Support NeoGAF

[Ashodin]

Haha Jobbs I told you. People want your game bad on GAF lol

 

[missile]

I finally have the 3d angular momentum equation down on paper! The equation
now includes variable mass. Holy! And yes, I will recompute the inertia tensor
each timestep, but not only that, I will also compute the time derivative of
the inertia tensor each timestep! How about that? That's pretty much a
criminal offense!! xD

Now I have to code it all up. If everything works out in code as well, I think
my equations here are one of the most, if not the most, accurate 3d rigid body
equations ever coded for video games.

Yielding what? Well, the linear momentum equation yields awesome motion of a
given rigid body under the applied forces. But much more important is the
angular momentum equation, which gives us the rotation of a rigid body under
the applied forces. This equation seems to be not treated carefully and as
such is actually the reason why rotations of 3d rigid bodies do suck in games.
They are often either too fast or too slow and often do not include proper
gyroscopic effects (rotation about an axis can induced a rotation about another
axis depending on the mass distribution with respect to the axes and the
angular velocity). That's at least something that takes me out of the loop
way too often, if objects do spin unnatural somehow.

With my attempt here, I now hope to capture all of these awesome rotations!

 

[BomberMouse]

I finally have the 3d angular momentum equation down on paper! The equation
now includes variable mass. Holy! And yes, I will recompute the inertia tensor
each timestep, but not only that, I will also compute the time derivative of
the inertia tensor each timestep! How about that? That's pretty much a
criminal offense!! xD

Now I have to code it all up. If everything works out in code as well, I think
my equations here are one of the most, if not the most, accurate 3d rigid body
equations ever coded for video games.

Yielding what? Well, the linear momentum equation yields awesome motion of a
given rigid body under the applied forces. But much more important is the
angular momentum equation, which gives us the rotation of a rigid body under
the applied forces. This equation seems to be not treated carefully and as
such is actually the reason why rotations of 3d rigid bodies do suck in games.
They are often either too fast or too slow and often do not include proper
gyroscopic effects (rotation about an axis can induced a rotation about another
axis depending on the mass distribution with respect to the axes and the
angular velocity). That's at least something that takes me out of the loop
way too often, if objects do spin unnatural somehow.

With my attempt here, I now hope to capture all of these awesome rotations!

Why does the shape/mass of the rocket change? Isn't it easier to use a rigid body/joints approach?

 

[Blizzard]

I finally have the 3d angular momentum equation down on paper! The equation
now includes variable mass. Holy! And yes, I will recompute the inertia tensor
each timestep, but not only that, I will also compute the time derivative of
the inertia tensor each timestep! How about that? That's pretty much a
criminal offense!! xD

Now I have to code it all up. If everything works out in code as well, I think
my equations here are one of the most, if not the most, accurate 3d rigid body
equations ever coded for video games.

Yielding what? Well, the linear momentum equation yields awesome motion of a
given rigid body under the applied forces. But much more important is the
angular momentum equation, which gives us the rotation of a rigid body under
the applied forces. This equation seems to be not treated carefully and as
such is actually the reason why rotations of 3d rigid bodies do suck in games.
They are often either too fast or too slow and often do not include proper
gyroscopic effects (rotation about an axis can induced a rotation about another
axis depending on the mass distribution with respect to the axes and the
angular velocity). That's at least something that takes me out of the loop
way too often, if objects do spin unnatural somehow.

With my attempt here, I now hope to capture all of these awesome rotations!

Have you tried Kerbal Space Program? It might be fun or inspiring for you even without certain features you may strive for.

 

[missile]

Why does the shape/mass of the rocket change? ...

First off, the use of a rocket here serves just as an example, it offers me
many advantages while implementing all the equations. For example, a rocket
goes (mostly) upwards, hence just one dimension involved, which is good for
verifying some of the formulas. Secondly, the shape is an easy one. And
thirdly, I don't need to generate kind of a complicated lift force. The only
lift force comes from the thruster and goes straight upwards (for the time
being), i.e. the nozzle is held down. Within my rocket example the lower
section of the rocket serves as the propellant fuel tank. The thruster burns
propellant mass and eject it at a given mass flow rate, i.e. m'(t). This mass
flow rate is negative. Hence we are losing mass, i.e. m(t) = m(0) + m'(t)*t
with m(0) being the initial mass of the rocket. Now since m'(t) is negative,
the rocket's mass decreases over time. And this influences all the equations
due to the fact that the mass isn't a constant any longer.

