Indie Dev Blog: Folk Tale

In a rush? Jump to the latest video on page 6, or read on to see our progress over the last two years.


After reading and seeing what everyone else is working on, we thought it was time to share some early screenshots of our team's first project together: Folk Tale. You can also watch the latest dev update video on YouTube

If you are a talented 3D, texture or concept artist, and are looking to get involved in an established project in exchange for a share of net profits based on the hours you contribute, please get in touch. We are currently five strong, aiming to develop a single-level playable demo by November 1st 2011, after which we'll be contacting publishers.

#1


#2


#3


#4


#5

 

Looks awesome. How did you make such labels on units, that always looking on camera and displayed over all layers (trres, terrains e.t.c.)?

 

@west131087:

I wrote a class that I attach to the game object, and in that class, place the draw code in OnGUI which uses GUI.DrawTexture in the context of a GUI.Window. Another class attached to the game object (e.g. Character.cs or Building.cs) also has an OnGUI function that checks if the object is visible from the camera, and set's the icon screen XY position based on the game object's world coordinates using Camera.WorldToScreenPoint

To check if a game object is visible to a camera, I iterate all the child renderers. If any has isVisible == true, it's visible.

 

Very nice environment work there, very cool!

 

Awesome game.

Too bad i am only a programmer because I would love to join the team.

 

I am awed. Looks really nice.

 

WTF??? I loved the first part! I've played it in 1998, and it was awesome game!!! I wish you to outdo your predecessors.

 

Looks fabulous

Id love to participate, Check out my 2d/3d work and if you think that Id be a useful asset to the team please get in touch, really does look amazing!

best of luck

-Charlie

 

Two new ambient zones added to the map this week. Here's a first peek of a mostly unpopulated volcanic environment. The next dev video will feature a first pass paint/sculpt of the complete level terrain including the yet to be revealed snow mountains.

We're recruiting talented artists across a range of disciplines (3D, UI, concept, texture). So if you can dedicate 20+ hours per week between now and November (share of profits), please get in touch.

 

We kicked off a busy month with a title change to "Folk Tale" to reflect the divergence away from a sequel, develop the project as a new line of intellectual property, and to better capture a feeling of fantasy, adventure and storytelling. A new concept logo aims to portray the simple lives of some of the games main characters: the villagers.



The focus remained on production of prerequisite assets to support the implementation of gameplay, namely the essential buildings. The warehouse (far left), smelt (to be reworked from scratch), and barracks all make their first appearance. The lumberjack occupation, with it's own model can now be assigned by selecting a villager and sending him to the logging camp (see video). The miner occupation model is nearly ready, and will be making an appearance in next month's dev video. Other additions in the village include the chicken coup, and the cartoon cow with associated milking animations.



The style of the original barracks was a little too evil looking, and so it was relocated to the dwarf-themed lava area, which now includes four runic buttons that will act as the precursor to triggering an iron golem, which the player must defeat to secure an important quest item.



The goblin swamp has some new defenses, and patrolling slavers and archers. The AI Pathfinding hasn't had nearly enough attention, so characters are still prone to a little invisible wall climbing and spinning, but nothing that can't be sorted out. The multi-threading added in A*Path 3.0 certainly helps, and a little thanks to Aron for supporting us in ironing out the initial release bugs.



Using the limited system profile data available in Unity (including the out-dated GPU fill-rate stat), the game now takes a reasonable guess at graphics settings, as well as persisting user settings. These eye-candy screenshots have nearly all the image effects enabled, and my poor iMac with it's mediocre GPU just about copes for capturing video on full settings. Those real-time Water4 reflections are FPS killers due to the additional cameras required to draw the render textures, halfing the frame rate. Still, better that players with decent GPUs get to enjoy their investments than not at all.



And finally, here's the latest dev video on the Games Foundry YouTube Channel:

God I hope UT implement hardware custom cursor support in 3.5, otherwise our project will be missing an essential RTS element. If you haven't already, please vote for it. Can't wait for the new GC and GUI implementations, because the current ones suck ass.

