for any length of time you'd always find some abandoned bit of old Industrial Revolution architecture. You'd never know what exactly it was -- it'd just be some crumbly old bricks attached to an old water wheel, or a chimney stack that connected to a driveshaft that is no longer there. So I figured I'd try and mash these together somehow. Games like Shadow of the Colossus have that same kind of feel; a big landscape that never quite gives away the history of it. If I can capture any of that then I'll be happy.

Gameplay wise, Portal was a big eye-opener for me. The way it let you do things from a first-person viewpoint but wasn't an first-person shooter really stood out. Now I'm not planning on making a game with portal mechanics or anything clever like that, but I like the way it let you have complex puzzles using simple ideas. I also liked how line-of-sight suddenly became important, something that wasn't really a thing in puzzle games until they could operate in first-person.

I suspect the closest thing I'll end up to though is something like the Talos Principle. I loved in that game how the player himself formed part of the puzzle state -- i.e. a puzzle is made up of many pieces, and where the player is at any given point limits what he can interact with. Move through a doorway and suddenly any puzzle objects you weren't carrying are inaccessible on the other side, that kind of thing. The player isn't just solving the puzzle, they become a small part of it. I think that's something I'd like to explore a little more.

[GrumpyGiant]

One of the challenges of making 3D games is that editing and building content suddenly becomes very difficult. Most 3D world editors are annoying as hell to place things in. Having a grid-snap option can help, but it's still awkward.

So I thought I'd try out a simpler approach here. My world is going to be structured as a big open expanse, with puzzle areas scattered throughout. I love how when exploring a big forest you might suddenly come across a strange structure in the middle of nowhere. A lot of country estates and parks in the UK actually have buildings designed just for this purpose -- Follies, made for no reason other than to be interesting. Places like Portmeirion are full of these. You'll be walking down a forest path and suddenly come across a lonely structure hidden away from everything else.

I want to make my puzzles along that kind of design. Each puzzle is going to be within a small enclosed space, which sits alone in the forest. Given that each of them are separate and don't need to interconnect too strongly with the rest of the world, it seems I'll make it a lot easier for myself if I have a way of designing them that doesn't need me to fire up some horribly complicated 3D scene editor each time.

So I'm going to try using a 2D tilemap editor instead. I'll build each puzzle using just a simplified representation (pictured below), and then have a way of importing that tilemap into the 3D game and dropping it where I need it. The tiles will then be replaced with 3D models.

There's still a lot of possible issues with this approach. What if I need more detail than the tilemap grid can provide? Is a flat 2D grid enough? Will I need that 3rd dimension after all? I'm planning on making the puzzles be mostly 'flat', so hopefully it should be enough, but I'm a little worried about it.

[Thaumaturge]

An intriguing project, thus far; posting to follow! ^_^

[Mark Mayers]

Looks interesting.

What do you have in mind in terms of puzzle mechanics?

First Person Puzzle games are challenging to design... Good luck!

[GrumpyGiant]

Looks interesting.

What do you have in mind in terms of puzzle mechanics?

Still very much undecided at the moment. Working on some prototypes right now and hopefully I'll be able to post some sketches next week. I'm thinking it might be nice to have a kind of symmetry between the giant houselegs and the small objects within the puzzle areas, so I'm looking into behaviours that mirror what happens outside the puzzles. (which right now is nothing, but I've got some thoughts on it)

First Person Puzzle games are challenging to design... Good luck!

Darn straight.

So far it's helped to stay committed to the vision, but not get held to any particular details. Directions to aim for, not plans.

[marcgfx]

I like your approach. It's something that works well, when you are the only team member! The forest looks really nice and the houselegs are fantastic 
I must admit, I can't think of a specific puzzle that would profit from a forest setting. I was thinking of fairy tales. maybe goldilocks, red riding hood and rapunzel could be found in those woods... hansel and gretel... I guess there could be fairy tale related puzzles. but yeah, houselegs imply something completely different.

[GrumpyGiant]

One of the earliest visions I had for the game was the image of these giant monsters searching the forest, looking for you with some kind of giant evil spotlight. I liked the idea of you walking through the forest and getting seen by these things. I want to make it so the player feels safe amongst the trees, but exposed out in the open clearings.

I drew this highly-detailed and exquisite concept sketch (spared no expense!) a few months ago. I didn't know how any of it was going to fit together back then, but I was drawn to the visual of the beam poking above the hills. The idea that you get a glimpse of the searchlight before spying the monster itself seemed like a fun element.