... Isn't it easier to use a rigid body/joints approach?

That's a different context. When I say the rocket loses mass, it loses an
infinitesimal amount of mass, dm, and not an extended one like it would be if
the, for example, lower stage of the rocket would break lose after burn-out,
which would also change the total mass of rocket, for sure. For the time being
the lower stage of the rocket is strongly coupled with the upper section. One
can see this as a joint of infinite bonding.

Joints (constrains) are on my list as well. Would be cool to have the
rocket brake apart while applying to great a force. Heh?

Collision detection is also on my list. Collision detection is going to be a
hard one, not only geometrically, but also dynamically, i.e computing the
right physical response. I've seen that some people simplify the equations of
motion drastically (or simply do not know that there are some additional terms
worth considering) to compute the resulting linear and angular velocity after
impact. However, that's a different story.

Have you tried Kerbal Space Program? It might be fun or inspiring for you even without certain features you may strive for.

Problem is, I'm already over-inspired just by my own stuff. I would actually
need two lives realizing it all. And as I said, the rocket is just an example.
I don't want to make a rocket simulator or something similar. Despite rockets
are fun, in the end their movement isn't fascinating until a major failure
happens. Indeed, a rocket gone wild is really pleasing to look at.

So whenever you shoot a rocket in a game, make it go a bit wild! There is
another advantage doing so. Like particles, rockets to resolve space to a given
degree enhancing the perception of the space the game takes place. That's good
demonstrated here;

Game: Cactus


I just want to add something to my previous post. The emphasis put into the
equations aren't made just to make rotations look good. The major reason is to
always apply the right force at the right attitude over time. You have to
know that a given attitude determines how some force are applied, leading to
a new attitude after integrating the equations of motion. For example, the
lift force of a wing depends on its current attitude. As I once said, I want
to have a very good mechanics of all the ZONE crafts. And getting the angular
stuff right is a major ingredients leading towards this goal. The rotation
will feel right with respect to the weight and axis of rotation, and you will
experience gyroscopic effects, which can also lead up to some tumbling motion
of the craft.

 

[RobotNinjaHornets]

Found a junk heap. Or maybe it's some kind of pack rat's lair..

I also made another gif that demonstrates combat against a different type of enemy than I've shown before. I think it's probably too big to embed so I'm just linkin it:

http://www.ghostsonggame.com/miscd/tdroid.gif

There aren't a lot of mecha/droids in my game, only those leftovers from various expeditions to this dangerous world through time. Things always seem to end the same with those.

Goddamn this looks good

 

[charsace]

Damn. This gif just sold me on the entire game.

Goddamn this looks good

I played it. Its gonna be good.

 

[V_Arnold]

I do not even know which one of you to congratulate on your own progresses, so here is to all of you

On another note, if any of you need a software scaling function in HTML5 that is pixel precise, but slow to generate the result (kinda comes with the territory...going over width*height per pixel), I can show you how to do it. It is fun, uses custom canvases...and after two days of coding/thinking, we decided to just use double sized PNG's in addition to the normal sized ones. Way less loading times that way.... especially with the amount of sprites we work with.

 

[Jobbs]

Thanks guys.

Tying up some old loose ends this afternoon.Here's something different.

 

[Feep]

DID IT! Any arbitrary number of soldiers and structures, correctly sorted from back to front, and all shadows falling correctly on all surfaces.

This is really the culmination of a LOT of research, asking for help, and weeks of experimenting. Doing what I'm doing is extremely weird, so I'm really happy it worked out.

Early performance tests were fine. I ended up using a generic sort on the Soldiers (just implemented IComparable<> and let .Sort() do its thing), the structures are presorted in the scene, and then a single merge pass without recursion.

WOOOOOOOOOO

Here's what it looks like in the editor:

And here's the illusion magic:

 

[Turfster]

And here's the illusion magic:

Magic indeed!
Looking good.

 

[Ashodin]

And now for today's update, I'll give you some of the world-building I've been working on for the game.

Apexicon: The book containing every word and the most powerful artifact in existence. Just holding the book imbues you with the ability to summon any word onto any surface, or empowering yourself with words beyond your comprehension. It is said this book was the origin of the universe itself, written spontaneously into itself since the dawn of time every event, every concept, every word.

Writing Era: Year categorizing metric. The War of Punctuation happened in 1127 WE.

The Apexicon has since been passed down through early society, cavemen, and the like, bestowing language and words upon man, so that he may express himself in an easier fashion than gestures and drawings. This book was highly coveted until it was lost in the War of Punctuation.

