-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Indie Game Development Discussion Thread | Of Being Professionally Poor
- Thread starter chubigans
- Start date
- Status
I feel like we are about three months away from showing the first build to a selected few - basically, what I am going to be looking for is feedback from folks that are interested in SRPG games that are closer to X-Com than Disgaea, but still retain the single screen combat instead of having huge maps. I cant wait to start sharing details, but my mind is a maze right now, and I still need to rewrite the whole engine for sanity's sake.... but I'll post anything that is worthwile once the time comes 
Oh, and one more sneek peak:
Oh, and one more sneek peak:
Permadeath and no stat/item creep, hey!
Thanks a lot!
Thanks so much!
We might do a free version again in the future, it does work to get more eyeballs but like someone else said, you'd still need tons of players to make any dough. Overall our feeling right now is to simply make the game great and be open about it, and "people will come". We've shipped enough iOS titles to know that even a well marketed one is probably going to flop. There's too much money from the big boys keeping them at the top, and only room for a few random hits that are actually viral.
I'll see if I can post some screens and videos later on.
RobotNinjaHornets
Banned
Fuck me making sprites takes time. I've managed to do the detail on one running frame, and it took an hour or two. Good god.
And I don't even like how I've done the shading D:
And I don't even like how I've done the shading D:
McNerdBurger
Neo Member
Thought I'd share a screen of a prototype I'm cooking up. It's a future stock trading game called BoyTrader. You use genetically modified boys that have inherent powers to manipulate the market. You then work the market with their abilities to increase your fortune. The abilities drain their lifeforce, so you must always be pushing them closer to death in order to maximize profits. Think Dark Crystal mashed with Ender's Game.
I'm only about 4 days into it and just trying to figure out how the core trading mechanics will work, but I've got the basics up and running. The bottom middle screen will eventually be replaced with the boy hunched over and plugged into a terminal.
I'm only about 4 days into it and just trying to figure out how the core trading mechanics will work, but I've got the basics up and running. The bottom middle screen will eventually be replaced with the boy hunched over and plugged into a terminal.
Thought I'd share a screen of a prototype I'm cooking up. It's a future stock trading game called BoyTrader. You use genetically modified boys that have inherent powers to manipulate the market. You then work the market with their abilities to increase your fortune. The abilities drain their lifeforce, so you must always be pushing them closer to death in order to maximize profits. Think Dark Crystal mashed with Ender's Game.
I'm only about 4 days into it and just trying to figure out how the core trading mechanics will work, but I've got the basics up and running. The bottom middle screen will eventually be replaced with the boy hunched over and plugged into a terminal.
Well you got my attention! What are you using to develop it?
McNerdBurger
Neo Member
It is indeed a dark concept but it'll be comedic in nature. I see the tone as being Edmund Mcmillen inspired, but likely a little more dry. The name BoyTrader might not stick, but... you know... the boys trade stock and you trade the boys...(you'll upgrade to more powerful boys as you progress) so it works for me. Maybe it's a parody of WAR FIGHTER in my mind because I'm still just delighted that someone named a game that.
Developing in Unity. I'm a sound designer by trade but I've been learning Unity/C# over the last few months in an attempt to round out my skills and shift into making smaller games. Loving it so far. Using NGUI for all the UI elements and loving that as well.
I should also re-iterate that this is 4 days old so it'll be a while before I can tell if the mechanics are even fun. I have a feeling I'm not the only one here who gets a bit ahead of themselves
nukembashi
Neo Member
Well, I define a small-time programmer as someone that programs on their spare time, and not trying to get into the industry as a salaried employee for an established firm like Yahoo!, Studio Pixel, or Mojang. Essentially, a minimum wage worker that programs when their shift ends each day pretty much fits the bill.
As for a small job, I'm planning on just having them do some basic coding work that shouldn't take much more than a few days at a time. I'm not particularly proficient at coding myself, but it's a job that would take me a couple days to do, so I'm using that as my primary frame of reference.
Sorry for being so general. I don't want to make my posts look like a Craigslist ad when I don't even know if most programmers would even be interested in such a job.
@McNerdBurger:
I love your interface. Did you use any special plugins for the interface/graphs or use Pro to rendertexture, or am I really that hopeless at doing anything shinycool
Edit: Derp. Blinded by the shinies and read over the NGUI bit. Did you purchase the full version, or use the free edition?