 

i am in awe

 

Looks amazing, love the atmosphere!

 

OMG, this is so beautiful!
Great job!

 

The new name works well and the screenshots look very nice. Good job guys

 

Fabulous looking game. Love it. Bookmarked.

 

Just...amazing!
Game like this give me motivation to continue game developin'.

I'm proud to be a Unity user!

 

Truely awesome, hope you get published.

 

Looks really cool Loved Beasts n' Bumpkins

 

Makes！

 

LOOKS ABSOLUTELY AMAZING !

i really like how you made the world seam so vast when looking at the scenery . Cant wait to see more

 

This is very awesome, so its like a 3d RTS, reminds me of Civ

 

Really, your project is getting good shape. I wish you the best.

 

2 months since our last update. Boy where does the time go?!

Here's Folk Tale Dev Update 2 on YouTube.

The dev video begins with a placeholder start screen. As this will be a single-level playable demo to pitch to publishers and investors, the only option will be to start the game which triggers the intro cutscene.

Moored at the new docks area is our migrant galleon, the key mechanism for population control. When your village has spare housing capacity, new villagers periodically arrive from far off lands. On the flip side, if you are a tyrannical leader, unhappy villagers will pack up and leave on the next boat off the island.



Swinging past the beach and flying by the new windmill building that we'll mention later, there's a birds-eye view of your village, where the first 20 minutes of the game focuses on resource gathering, construction and economy. Breaking a wide angle shot down, we have a single camera doing all the work, rendering the skybox, low-detail mountain geometry which responds to fog and lighting, a margin of ocean around a central island , and the main level map. The far clipping plane is at a whopping 4500 to render the far mountains. Unity's frustum culling removes objects off-camera, with our own distance culling system removing far objects that would otherwise be on camera due to the huge far clip plane. We've played a little with the occlusion culling, but given it's buggy state, we're leaving this until Unity 3.5 is released. A Preferences Manager handles the selection of default graphics settings, managing post render effects, terrain detail and culling distance based on the capabilities of the GPU combined with the screen resolution to ensure a decent range of machine specifications can be supported.

Back at the docks, and the migrant galleon rings it's bell to depart with an unhappy villager up on deck. Boats use a Taxi Manager that allow villagers to buy tickets and wait for the boat to arrive. It also runs the Goblin Ferry in the Stinking Swamp. Boats currently use UnitySteer but for some reason the galleon is really jerky. Something to investigate at a later stage.

As night time descends, several pesky goblins find their way into your village. Nothing that a few Knights trained at the barracks and a wizard can't cope with. As the battle begins, the sound track changes to reflect the heightened threat level. You might recognise the Knight and Wizard models as purchased stand-ins, and not the work of our team. Our Knight replacement is still being animated and just missed the cut-off for this build.

Quite a bit of work has been done to improve battles, but there's still a long way to go. Melee combat is a challenge to implement well, especially when one unit is outnumbered and needs to be surrounded by opponents. We're getting there, but more refinement is needed.

One more birds-eye view, this time at night time. The day-night cycle uses the same system as our ambient zones which controls pretty much everything: fog, ambient light, sunlight, water, and soundtrack. One problem we have is during the transition from day to night. If you don't rotate the camera, the terrain billboards fail to update their lighting colour, which means you get daylight lit billboard trees at night time. Other people have this problem, but as yet we've not found a fix. We hope UT fix this in 3.5, otherwise we'll have to ditch the Unity tree system.



Our first new character introduction is the Lumberjack. Villagers perform a number of menial jobs around your village, or can be assigned to specialist occupations by sending them to their respective occupational buildings for training. Once trained, they diligently go about their task. Here you can see our old crappy looking trees, which we're slowly improving and replacing as we go. One thing we need to do is ensure the positioning of the lumberjack when he fells a tree, so that he's not obscured by the tree foliage. To enable interactive trees, we built a Terrain Manager that hooks in to the Unity Terrain Trees, and replaces each one on demand with a mesh tree that can be chopped down. That way the majority of trees benefit from distance billboarding.