Several different groups, each heralding the various punctuations which enhance the magic already inherent within the words, fought each other over which punctuations were the most beneficial. Each type had its own extra power added to words, and each type was lauded as they assisted in jobs within society.

THE PUNCTUATICUS FIVE were the known groups that arose out of the various supporters of punctuation, and were the ones that instigated the War of Punctuation.

Exclamation Proclamation - A ragtag band of warriors, explosive and violent, always attacking with various words attached with exclamation marks to dangerous and powerful effects.

Comma Chameleons - A solemn group of individuals who insist that commas are the best way to merge all ideas into one, and utilize them to join various spells into one. This group stays hidden more than most to safeguard their precious thoughts.

P.E.R.I.O.D. - The “Professional Enders of Rationally Intelligent Organized Discussion” group heralds their favorite punctuation by using it to end debates, conflicts, and other various battles. Their “end of story, period” catchphrase is feared by many, as these individuals believe they are the peacekeepers of everything, whether you wanted it or not.

Question? - Ever vigilant individuals who challenge the notion of anything, always content with wanting to change the status quo. Ending things with a question leads to more questions than answers, and this group would want it to stay that way.

The Parenthetical (Hierarchy) - This group is the most numerous of the five - with various splinter groups using brackets, parentheses, and other ways to box in words. Utilizing these punctuation marks allows one to do what the other punctuations can’t - transferring words to and from surfaces with a flick of a wrist. This group believes it has a higher power than all the other punctuation groups, and it shows.

 

[Vard]

Thanks guys.

Tying up some old loose ends this afternoon.Here's something different.

Hard to believe this was done in Stencyl! The small effects complement each other really nicely - the clouds passing by, the shaking screen + electricity + flickering helmet, and the floaty particles... gotta appreciate all those little details seen in such a short animation.

 

[Jobbs]

Hard to believe this was done in Stencyl! The small effects complement each other really nicely - the clouds passing by, the shaking screen + electricity + flickering helmet, and the floaty particles... gotta appreciate all those little details seen in such a short animation.

Thanks. Stencyl's a tool -- You can spawn objects, you can make them do things, you can set up logic that tells things how to behave. You can put any art in you want. So there's no reason you can't do all kinds of things. For example, I spawn an object with a pattern graphic, fix it on top of the screen, and set blending modes to it, and that's how I get an overlay effect for my game.

 

[Chris_C]

Here's some drawings I did for the human characters of ghost song, because I'm about to work on the part of the game where they appear. It'll take me a while to do the game art and all the animations for them, but hey, it always does.

These look quite nice! A couple of things:

Be careful with your first character (the one with the onsie tied around her waist). Because of your colors and silhouette, you're veering into "Chell" territory.

Also, I'd recommend making your characters arms a hair longer, as they're a little short at the moment. When extended, human hands end just a little above the knee.

That said, the game looks great, especially liking the last Gif you shared

 

[Jobbs]

These look quite nice! A couple of things:

Be careful with your first character (the one with the onsie tied around her waist). Because of your colors and silhouette, you're veering into "Chell" territory.

Also, I'd recommend making your characters arms a hair longer, as they're a little short at the moment. When extended, human hands end just a little above the knee.

That said, the game looks great, especially liking the last Gif you shared

To be honest, I didn't know who Chell was and had to look it up. Never played Portal. I know this is blasphemy to some of you, I swear I don't mean any harm, I just haven't played it!

My concept of arm length is to extend them a bit past the groin, if that's wrong then maybe old habits just die hard. I used to illustrate comics before I got into indie development.

Thanks for the feedback!

 

[missile]

DID IT! Any arbitrary number of soldiers and structures, correctly sorted from back to front, and all shadows falling correctly on all surfaces.

This is really the culmination of a LOT of research, asking for help, and weeks of experimenting. Doing what I'm doing is extremely weird, so I'm really happy it worked out. ...

Am continuously amazed how much emphasis we put into all of this, into video
games. No valley too deep, no mountain too high! Heh?

At times I have to recall myself that what I'm doing is actually stuff for video
games. If I'm done with the angular stuff, and if I would implement the
rotating earth coordinate frame (will do later on), i.e. accounting for the
additional centrifugal- and Coriolis force acceleration terms, one could
compute the ballistic of a rocket by just applying the right external forces.

Video games are perhaps the only one resort where everything has to come
together at once. It spans so many disciplines.