Anyway, I love it so far.
I love your interface. Did you use any special plugins for the interface/graphs or use Pro to rendertexture, or am I really that hopeless at doing anything shinycool
Edit: Derp. Blinded by the shinies and read over the NGUI bit. Did you purchase the full version, or use the free edition?
Anyway, I love it so far.
Yay, I'm back! With another Unity question.
(Also, welcome to fellow GAFfer DM_Uselink, who has joined us in the Iridium offices! He's being awesome making awesome art.)
If I want a manager class to maintain a list of *not instantiated at runtime* prefabs/Objects, what's the best way to do it? I thought making a list of GameObjects and then trying to populate it using the GameObject.GetObjectsOfTypeAll() or something would work, but apparently trying to cast the returned type Object to a GameObject is illegal, or something.
Best practices? I know how to do this with instantiated stuff, but this seems weird.
(Also, welcome to fellow GAFfer DM_Uselink, who has joined us in the Iridium offices! He's being awesome making awesome art.)
If I want a manager class to maintain a list of *not instantiated at runtime* prefabs/Objects, what's the best way to do it? I thought making a list of GameObjects and then trying to populate it using the GameObject.GetObjectsOfTypeAll() or something would work, but apparently trying to cast the returned type Object to a GameObject is illegal, or something.
Best practices? I know how to do this with instantiated stuff, but this seems weird.
McNerdBurger
Neo Member
@McNerdBurger:
I love your interface. Did you use any special plugins for the interface/graphs or use Pro to rendertexture, or am I really that hopeless at doing anything shinycool
Edit: Derp. Blinded by the shinies and read over the NGUI bit. Did you purchase the full version, or use the free edition?
Anyway, I love it so far.
Thanks man! I'm no visual artist but I try to make the prototypes look nice so that I can imagine the atmosphere. I use the full version of NGUI and a Unity Pro licence that I'm borrowing from a fellow dev friend (check out his next project here, it's a great concept). I'm on the fence about Unity Pro. Not sure I can do this game without bloom though
.The graphs are drawn in a terribly inefficient way right now, but they work for the time being. I'll probably need to change them to use the line renderer or something later on. I actually hadn't thought of using rendertexture for the screens... that might be a better way to do it in the long run. I'd like them to be very dynamic. Moving around on robotic arms as you switch out stocks and so forth.
Do you mean something like object pools? If so, I have some source code for a generic GameObject pool I've cleaned up a while ago. It only uses an interface to make sure isntances are activated and deactivated properly as they're taken out of and back into the pool, respectively. I might have misunderstood you, though. The objects are instanced as soon as the level loads to avoid the overhead of instancing them on the go, so that might not fit what you wanted to do.
Nah...I want them permanently in the scene view, like lights and whatnot, so I can move them around easily. They're permanent fixtures. I still need to iterate through them in code, though, for a few reasons.
You probably already have, but have you tried casting the results to a Transform, and then doing transform.gameObject?
Hmm... so you mean you want them to activate and deactivate as they get close to the camera? Or is it that you want to place them all in scene view and have them stay inactive until some other event? Or something else entirely?
Teknopathetic
Member
Off the top of my head you might be able to have them all as child objects to an empty gameobject and then use their transform to find the children. See:
http://docs.unity3d.com/Documentation/ScriptReference/Transform.html
http://answers.unity3d.com/questions/10417/how-can-i-access-the-children-of-a-transform.html
Edit: Or what Turfster said.
http://docs.unity3d.com/Documentation/ScriptReference/Transform.html
http://answers.unity3d.com/questions/10417/how-can-i-access-the-children-of-a-transform.html
Edit: Or what Turfster said.
I thought you were using Rendertextures.
That's what I would do anyway, but then those used to be my answers for everything. I really, really miss being able to use them since I moved to unity free...Eon Altar looks great.
McNerdBurger
Neo Member
Could you give them a common tag and use FindGameObjectsWithTag?
It returns an array of the matching objects.
McNerdBurger
Neo Member
Hmm I didn't know that... that's goofy.
Doing another quick stream today. Going to be showing my dynamic sky lighting model that builds spherical harmonics based on the sun's direction:
You can watch it on Clash Tournaments:
(Click Here)
Feel Free to ask questions.
You can watch it on Clash Tournaments:
(Click Here)
Feel Free to ask questions.