Farming has featured before, so we'll not dwell on this other than to mention that the animations need revising, and animation events adding to trigger the sfx on cue instead of the current looped sample that don't sync. The farmer plants, grows and harvests wheat before taking it to the windmill to be ground in to flour.

Our second character introduction is the female villager, seen in the video milking a cow, which also makes its first appearance. Milk will be taken in churns to the yet to be modelled farm, where it will then either be sold, or we might expand the economy to include cheese making.

Finally we have the Miner. Currently he hacks away at a pile of iron ore, but ideally we want him to go into the iron mine. Once he's collected the ore, he'll return to his yet to be modelled occupational home of the Smelt, where it will be turned in to iron bars for use in construction.

Entering the lava pools of the Old Forge, we have the makings of our first simple puzzle. I don't want to talk too much about this, saving it instead for a future update when we've got the voice over recorded and can show the whole piece in context. We're using Boxman's arc lightning effect from the Muse-games Fabricator contest. We hope you enjoy the eye candy, and the moody soundtrack.



Wading into the Stinking Swamp, the goblins have built a new palisade fortification, with watchtowers patrolled by archers. The walkways have been widened to work better with the resolution of the AI Pathfinding grid. As the camera pans around, we catch our first glimpse of one of Slavemaster Urzal's ogre bodyguards, wielding a hefty looking spiked club.



Our final location of the day is the Ogre Mines which we only introduced earlier this week. Our little band of adventurers has gained some experience by now, and is ready to tackle one of these brutes. The Knights take up positions around the ogre, while the wizard remains at range to lob off a few fireballs. Hovering over an enemy will show a red health bar, and the monster's level. We didn't want to clutter the screen with UI.



Most of the segments in this dev video were produced using our CutScene Manager, giving us control over iTween camera paths, image effects, actors, sound effects, voice over, and world events.

Well, that's it for another month or two. Next update we should be able to unveil the snow mountains, and possibly the coolest characters yet.

 

Really nice stuff. One thing I would try to improve would be the rock tiles on the terrain. Make them have chunkier rocks would help.

I dig the style and the day-night and docks color mood.I also like that Ogre.

Keep it up.

koyima

 

Really nice video! Love the mood of some of the images and settings. Especially the color and light scheme of the swamps!

 

when i saw ther first screenshot i was like ow wow and amazing stuff man just amazing

 

As a prelude to early stage marketing and community building, we've established ourselves on the usual suspects. Can we ask the community's support by helping to get us started with a fan base?

IndieDB: www.indiedb.com/games/folk-tale
Twitter: twitter.com/gamesfoundry
Facebook: www.facebook.com/folktalegame
YouTube: www.youtube.com/user/gamesfoundry

If anyone has any suggestions of other places we should setup, please let us know.

Many thanks!

 

Hi! i love the video, and i love your work.
thanks

 

Awesome!

 

Lovely looking game, very similar to my idea I'm working on . My question although is did you use the terrain engine to place all your trees, because I can't use the terrain engine with my trees as they have scripts - resulting in either hand placement (not fun) or a third party tool!


Again, great game!

 

We started off using Terrain Trees. But the shader bug that means tree billboards don't respond to changes in ambient light (i.e. day/night cycle) unless you change the Y-axis camera rotation, meant that we had to rip them out. Originally we hooked into the terrain api to cache a TreeInstance[] array for searching for the nearest tree to a character, and then at run-time would replace them with our mesh+script trees on-demand. However, since ripping them out, we now use mesh trees, with our own distance culling system. If the shader bug isn't fixed in 3.5, we'll look at implementing our own tree billboarding system probably using RenderTextures.

 

Sounds complicated, maybe you could eventually sell your tech on the asset store . Other than that, fantastic looking game - I'd buy it

 

This looks awesome, good luck!

 

@Alec - unlikely. That would require a commitment to support which is too much hassle.