 

[Feep]

But set up the right reference frames, Missile, and a Coriolis force term appears, plain as day! ;-)

Yeah, almost every creative discipline ever invented goes into video games. It's rad.

 

[Ashodin]

DICTIONARY PROGRAMMING IN PROCESS. WORKING... ACCESSING

A GIF appears! DAMAGE Is working!!! Also seeecret pause menu.

WOOHOO

I'm so thrilled things are coming together.

 

[CreativeSK]

I'm trying to implement an entirely custom rendering pipeline in Unity.

I want to have dynamic global illumination. I'm working on an implementing one of the Dachsbacher papers, for single bounce indirect lighting (high and low frequency).

After yonks and yonks of painful shader debugging I've gotten this far:

It's obviously not quite right - the influence of reflected light from the floor is rather overwhelming compared to the walls - but it's the 'most right' I've got it so far.

My question - does anyone know if anyone is trying to do similar, or already has done? I've tried to find an off-the-shelf solution in vain. It would save a lot of hassle if there was something out there working and tested.

Well what paper exactly are u trying to implement here? Splatting Indirect Illumination, Refelective Shadow Maps or another one? I've implemented a bunch of them for my bachelor thesis leading up to Light Propagation Volumes. I still got some GLSL code and an OpenGL Project aswell. Im sure i can help u out if u have any specific questions. If ur interested just hit me up.

 

[Raonak]

got the combat system down on my DMC-RPG. can now do 3 combos, a stinger, launcher, aeiral split, charge attack, and sword throw. Actually managing to get SSStylish rank with only 1 weapon when I try hard enough (usually requires getting enemies to line up).

Im having trouble deciding a good progression system for my actionRPG. main problem is both my systems are conflicting. see, in RPGs, you want to get strong, so you defeat your enemies faster. but in stylish character action games, you want to keep enemies alive as to increase your stylish level.... Im afraid people will overlevel too much to the point that enemies start dying in a few hits, totally breaking the stylish engine, as they won't be able to get enough hits to actually increase their rank enough.

maybe allow enemey HP scaling, or damage caps. or something...

 

[gofreak]

Well what paper exactly are u trying to implement here? Splatting Indirect Illumination, Refelective Shadow Maps or another one? I've implemented a bunch of them for my bachelor thesis leading up to Light Propagation Volumes. I still got some GLSL code and an OpenGL Project aswell. Im sure i can help u out if u have any specific questions. If ur interested just hit me up.

Hey! Thanks for replying

I'm trying Splatting Indirect Illumination.

The main thing I'm trying to address at the moment is the dominance of the white indirect light, as in the pictures posted earlier.

The light has a field of view of 60 degrees, so in fact, most of the direct light is hitting the white surfaces of the floor and box tops. So in a certain sense I guess it makes sense for white to be dominant. But I think the strength of the white and the fall off isn't correct, which makes me think I'm perhaps missing a scaling somewhere in the code. Like maybe you are supposed to scale the reflected flux using the size of the RSM or number of splats used? Or the field of view of the light?

I also would like to widen the field of view of the light, so direct light will hit more of the side walls etc. But when I do this, my splat distribution is all wrong. Basically because of the perspective projection in the RSM, splats get distributed in a concentrated fashion closer and closer to the light as the fov widens. So I think my splat distribution needs to be modified also in some way, to account for the perspective projection (?)

I actually contacted Carsten Dachsbacher and he's been gracious enough to answer some of my questions. He mentioned also that with a wider field of view, I'd need to account for the solid angle of the RSM texels in calculating the reflected flux. I'm not sure if that's something you accounted for in our implementation?

edit - I'm also not using importance sampling, but I have doubts that the lack of this would be causing the problems I'm seeing...

 

[gofreak]

On that last note I decided to experiment with scaling the indirect light by a function of the number of splats and results look...slightly more plausible?

There is red and green bleed now on the floor.

Quite strong white bleed off the top of the box, and not sure how plausible that is..or the shape of the reflected light... Also not quite sure why there's that blue-ish tinge to the reflected white light. I am compressing the flux in the RSM but not sure if that would explain that.

edit - Actually, I think the light reflected off the top of the box is being exaggerated a bit by the fact that for direct lighting I'm treating the light as a point light when really it should be a spot light...so the direct light around that area at the top near the light is fairly strong...

 

[Blinck]

Just wanted to say that this thread is very awesome and inspiring Keep going guys!