Yep. I've got code I reused for an experiment today, that basically generates an entire world full of meshes with tags "walkable" or "obstacle" based on a few parameters, and Unity barfed all over my computer when I tried to run it, generating nothing at all because the tags didn't exist in this new project and thus the meshes couldn't be added.
I had of course forgotten the tag naming was even in there, since I didn't really need it for this project. Cue me having to add the tags manually. I'd be happy with something that lets you add stuff to the tags collector in code, it doesn't have to add new tags if it encounters one in code (because that would of course lead to random typos and differences in tags), but having it just bomb and do nothing at all because there's a tag there it doesn't recognise is rather silly.
I had of course forgotten the tag naming was even in there, since I didn't really need it for this project. Cue me having to add the tags manually. I'd be happy with something that lets you add stuff to the tags collector in code, it doesn't have to add new tags if it encounters one in code (because that would of course lead to random typos and differences in tags), but having it just bomb and do nothing at all because there's a tag there it doesn't recognise is rather silly.
vladimirdlc
Neo Member
We posted in some forums yes. We have 250 downloads in a week and 5 sold. 3 of this downloads are "real" sells. The other 2 are from me and my friend.
Today I just found out that the game is crashing in Windows Phone 8 making it unplayable.
I'll have to correct this ASAP. I don't know exactly how, because I don't even own a Windows Phone 8.
That's not soo bad. If you believe that the game have good selling material try to pitch it to reviewers and people directly or in gamedev events.
Have you tried at least the game with an emulator?
http://msdn.microsoft.com/en-us/library/windowsphone/develop/ff402563(v=vs.105).aspx
I have windows 8 but sadly just in a laptop :/, if you're not lucky with the emulator I will suggest you to open a topic in the windows 8 developers forum, maybe you could even find someone to help you make the tests in a phone.
http://social.msdn.microsoft.com/Forums/en-US/category/windowsapps
Some ideas and techniques behind ZONE
I made some good progress to the equations that will drive the AG crafts of
ZONE. Everything is still done on paper. After getting all the forces and
masses (linear, angular, inertia) right, I'm now actually at the point to
integrate the derived equations of motion using a numerical scheme. I'm opting
for Runge-Kutta of order 4 (rk4) instead of simple Euler or the midpoint
rule, since the increase in stability is very much needed because the crafts
will undergo some hard/large forces during race making the equations a lot
stiffer and as such much harder to solve, much harder to track by a simple
integration scheme like Euler, which would simply blow-up making the craft
instable and as such uncontrollable. But the problem is, it comes at a price.
It requires me to compute the force and torque term four times per time step
(, per craft). Woahh! That's quite a heavy load. Usually the force terms are
way simple and multiple evaluation isn't so serve, but I want to go a lil
further with the AG crafts of ZONE than what seems to be the standard. For
example, I want to account for a cool propulsion system burning kilograms of
fuel exhausted from some nozzles propelling the craft. Burning fuel means a
change in weight of the craft, which makes the craft's mass non-constant, i.e.
m(t), and as such the center of mass becomes non-constant as well. So one has
to recompute the entire mass distribution, volume, moments etc. again, i.e.
one has to recompute the expensive inertia tensor and its inverse again and
again. Usually the weight and the center of mass, and as such the inertia
tensor are constant. This lessens the computational requirements in evaluating
the force term, which is still expensive if you have to do it four times. But
given the non-constantniss of mass amplifies the problem in a big way due to
the recalculation if the inertia tensor.
That's where I'm standing.
Now one could say; hey, dude, skip the loss in weight due to the burning fuel!
Sure, that's the easy way out. But let me explain way I think this is actually
all cool stuff. Btw; You can make a comment about it if you think otherwise or
if you share the idea in some way.