However, one thing we are considering is once the demo is ready for community testing, sharing knowledge including documentation (GDD, legal contracts, voice over scripts for briefing actors, tech overview) and a dev video that talks about the tech such as real-time terrain deformation, melee combat (inc. surrounding of out-numbered enemies) and group movement, and how we worked around issues in Unity and tweaked performance. The forums and answers sections have been invaluable in helping us when we first started using Unity, and we really want to give something back. One of the biggest hurdles people seem to be encountering is how to get a collaboration project up and running, and to continue to execute through to completion, so hopefully the documentation and advice will be useful to some.

 

We've been keeping a keen eye on the Developer Preview thread over the holidays while out of the office stuffing our faces with turkey, hoping to gain some insight into the likely impact 3.5 will have on Folk Tale; the features we will use, and those we'll ignore. This feedback does not cover features we don't use.

AI Pathfinding
Currently we use Aron's excellent A*Pathfinding 3.x solution in grid mode. For a long time we've been looking to evaluate a navmesh system hoping that it would be more memory efficient, but with the requirement to re-bake a small area of the navmesh at runtime (a requirement for when the player places a new building on the terrain, as characters can walk on buildings, and can't just avoid them). Alas, the 3.5 solution seems to be fairly basic, so we'll be sticking with Aron's solution in grid mode. Another AI feature that looked promising was the crowd simulation. We've already implemented quite a lot of our own crowd management and avoidance system, and unfortunately the 3.5 solution doesn't seem to offer enough to warrant replacing our own code.

Particles
We deliberately lowered particle-based project tasks to low priority in anticipation of the new system due in 3.5. From all accounts Shuriken looks to be a significant improvement, so we're excited about cracking that open, and will probably start work in parallel to our 3.4.2 dev using beta.

Multithreaded Rendering
This is the one feature we expect to have a major impact on performance. Having already performed a lot of optimizations, our fingers are crossed hoping that by splitting off the rendering onto a separate core, we should see smoother frame rates (we're currently vsyncing to around 30fps).

LOD
Hmm, what to say. We had hoped for auto-decimation of skinned meshes similar to Simplygon (which is too expensive). That obviously got descoped, so what we're left with is somewhat useless for us. We already have our own distance based culling system on top of Umbra (camera distance is 6000+ to render the far mountains) that is calculating square magnitude, which we also use to hide detail other than at close distance (e.g. window frames, crates, sacks etc.). We don't want to be duplicating that calculation, so we'll probably add the little code required to add our own LOD support.

GUI
Disappointment once again. So the plan will be to assess and implement nGUI.

Small Wins, Small Moans
Async progress - whoop. We can finally have a nice loading bar.
No Hardware cursor support - boo. Means RTS games will always be missing one of the most important UI features.
Fixed OC - good.
Same old slow Mono GC - massive boo.
Text-based Prefab System - replaces Merlin's Prefab Lab that has served us well.
Multi-Object Edit - good.

Overall, a mixed bag. After the initial disappointment had passed and it became obvious how we might proceed, we're what I can only describe as mildly excited by 3.5. We'll wait until the production version is released before migrating. In the meantime there's plenty to be working on implementing LOD, nGUI and finishing crowd simulation. The first-pass voice over recording should be completed in January, and work is progressing on the assets required for some of the cutscenes, so we will have some cool things to share in the next WIP video.

Our recommendation to UT continues to be shift focus onto fixing the excellent stuff that's already in, instead of releasing half-baked implementations of new functionality, especially where superior third-party solutions already exist.

 

This game looks so good i like this game verry much it could defently compete with the current RTSs out there it is beutifull

 

Amazing stuff. I especially love this screenshot
View attachment 23759


Nicely done stuff. The graphics in your game look exceptional. I would imagine it would not be easy to locate another 3d artist of the same caliber

 

Fantastic! Good lord what I'd do with 1/10 your graphic skills!?!

Count me in when this one launches!