 

[BlueMagic]

Question for the people who worked with XNA before:

How bad of an idea is it to draw all the sprites on the screen at once? I mean calling SprBatch.Begin() and SprBatch.End() only once (for the objects mainly, maps and HUD will be done separately probably). The way I was doing it (since I saw it done like this in many places) was calling Begin() and End() for each sprite that was going to be drawn. However this would not really let me use the depth buffer (and sorting the list is quite slower). Since my game is top down (edit:isometric actually) I could really benefit from this, but I'm sure there are at least some downsides (can't practically use shaders? I'm not entirely sure).

 

[Jocchan]

You are actually supposed to draw multiple sprites between Begin() and End().

 

[BlueMagic]

You are actually supposed to draw multiple sprites between Begin() and End().

And I was doing this to an extent, but not 1 call for every sprite on screen. I guess I was just wondering if there was any heavy downside.

 

[charsace]

Question for the people who worked with XNA before:

How bad of an idea is it to draw all the sprites on the screen at once? I mean calling SprBatch.Begin() and SprBatch.End() only once (for the objects mainly, maps and HUD will be done separately probably). The way I was doing it (since I saw it done like this in many places) was calling Begin() and End() for each sprite that was going to be drawn. However this would not really let me use the depth buffer (and sorting the list is quite slower). Since my game is top down (edit:isometric actually) I could really benefit from this, but I'm sure there are at least some downsides (can't practically use shaders? I'm not entirely sure).

You ideally want to get as many sprites called in one sprite batch. Its more efficient.

 

[Paz]

So incredibly psyched, right now Tim (one of the two other people working on Cactus with me) is at AVCon showing the game publicly for the first time, and next weekend we will be at PAX Aus

Now I just need to figure out how to let some journalists know...

 

[Terra_Ex]

And I was doing this to an extent, but not 1 call for every sprite on screen. I guess I was just wondering if there was any heavy downside.

Lot's of info on it here, as others have said, the fewer begin/end pairs, the better.: http://gamedev.stackexchange.com/qu...ions-of-calling-spritebatch-begin-end-in-many

And you can still use shaders with different parameters/ techniques on each object in your scene, if that was a concern of yours, use something like this before spritebatch.draw(...

theEffect.Techniques[0].Parameters["paramName].SetValue(value);
theEffect.Techniques[0].Passes[0].Apply();

(Don't have VS in front of me right now but it's something along these lines (been a while since I last used it):

 

[Limanima]

Snails is getting featured on the brazilian Windows Phone store!
For what I've read, this picks are not random, they select only games that are worth it. This is good. Let's see how many downloads the game will get.
It's nice to see my game on the same page as Angry Birds Star Wars!

 

[Dynamite Shikoku]

Snails is getting featured on the brazilian Windows Phone store!
For what I've read, this picks are not random, they select only games that are worth it. This is good. Let's see how many downloads the game will get.
It's nice to see my game on the same page as Angry Birds Star Wars!

Congrats. Getting featured must be awesome.

 

[Jobbs]

Screenshot saturday?

Here's something different, one of the rare surface areas of the game. And the world's kinda floating away.

There's a mutant witch in the area. Destroying her makes all these floating things drop, which is kinda neat, but potentially also permanently removes pathways to optional areas.

 

[Bog]

Screenshot saturday?

Here's something different, one of the rare surface areas of the game. And the world's kinda floating away.

There's a mutant witch in the area. Destroying her makes all these floating things drop, which is kinda neat, but potentially also permanently removes pathways to optional areas.

Spoilers, damn.

 

[Blizzard]

Spoilers, damn.

I know, sheesh, some of us are trying to go into this game blind! Who does this poster think they are.

 

[Dark Octave]

Screenshot saturday?

Here's something different, one of the rare surface areas of the game. And the world's kinda floating away.

There's a mutant witch in the area. Destroying her makes all these floating things drop, which is kinda neat, but potentially also permanently removes pathways to optional areas.

Looks amazing. I really can't wait to play this. I would probably shoot for Greenlight or something if I were you.

 

[Raonak]

I've been working on this about 2-3 weeks now.

basic combat completed... time to work on more enemies and actual levels.

 

[Vazra]

RPG Maker VX Ace 66% off on Steam lowering it down to 23 bucks. I wonder if I should jump at that price and learn to do sprite art.

EDIT: Got it.

 

[McSpidey]

So incredibly psyched, right now Tim (one of the two other people working on Cactus with me) is at AVCon showing the game publicly for the first time, and next weekend we will be at PAX Aus

Now I just need to figure out how to let some journalists know...

You'd think journos would be interested in covering the first game confirmed to make humans go Super Saiyan.