Well, when you carry just a few kilograms of fuel with respect to the total
mass of the craft, you won't recognize a huge change in angular momentum if
the fuel tank isn't a kilometer away from the center of mass. However, guess
for example half of the craft's weight is fuel! A slight displacement off form
the center of mass (hence changing it) has a large effect on the angular
momentum of the craft. Now if the fuel burns then its weight will be reduced
and as such the weight of the craft, leading to a change in its center of mass
(usually) and a change in inertia of the craft. Hence, during race, the
dynamics, the behavior of the craft will change and you as the player have to
account for it, making it way more challenging to race. So for example, at the
start of a race, your AG craft is heavy loaded and the dynamics/handling of
the craft reflects this load. During race you will lose fuel making the craft
lighter and as such faster while the dynamics will change as well due to the
change in center of mass and inertia over time. So you have to know your
craft! It becomes a strategic factor! You may save some fuel (and as such
carrying the weight with you all the time) till the end of the race to ignite
your afterburner for victory! Nothing comes closer winning the race last turn,
I tell you! So it may pay off being able to handle the craft well under a
heavy load to save fuel for some precise afterburner bang-bang maneuvers. Oh,
did I told you may get out of fuel?
Hence, instead of having just one way of how to handle the craft, the craft's
handling characteristics extends over a certain range. This differs in a huge
way from F-Zero and WipEout HD. Even without the change in weight and
inertia, simulating the equations of motion like I've stated above is already
way beyond these two games (combined).
It's my believe that this will enhance the gameplay due to the sophisticated
dynamics of the craft and due to the induced tactical element which comes
along with it -- can be split over certain levels of difficulty.
And it gets even better. Once you are computing the inertia tensor again and
again, many cool things come for free. For example, guess someone picks up or
carries some bombs, mines, rockets etc. Those things will influences the
weight and the dynamics of the craft immediately (depending where on the craft
the cargo bay is located). So someone carrying a bomb may get a "penalty" due
to the increase in weight (and a reduction in speed) and change in handling.
Hence, weaponry can be taxed according to their devastation. Another part
could be that someone wracks part of your craft or vise versa. The loss in
weight immediately chances the handling of the craft. Etc.
But then there was this 4x evaluation of the time-depended force/torque term.
Some strategies to reduce the burden are;
- adaptive time-stepping (not so easy)
- adaptive order of integration (not easy)
- reduced order of integration for linear and angular position (easy)
(So only the linear and angular velocity will be integrated with rk4. However
the saving won't be so high.)
These are things I have to consider in the future. Since, from all the
formulas on may paper, I can readily taste the computational requirements
needed. If someone here has or had some similar problem of integrating some
more complicated differential equations with the urge to reduce the
computational burden without given up on the rk4, I'm would be pretty much
interested of how the problem was treated.
Of course, vectorization and parallelization of the code will speed up things,
but that's already implementation specific. If the theoretical method sucks,
its implementation will suck either.
One thing is for sure, I can put the PS4 to good use! xD
Damn Sony, send me a devkit, please!
That's all for now. Thx for reading!
I made some good progress to the equations that will drive the AG crafts of
ZONE. Everything is still done on paper. After getting all the forces and
masses (linear, angular, inertia) right, I'm now actually at the point to
integrate the derived equations of motion using a numerical scheme. I'm opting
for Runge-Kutta of order 4 (rk4) instead of simple Euler or the midpoint
rule, since the increase in stability is very much needed because the crafts
will undergo some hard/large forces during race making the equations a lot
stiffer and as such much harder to solve, much harder to track by a simple
integration scheme like Euler, which would simply blow-up making the craft
instable and as such uncontrollable. But the problem is, it comes at a price.
It requires me to compute the force and torque term four times per time step
(, per craft). Woahh! That's quite a heavy load. Usually the force terms are
way simple and multiple evaluation isn't so serve, but I want to go a lil
further with the AG crafts of ZONE than what seems to be the standard. For
example, I want to account for a cool propulsion system burning kilograms of
fuel exhausted from some nozzles propelling the craft. Burning fuel means a
change in weight of the craft, which makes the craft's mass non-constant, i.e.
m(t), and as such the center of mass becomes non-constant as well. So one has
to recompute the entire mass distribution, volume, moments etc. again, i.e.
one has to recompute the expensive inertia tensor and its inverse again and
again. Usually the weight and the center of mass, and as such the inertia
tensor are constant. This lessens the computational requirements in evaluating
the force term, which is still expensive if you have to do it four times. But
given the non-constantniss of mass amplifies the problem in a big way due to
the recalculation if the inertia tensor.
That's where I'm standing.
Now one could say; hey, dude, skip the loss in weight due to the burning fuel!
Sure, that's the easy way out. But let me explain way I think this is actually
all cool stuff. Btw; You can make a comment about it if you think otherwise or
if you share the idea in some way.