Cheers,

Galen

 

Thanks for the nice comments guys. I'll be sure to pass them on to the artists at this weekend's team meeting. After a break in December to recharge the batteries, we're starting the New Year with an expanded team for the final push towards a playable demo. Hopefully we can share some of the cut-scenes in the next WIP video, featuring some of the newly recorded voice acting.

 

Quite a bit in this WIP update as it's been three months since our last.

Under the hood, we've migrated to 3.5 beta, with the impact pretty much as expected (see earlier post). On top of the 3.5 performance boost provided by rendering on a second core, we've started cleaning up assets to use a few large texture atlases and a couple of shaders to reduce the draw calls.

Once UT announced the new GUI wasn't going to be shipped in 3.5, we grabbed an NGUI licence, and it's working out nicely now we've got our heads around the coordinates system. Extremely good support from Aren, the author. We ran some tests on the old OnGUI - oh my god. It's evil, stay clear and buy a third-party GUI solution unless you want frequent GC spikes.

We've added our own LOD extension to the distance-based culling system for trees (and other stuff), and thrown out Unity's tree billboard system because the shader doesn't respond to ambient light changes. Ok, so it's currently not as elegant, but you notice it a lot less than sunlit trees on a moonlit landscape. Only the green fir trees currently have LOD.

The video begins with the start screen, during which the object pool manager handles all the object instantiation to front load most allocation overhead and reduce garbage collection calls during gameplay.

Hitting start drops us into the voice-over intro using our cutscene manager. Basically it pushes a sequence of objects onto a queue in chronological order with each object having a cue-time. A Coroutine while loop pops all objects with the current timestamp, and then sleeps until the next cue is due, or until the cutscene finishes. All event objects inherit from a base class, with switch blocks managing each event type enabling us to do anything that is possible through the API.



We implemented quality settings several months ago, and the poor iMac used to render the videos is hard pushed with real-time shadows and reflections running so apologies if it's not uber smooth. We wanted to show off the visuals more than the frame rate. The game runs great on a gaming PC, but also runs on a Macbook Pro on low settings.

The intro ends and drops into the voice-acted tutorial system driven by the quest manager. All pretty simple, just needed to handle all those special cases where a player can skip things, or might trigger another in-game event that conflicts with the active GUI.

Night-time descends and drops us into the second fully animated cut scene. Pete our writer insisted we have London gangster goblins, and since that point the script moved further and further away from stereotypical fantasy. So much so in fact, we have a Mexican Wrestler based Goblin Witchdoctor with a height complex for the end of level boss, and a camp wizard who's Torrets-suffering familiar has been stolen. Goodbye fantasy, hello comedy.



Once the goblins have done their little acting debut, there's some placeholder combat to highlight the new character voice responses for selection and command. At least until the next bit of goblin cutscene is finished where they raid your village and run off with our reluctant hero's wife. When characters die, they now have their rotation adjusted to match the normal of the terrain under their feet to stop their corpse intersecting the terrain too much.



The sound effects have received a much needed overhaul thanks to our new sound designer Ro who's done a fantastic job. Still a few to add in, but most are new.



After a detour through the volcanic region, there's a first glimpse of the goblin occupied derelict barbican blocking the entrance to their swamp. The player will be needing the iron golem seen earlier to pound through those gates being clearing a path through the goblin village to free the slaves and confront the boss and his two ogre bodyguards.



And finally we have a brief glimpse of the snowy mountain region. Mostly early stage assets at the moment including the geometry for the Were-fu practicing monks.

 

The new video looks great, and the overall polish makes the game very appealing indeed - it makes me want to sit down and play it which is a good sign! I love the gangster Goblins, and the overall mixing of classic fantasy with different themes. Sounds like a lot of fun!

I was a little disappointed that the new GUI didn't make 3.5 either, and since using the Unity GUI in our last project (shudder!) I've been considering the various 3rd party GUI alternatives. Is everything in the latest video done with NGUI? Do you recommend it?