So the houselegs needs some kind of beam or searchlight to look around with. There's a couple of criteria we know it needs:

• The beam must be visible from a long way off, so you can see what it's looking at from wherever you stand.
• We need to impart to the player that this beam is in some way 'special' and 'evil' -- i.e. not just regular light, but something that will trigger an action if you get caught in it.
• Must not be janky. (as-if!)

Historically in games, a classic way to draw a spotlight is using just a single flat polygon ("billboard") with a fading spotlight texture on, like so:



Once you use the right blending, it can look pretty good from the side. Lots of games have used things like this in the past. The trouble is when you try and look directly along the beam, you see that's it just a flat polygon. A common solution is to add more billboard images at 90 degrees to it, which kinda works, but it's a bit janky. It works well enough for seeing spotlights from a distance, but isn't much use when you can stand within the spotlight.

Another approach is to use a stencil buffer. With that you can just draw the spotlight cone as a polygon model, and color in the pixels underneath it. We're getting somewhere now, but it's going to be either ON or OFF -- i.e. the parts under the spotlight will be 100% lit up, and the parts outside will be 0%.

So what I'm doing instead is to use volumetric shading. For each pixel, I trace a ray through the spotlight and find where the ray intersects the outer cone. Using the distance between these two points, I can work out how much of the spotlight is visible at each pixel. Once I have that, I can shade it appropriately, using a standard distance-fog algorithm.



This has three advantages:
1) The beam looks solid and non-polygony from all angles (including inside)
2) I get a nice fade effect towards the edge
3) As you walk in/out of the beam it fades away correctly.

But this still leaves the question of exactly how we color it in. Now if it were a real spotlight, we would simply add the light from the spotlight on top of everything else. And that'd work fine, except my game is set in the daytime.

If you've ever tried to shine a flashlight outside on a sunny day, it basically does nothing. The sun is so bright that it overwhelms everything else. In fact, it's kinda stupid for my monsters to be using a searchlight during the day. So I could change the game to nighttime, I guess. But I don't want to do that. I haven't ruled out changing the look of the thing, but right now I have a really nice bright sunny daytime feel and I kinda want to keep it.

But our light doesn't need to be physically accurate; in fact, I want it to contrast and stand out, to make it look more sinister. So instead what I'm doing is to use a multiplicative blend -- i.e. tint the scene red, like in The Terminator. I don't think there's any kind of beam in the real world that could do that, but it gives the look I want so screw it

Here's the final thing in motion:




So there we go, our completed searchlight. I guess the next step is to start giving the houselegs some AI so it can actually use it to look for you.

[Ninety]

Interesting approach you're taking here. I'll be following

[GameBubble06.]

Interesting style. For the searching beacon what you can do to enhance the effect is make (if possible) the light to highlight the scanned objects, trees, rocks, etc. It gives imo a "scanner" effect rather than a "flashlight" effect.

[GrumpyGiant]

Interesting style. For the searching beacon what you can do to enhance the effect is make (if possible) the light to highlight the scanned objects, trees, rocks, etc. It gives imo a "scanner" effect rather than a "flashlight" effect.

Yeah that's an interesting idea, I'll give it some thought certainly.

[GrumpyGiant]

Testing out my tilemap puzzle editor:





I wrote a simple rubbish loader for Tiled's .TMX files and dropped it into the game. It spawns a new entity for each tile in the map. Right now I don't really have any actual models to put in there except for a simple brick wall, but it works!

One of the decisions I made very early on was that everything in the game is hotloadable (code, textures, models, and now tilemaps). I can't recommend doing this enough. It's so gosh darn useful to just be able to edit something in an external editor, hit Save and watch it appear right in game. Once you get used to it there's no going back!

Hopefully I'll be able to start pushing fowards now and test out some of the puzzle ideas. I haven't written much about that yet. I'm not even sure if I should or not? I don't want to accidentally give away all the spoilers in the devlog... but I think it should be fine to explain the basic mechanics of things.

[nathy after dark]

One of the decisions I made very early on was that everything in the game is hotloadable (code, textures, models, and now tilemaps). I can't recommend doing this enough. It's so gosh darn useful to just be able to edit something in an external editor, hit Save and watch it appear right in game. Once you get used to it there's no going back!

How'sabout some technical side info on designing/coding a hotloadable system? Seems really interesting & would be down to hear your strategies/lessons from it along the way.