Well, when you carry just a few kilograms of fuel with respect to the total
mass of the craft, you won't recognize a huge change in angular momentum if
the fuel tank isn't a kilometer away from the center of mass. However, guess
for example half of the craft's weight is fuel! A slight displacement off form
the center of mass (hence changing it) has a large effect on the angular
momentum of the craft. Now if the fuel burns then its weight will be reduced
and as such the weight of the craft, leading to a change in its center of mass
(usually) and a change in inertia of the craft. Hence, during race, the
dynamics, the behavior of the craft will change and you as the player have to
account for it, making it way more challenging to race. So for example, at the
start of a race, your AG craft is heavy loaded and the dynamics/handling of
the craft reflects this load. During race you will lose fuel making the craft
lighter and as such faster while the dynamics will change as well due to the
change in center of mass and inertia over time. So you have to know your
craft! It becomes a strategic factor! You may save some fuel (and as such
carrying the weight with you all the time) till the end of the race to ignite
your afterburner for victory! Nothing comes closer winning the race last turn,
I tell you! So it may pay off being able to handle the craft well under a
heavy load to save fuel for some precise afterburner bang-bang maneuvers. Oh,
did I told you may get out of fuel?
Hence, instead of having just one way of how to handle the craft, the craft's
handling characteristics extends over a certain range. This differs in a huge
way from F-Zero and WipEout HD. Even without the change in weight and
inertia, simulating the equations of motion like I've stated above is already
way beyond these two games (combined).
It's my believe that this will enhance the gameplay due to the sophisticated
dynamics of the craft and due to the induced tactical element which comes
along with it -- can be split over certain levels of difficulty.
And it gets even better. Once you are computing the inertia tensor again and
again, many cool things come for free. For example, guess someone picks up or
carries some bombs, mines, rockets etc. Those things will influences the
weight and the dynamics of the craft immediately (depending where on the craft
the cargo bay is located). So someone carrying a bomb may get a "penalty" due
to the increase in weight (and a reduction in speed) and change in handling.
Hence, weaponry can be taxed according to their devastation. Another part
could be that someone wracks part of your craft or vise versa. The loss in
weight immediately chances the handling of the craft. Etc.
But then there was this 4x evaluation of the time-depended force/torque term.
Some strategies to reduce the burden are;
- adaptive time-stepping (not so easy)
- adaptive order of integration (not easy)
- reduced order of integration for linear and angular position (easy)
(So only the linear and angular velocity will be integrated with rk4. However
the saving won't be so high.)
These are things I have to consider in the future. Since, from all the
formulas on may paper, I can readily taste the computational requirements
needed. If someone here has or had some similar problem of integrating some
more complicated differential equations with the urge to reduce the
computational burden without given up on the rk4, I'm would be pretty much
interested of how the problem was treated.
Of course, vectorization and parallelization of the code will speed up things,
but that's already implementation specific. If the theoretical method sucks,
its implementation will suck either.
One thing is for sure, I can put the PS4 to good use! xD
Damn Sony, send me a devkit, please!
That's all for now. Thx for reading!
Well, it seems that I can't test Wp7 apps on the Wp8 emulator because the SDK doesn't support it.
I'll try the forum, maybe I can get some help. Thanks!
Meanwhile, I've temporarely removed the game from the store. Better that then having unhappy users because the game crashes...
Dynamite Shikoku
Congratulations, you really deserve it!
Is there such a genre as 'twitch platformer'? Or am I imagining it. I'm thinking about how to describe my game for when I start pimping it soon.
FormallyMegadrive
Member
Definitely, Super Meat Boy or perhaps I Wanna Be The Guy could be in this subgenre.
Dynamite Shikoku
Congratulations, you really deserve it!
Cheers. That's what I was thinking of.
Missile: Have you thought about pre-compiling data on the loading of levels/races/whatever?
I am not saying that you should not do real time computations on it, but those should be left to things that really matter (i.e. explosions, crashes, stuff like that). When you have a race between n vehicles, you can generate a lot of stuff without knowing what weapons they will pick up or where they will go. You can still compute their basic datasheets for linear movement, for some basic angular movements (and mesh different angles together if you want a more precise data), for being in a fully loaded state, and half-loaded/near empty as well. Maybe have a set of pre-determined damages as well, and only calculate what state the vehicles will end up being in, freeing up tons of computational power.