Assault Android Cactus at AVCon

 

[lordfroakie]

RPG Maker VX Ace 66% off on Steam lowering it down to 23 bucks. I wonder if I should jump at that price and learn to do sprite art.

EDIT: Got it.

We should learn Ace together also, download the Ace packs and music packs since they're cheap. I downloaded modern tile set, might use them with custom tile sets.

Silo 2 is currently 29.99, construct 2 59.99 and art rage 4 is 24.99

 

[BomberMouse]

Screenshot saturday?

Here's something different, one of the rare surface areas of the game. And the world's kinda floating away.

There's a mutant witch in the area. Destroying her makes all these floating things drop, which is kinda neat, but potentially also permanently removes pathways to optional areas.

Please reconsider that. Don't punish your player for advancing in the game.

 

[Corsick]

For a complete newbie, is RPG maker better than Gamemaker? I'm not really sure if the added freedom/flexibility in Gamemaker would be good for me as a beginner. Then again I'm sure I'll want to branch out from RPGs at some point.

Edit: I did read the op as well, just conflicted is all.

 

[Ran rp]

Wow, I need to follow this thread more often.

 

[Vazra]

We should learn Ace together also, download the Ace packs and music packs since they're cheap. I downloaded modern tile set, might use them with custom tile sets.

Silo 2 is currently 29.99, construct 2 59.99 and art rage 4 is 24.99

Well we should probably PM each other with our findings. I will be watching Youtube tutorials for the maker and also scripting tutorials.

I need more money to get other engines as Im short on money thanks to the steam sale.

 

[charsace]

We should learn Ace together also, download the Ace packs and music packs since they're cheap. I downloaded modern tile set, might use them with custom tile sets.

Silo 2 is currently 29.99, construct 2 59.99 and art rage 4 is 24.99

Why buy Silo 2 when there is Blender and Wings?

 

[Raonak]

For a complete newbie, is RPG maker better than Gamemaker? I'm not really sure if the added freedom/flexibility in Gamemaker would be good for me as a beginner. Then again I'm sure I'll want to branch out from RPGs at some point.

Edit: I did read the op as well, just conflicted is all.

I found RPGmaker to be an excellent starting off point. you baiscally have a playable build
as soon as you create a map and set a party starting point. everything is readymade for you, items, monsters, tilesets, heroes, etc. which you can go erase them and use your own once you're ready for a proper project.

 

[Jobbs]

thanks all.

Please reconsider that. Don't punish your player for advancing in the game.

Something that's constantly with me when I work on this game is the idea that actions should have consequences, and that I'm not necessarily bending over backwards to make the player feel welcome in the world. This isn't meant to be a cuddly game.

In this specific example, the pathways aren't going to lead anywhere you *must* go, it's just a way of adding missable content and hopefully fostering community and what I call "supra-fiction lore" as not everyone will have the same experience or see all the same things. It's definitely walking a fine line, of course. Instead of looking at it as losing something for killing the witch, look at it as gaining something if you find it before you kill her. I understand this may be controversial and I'd be interested in hearing other points of view on this type of thing.

Another example of something I wanted to do but probably won't make it into this first game -- But another idea that interested me was a water level that spans the map and can be lowered, but never raised. To progress in the game it must be lowered, but in doing so you also close off areas of exploration forever -- Areas that could only be swam to. And by changing areas from an underwater area to a dry area, you also change the plant and animal life and may miss things that you'd only have seen if you were underwater.

This introduces a whole bunch of complexity and is something I want to do, but probably can't do properly in this game, especially considering I haven't included underwater gameplay at all yet and am not even sure how it should handle. But it's in the back of my mind to base the next game around it, if there is one.

 

[Jocchan]

Missable content is fine as long as it's used with moderation, it's absolutely non-essential to beat the game, and the requirements are not too obtuse (ie. requiring specific chests to not be opened).

The witch example sounds fine to me, of course assuming the connection between her and the floating rocks is clear enough before the player kills her.

 

[Dynamite Shikoku]

I think making the game's music has been the most enjoyable part of the whole thing. With the other parts of the game, I felt really focused on getting all the little things exactly right, whereas with the music I feel the process is a lot more fluid. I just play and let things take shape. It's very free-flowing.

 

[gofreak]

Some updates to my global illumination implementation:

Changed the direct lighting to a spotlight model, which is more 'correct' than using a point light in this context. Though a 'box' spotlight would be better again than the circular one here...

Still haven't figured out why indirect white light carries this blue tint.