 

Looks great, good luck with your project

 

@NomadKing I'd say about 50% of the GUI has been converted to NGUI so far. What hasn't: the minimap, character portrait and damage announcements. The minimap (disabled for the video) is going to take a little more time. We only had 4 days prior to the latest WIP working with NGUI.

The resource panel, construction button, construction dial, tutorial windows, character healthbars/icons and enemy healthbars are being drawn with NGUI.

I can recommend it purely based on the level of support currently provided by the author Aren. It's also cleanly coded and well thought through which makes extending it easy. There were a few gotchas, like importing your atlas texture as ARGB32 if you want true pixel perfect GUI. DXT5 compression caused a number of issues on the project initially. I believe the new inbuilt texture packer fixes this. As a GUI framework it may still be fairly immature, but Aren's commitment to the product suggests it won't be long until it matures.

We can't comment on comparisons with EZGUI as we haven't tried it. When deciding which framework to go with we read that EZGUI was poorly coded and that alone dissuaded us. Your choice of GUI solution will ultimately be decided by your short term requirements.

 

great work! just keep it up.

and as far as tiling of the rock texture is concerned which was mentioned earlier i would suggest to have a look at:
http://forum.unity3d.com/threads/116509-Improved-Terrain-Texture-Tiling?highlight=terrain+tiling

lars

 

@gamesfoundry Thanks for that. While our next project isn't very demanding from a GUI perspective (although after working with the inbuilt GUI anything seems demanding!) I was looking for a solution that would cover a large range of GUI options to save re-learning a new system for future projects. I've been taking a look at the NGUI examples, tutorials and documentation, and it does seem very clean from a design point of view. If the level of support is as good as you say, then I'm sure we'll see some good additions to it as well. Seems like an easy choice now

 

It looks better and better. Great job

 

While gearing up to shoot a dev video showing folks around Folk Tale in the Unity Editor, we thought we'd share a sneak peak at some of the progress made since our last update.

Currently in play-testing is a modified camera system that lets the player seamlessly flip from RTS overview mode to pseudo FPS, letting you get up close and personal with your villagers (oo er missus).



Being able to get this close means we're going to have to have to rethink a few things. The existing art assets have been developed with the camera much further away, so it's time to think about populating the world with more close-up detail, and fixing a few UVs. Adding detail will impact memory, and that may lead us into another battle with the garbage collector.



The new Devil's Causeway section makes extensive use of one of the new sets of texture-atlas optimized static-batching environment props. We're waiting on the polygon-reduction plugin by bibbinator and we can set about adding in LOD models into our existing distance-based-culling-and-LOD system. That should provide another performance boost.



The art team working on Folk Tale continue to do a grand job of creating beautiful characters. Here's a close up of the Golem, posing for a holiday snap on top of the abandoned forge battlements. Rich - one of the Technical Artists - is doing a great job achieving a hand-painted shiny metal look without the benefit of specular maps ( we chose not to include them in our character shader ).



And finally the entire Stinking Swamp zone has being redone to play nicer with AI Pathfinding, and to bring consistency to the textures. Previously some of the walkways were too narrow for the granularity of the AI grid. That's all fixed now.

 

More great screenshots - I'm quite looking forward to playing this Any clues about a release date yet?

 

@NomadKing Unfortunately not. In the single-level playable demo we plan to develop most of the code that will be used to produce the final game to mitigate any technical risks (important as we have third party risk vis-à-vis Unity over whom we have no control). Having the functionality complete means delivery can be somewhat-guaranteed, which we'd want to be sure of before signing up to any contract that gives us exposure to financial risk in the event of non-delivery (e.g. recouping of advanced royalties by publisher which we'd already have spent on staff salaries).

By taking this approach we hope the demo will be polished, offering the best in game play and visuals, running on the widest range of capable hardware, all of which should be of benefit in pitching for funding or investment. Not forgetting of course the fun factor and pride in re-engineering something to run better. Looking at your games on Mac App Store, I'm sure you are very familiar with this feeling

 

Technical walk-through presentation worth doing? Please read this thread.