Not to mention that if you have these covered, you can still use real-time simulation for all the missiles and all the vehicle parts that are flying off after an explosion/accident/Whatever. But everything else is just a set of similar motions, happening in a 3 dimensional plane where the only thing that changes is the space offset (I.e. whether you turn left in x:500,y:500,z:300 is exactly the same as if you were turning left with the same vehicle at stage point x:1500, y:1200, z:-300)...
Just some ideas. I hope they make some sense
I am not saying that you should not do real time computations on it, but those should be left to things that really matter (i.e. explosions, crashes, stuff like that). When you have a race between n vehicles, you can generate a lot of stuff without knowing what weapons they will pick up or where they will go. You can still compute their basic datasheets for linear movement, for some basic angular movements (and mesh different angles together if you want a more precise data), for being in a fully loaded state, and half-loaded/near empty as well. Maybe have a set of pre-determined damages as well, and only calculate what state the vehicles will end up being in, freeing up tons of computational power.
Not to mention that if you have these covered, you can still use real-time simulation for all the missiles and all the vehicle parts that are flying off after an explosion/accident/Whatever. But everything else is just a set of similar motions, happening in a 3 dimensional plane where the only thing that changes is the space offset (I.e. whether you turn left in x:500,y:500,z:300 is exactly the same as if you were turning left with the same vehicle at stage point x:1500, y:1200, z:-300)...
Just some ideas. I hope they make some sense
Indeed. It's not going to be easy going through all of this. Guess for example
you are four people without a track record in the gaming biz.... And on the
other hand Sony wants to see a business plan and a throughout proposal of your
game, tools used, and background in the field. So you better come up with
something real kicking (being new), I guess. I think going the standard way
will be as difficult as ever. But that's where the indie lane comes into play.
But I don't expect that we can execute native code in the PS4 or allocate any
huge resources like the registered developers can. Anyhow, first things first.
I would recommencement doing so. Most programmers here aren't willing to spent
their (free) time on a boring job. Being indie means doing something very
exciting, something you can't let go off even if you won't get any money out
of it at first or second try. Well, I think, money isn't the driving factor
here. You have to know, most devlopers over here do already have a boring
(day) job working for the money. Hence catching some programmer interested
you'll better come up with something exciting. For example, you may present
your idea and show how a programmer from over here may fit in. That's the way
I would do it.
Game sounds horrible. But it looks good!...
Pretty interesting, mate! Love the model! How long does it take you to modelDoing another quick stream today. Going to be showing my dynamic sky lighting model that builds spherical harmonics based on the sun's direction:
...
something like this? Spherical harmonics are cool. :+
That's how it was done in the past.
Well, indeed, doing pre-computingwill help. I could build tables of valuable data and interpolate in-between to
get the values (forces) needed at a certain stage. That's something I have to
do later on nevertheless, i.e modeling. May plan of attack is just different
this time. Instead of building a model and fix it right at the start on some
pre-computed data or artificial damping, I though I will implement the
general model first and cut down later on until everything fits. This serves
the game much better because going down on the formulas, i.e. in doing more
rough approximations, is way easier than the other way around. So for example
I can say we may approximate one or another feature/characteristic of the
craft much coarser but only up to the point where some other feature are
still working rightfully/truthfully. Cutting down is much easier that cranking
up.
Let me give an example where I'm using this approach already; The general idea
is to have the craft modeled by a tetrahedral volume mesh. Each tetrahedron
can be given different physical properties describing the material at the
location the tetrahedron is places within the craft. Now it would be awesome
to be able to compute the inertia tensor of the craft over all the tetrahedra
gaining a pretty good inertia tensor leading to a very precise feeling for
the craft. This is the ultimate situation, but unfeasible on anything < PS4.
For one, the amount of tetrahedra will be quite huge. Whereas surface
polygonalization increases on the order of n², for a volume it goes with n³.
All these tetrahedra have to be pulled from RAM (did I say we need lots of RAM
as well) to the compute units rather quickly (176GB/s GDDR5, oh lord!) where
the inertia tensor is computed over all the tetrahedra and summed together.
The problem here is that the data is of order n³. So it's pretty important
being able to pull all this data from RAM very quickly feeding the units.
That's where the PS4 will really shine, since most of the problems within
physical computations are memory bound. So this 176GB/s of system memory
bandwidth is really a huge deal. The geometric fidelity will increase in order
of magnitudes on the PS4. However, since this approach isn't feasible now, I
have to make an approximation. But the approximation should be done in a way
to retain the main characteristic of the inertia tensor (the dynamics the
craft undergoes with respect to its weight distribution). That's something I
don't want to lose, since it makes the movement of the craft feel so natural.
Now what? Cutting down is easy, as I stated above. And indeed it is. Depending
on the performance of the system in question many of the tetrahedra will be
lumped together and approximated by a much simpler geometric shape for example
by a box, cylinder, or sphere getting averaged material values. These shapes
are now used within the computation of the inertia tensor speeding up its
computation in order of magnitudes while relying on simple larger shapes
easy to compute an inertia tensor for. This will cut down on the required
memory bandwidth as well as the demand put on the compute units. Despite the
fidelity may suffer the over physical behavior is still there since a real
physical computation is done even on the simplified model.
Now we can cut down on the physical model / equations as well, for example
using a simpler integration rule (Euler for example) as I've described above.
But this will ultimately blow-up the simulation in my case and will lead to
instabilities. So better to cut down on the data first before considering
simplifying the method. But once the data is at its minimum, the method has
to be simplified, of course.
That's the approach I use; sticking to the real physics but simplify the data
as much as necessary. For me, this is kind of a perfect approach. The kicker
here is that you have to consider a more general case beforehand, something
which will cost you time, but may save you a lot of time later on. Even if
not, I do it for the gamer. Further, I know that tweaking parameters in some
less physical model takes huge amounts of time for getting the behavior right.
And that's something I'm not willing to go into. Tweaking parameters of a
stiff "physical" model is hell, and the player recognizes it like nothing
else, i.e. if the vehicle does an abrupt and unexpected maneuver. And, if one
is really into the game, one recognizes the limits (the stiffness) rather
quickly if it comes to the physics. ZONE should be different here. As better
the physical model, as less tweaking is required to get the feeling right. But
tweaking will be required nevertheless, but in a way making the game more fun
instead of tweaking it in such a way to get the basics right.
Sounds hell complicated. And I have to admit, some things are a lot more
difficult then necessary with respect to games. At times I can't believe it
either what's necessary to pull out to get something really cool. I mean, for
example, Spherical Harmonics are deeply rooted mathematical and physical
functions. Now they belong to the standard set used by designers and graphics
programmer producing cool looking indirect lightning reflections in realtime.
Hell, that's awesome, isn't it?
BlastProcessing
Member
You know that thing that happens where you have been working on something for so long that you have no idea if it's any good or not? Well, we may or may not have totally hit that wall and are looking for some playtesters to give feedback.
The game is called Ephemerid. It's a breezy sort of musical adventure game. The intent is to make something that can be played through in one sitting like an album. We have a pretty old trailer here, but it still give a good idea about the vibe we are going for.
The game should run on any iPad 2 or newer. If anyone is interested in setting aside a half hour or so to check it out just click on the ol' Test Flight Link. If you have any questions or feedback hit me with a PM or shoot an email to mail@superchopgames.com.
The game is called Ephemerid. It's a breezy sort of musical adventure game. The intent is to make something that can be played through in one sitting like an album. We have a pretty old trailer here, but it still give a good idea about the vibe we are going for.
The game should run on any iPad 2 or newer. If anyone is interested in setting aside a half hour or so to check it out just click on the ol' Test Flight Link. If you have any questions or feedback hit me with a PM or shoot an email to mail@superchopgames.com.
Has Sony said anything about an indie publishing platform for the PS4 similar to PS Mobile?
It looks good. I liked it. I would help beta test it, but I'm terribly busy at the moment. I'll think about it anyway.
Up to my knowlegde, nope. But I assume it boils down to exactly this. For one
reason or another, I am still not convinced about how Sony approaches the
indies. But I'm pretty sure that those shown on E3 all had devkits at their
disposal. So how is this going to work? Well, I think, either your are
sandboxed or just another developer/company having signed an NDA with Sony.
I don't expect something huge so to speak, but I hope they will come up with
a strong SDK and a service making developing and publishing indie games a snap
on the PS4. That would be great. Perhaps the indies at E3 already beta-tested
the SDK / infrastructure? Dream on, missile!
At least one PS4 indie at E3 didn't actually have a dev kit. They just got a PS4 controller working with their existing code.
^ Game might have been running on a "PC" during presentation?
I'm not sure what to make out of those indies playing their games on E3. Were
all these games sandboxed, from PSM over to PS4? How do they got on the
platform without going the standard way via http://us.playstation.com/develop/,
which would obviously draw no distinction between a standard PS3/4 developer
and an indie developer. So how is the concept envisioned?
I'm not sure what to make out of those indies playing their games on E3. Were
all these games sandboxed, from PSM over to PS4? How do they got on the
platform without going the standard way via http://us.playstation.com/develop/,
which would obviously draw no distinction between a standard PS3/4 developer
and an indie developer. So how is the concept envisioned?
How "indie" were all those companies? Could they have been given loaner dev kits like the Sony site mentions? From what I understand PSM runs on devices that have really low specs, (Android phones and a gimped Vita) would it make sense to port something from it to the PS4?
RobotNinjaHornets
Banned
My progress on that sprite, in between coding, GAFing and packing. So much effort D:
Not really happy about the shirt at the moment, seems too flat. Oh well.
nukembashi
Neo Member
I didn't want to recommencement (I love that word, so versatile) to a job website simply due to the fact that I would have had to spend just as much time wading through worthless applicants as actually doing the job myself. If I went to a place like Jobs.com, Monster, or Craigslist to find someone, it'd have to be for a large project, or else the investment I put into looking for a qualified applicant wouldn't be worth the effort.
I suppose that you answered my question. Thanks.
Switching gears, despite not being able to find some hired help to do small stuff, does anyone have any recommendations for finding non-programmer staffing for indie projects? I can do writing on my own, but what about voice acting and artwork? Small-time voice actors and on-commission portrait artists are usually available, but I'd need a good storyboarder or a voice actor that's used to reading full act-sized dialogue lines. This thread may be largely about the programming side of indie game development, but what about the rest of the game?
McNerdBurger
Neo Member
Most of the pro voice actors are part of a union and they get expensive very fast. I'd recommend you look at theatre groups in your local area if you want someone affordable. If you have the cash though I'd recommend contacting a casting agency and setting up auditions.
If you're looking for sound effects work (or help with implementation), IM me. I'm freelance now but I was full-time at Bioware for the last few years. I'm always interested in trading skills (audio work for art work, etc.)
BomberMouse
Member
Are you a coder? Have you ever programmed professionally? Because otherwise you have zero rights to estimate the bolded parts. I know you address that on the next paragraph but it needs to be said again. Those signs in your "looking for programmer" post are a huge turn off to any freelance programmer. Anyone who makes a living coding knows to avoid ads that basically boil down to "I have a great idea, it just needs a little programming. I can't pay much but you can do it in your spare time".
No one wants to work for other people projects, specially when they are something as personal as a game. Just imagine a painter outsourcing the actual painting, "I have this image so clear on my head, you just need to paint it". Ridiculous.
If you really want to get someone to help you:
- Don't devaluate other's people work from the start saying "it's something small". That's not up to you to actually evaluate.
- Give some factual data of what you need to be done. If you can't elaborate then you have no idea what you are talking about and need to do some research.
- Programming is an art. If you want a "non professional" programmer you'll get messy results. If it's something so easy a "small time programmer" why not do it yourself?
- Have more respect for other people spare time. Saying "Just work on you spare time and be happy to be paid minimum wage (or less)" is a no-no.
I guess what I'm trying to say, your ad looks disrespectful to any potential employee. It basically boils down to "I'm looking for a small-time programmer who's looking to work in their spare time for as cheap as I can on a non descriptive task that I can't articulate but it's very small and easy so I shouldn't pay you much for it.
I'd also recommend Voices.com for VO work. Basically you post an ad and get VO samples and bids by each interested voice actor. I used a few 5-star rated (trusted) actors on there for an iPad project to great success. Cost a few hundred bucks. I think I recommended them before in this thread but worth mentioning again.
Edit: trading skills is a great idea by the way. Never thought about that approach to game dev before.
Rest assured: it does!
But on the bright side, once you have the ability to emulate physics in a certain scene with a simplified model or the real model, and scale down the tracked objects or up, depending on the hardware you are playing, you pretty much got the engine that will last for quite some time : )- Status