Game Development - 2015-07-30 (page 2 of 4)

« first day (1820 days earlier) ← previous day next day → last day (3197 days later) »

1:00 PM

@Tyyppi_77 get the hit position, instantiate a sprite, make it slightly transparent, rotate it randomly, change the scale slightly to stretch for some variation

@Joe you can't switch on non-const cases in C#

basically in each case whatever:, the `whatever has to be compile-time constant

@Joe I have a Material, and a bunch of declared Materials, and I want to compare the material to see which material it is, but C# is saying it has to be an int, string, enum, or something basic

you can cheat it I think... switch(true){ case (newCondition >999):

dont know if youd want to, or if that helps

meh I'll just if it

user92578

We were talking about this blood right?

user92578

1:03 PM

user image

@Chris So ... Dictionary<Material, MaterialType>?

the blood against the solid ground blocks are heavily skewed/scaled

if anything I think I change the rotation on them

user92578

How do you make sure that it doesn't render the blood sprite on the wall?

@MartinSojka that'd be a good way actually, store the key and what to do

@Tyyppi_77 I don't. there's always a wall, plus it's heavily shadowed, so I can get away without having to mask or clip any calculations

so it's always partially overlapping the block onto the wall, but it looks fine -- is part of my art direction I chose, bit of an easy solution

As a recovering C addict in the C++ phase, it's a little painful to invoke a hash table lookup when the type ID could be jammed into the Material struct and recovered with a single dereference

1:08 PM

if I had a scene like your outdoor stuff, I'd probably choose a few pre-made sprites that are cut off, and choose them to fit

hmmm have to do android debugger attaching over the network... doesnt work via usb or bluetooth

user92578

Oh right the shadow borders between the ground blocks and the back wall hidens the overlap I suppose

@Joe wtf lol, I had to use USB and the other options didn't work :P

@Joe you could try using the ADB Log thing and just see?

user92578

Yeah I can't really do that in GunHero

1:09 PM

@MickLH i can use unity remote on USB

user92578

But thanks for telling me how you did it!

but I cant get the damn usb working for on device app build

user92578

Okay G-gaps won't show atleast for now. SDL_TTF is stretching the font so the gap is disappearing

@Tyyppi_77 You can scissor your blood rendering to get the same clipping effect

user92578

That'd probably be quite tricky to implement

1:10 PM

@MickLH That would be my preferred solution if I can change Material. I might not though - it might be a third-party library, for example. A dictionary lookup is the "safe" (if slower) variant.

@MartinSojka Oh I saw the validity in your reasoning, I'm not calling anyone out except myself :P

@Tyyppi_77 Are you using SDL rendering functions, or OpenGL

user92578

SDL

Then I have no idea, but in OpenGL it's literally as simple as:
glEnable(GL_SCISSOR);
glScissor(x, y, w, h);
// render stuff here
glDisable(GL_SCISSOR);

So I'd expect SDL to expose similar functionality

user92578

Okay

user92578

Is the area the area where it is allowed to render, or the area where it is not allowed to render?

1:13 PM

Where it is allowed

user92578

Ah, okay

user92578

Well that sounds quite simple then actually

opengl.org/wiki/Scissor_Test

It's usually hardware accelerated too

user92578

I gotta write this down for when I actually try to implement the ground blood

user92578

Thanks

1:15 PM

np, good luck lol

@Chris lol that log thing is going nuts

filter it with Unity

tag:Unity I think?

and that will display debug.log messages?

I think so, unless it's reported as your app name or something, pretty sure it will though

that was way easier than what i was trying to do *Thank you (was the purpose of this message)

1:30 PM

all good

this CSharpatron is crazy. as I toggle the C# scripts it created to become active (meaning it hides away the UntyScript file and enables the C# file in its place), if there are dependencies of files that I haven't toggled live yet, it creates a dummy copy of the dependent script it needs, just a barebones class of the same name with empty methods and empty vars, just so the compiler doesn't cry, meaning I can carry on to test the next file, it's great

1:45 PM

sounds awesome what did that cost you

$45 USD

really? thats cheap if it works well

so far so good, not done yet though (so will reserve final judgement til it's all converted and done lol)

yeah, has like twenty 5-star ratings

people saying it trumps all others they've tried

think I wasted one week last year trying to convert this over

idk call me crazy, theres something cool about porting... seems like an interesting way to learn a new language

and just to have all my prefabs remain intact, it's some sort of crazy wizardry, much impressed

yeah

1:51 PM

are you planning on segmenting this into steps

or just making a change over all at once

I honestly learned a lot about C# through the chat in here hahaha

or at least seeing it and being exposed to it, warmed up to it ;)

i never touched C# before this chat

sxufdghizdurfhgzidthpzxidutjhzodtuhj FL STUDIO WHY DO YOU GLITCH CORE CRASHED SOUND CARD REMIX EVERYTHING

:(

Da-DingDa-DingDa-DingDa-Ding HAHA ITS GOOD HAHAHAHAHA HAHAHAHAHAHAAHAHAHAHAHAHAHAHAHA HAHAHAHA

killall -9 FL.exe :(

1:52 PM

Well, before I started having to enable these dummy classes, it had moved over my non-dependent classes all fine, prefabs intact, game ran

@Chris actually i was writing js in unity for about 4 months before i saw your first message about compile times and I switched that day

so atm I'm in non-testing mode, assuming it'll all work while I convert some 400-odd files remaining

@Joe hah, good on ya.

uh oh @MickLH has finally lost it, what happened

$ killall -9 FL.exe happened

You write something amazing, FL studio crashes, you lose all your work. Just another 10 minutes in the life of a musician.

some things are so amazing, the world doesnt deserve them? (he said realizing this was the most emo thing ever written)

1:57 PM

*Every time I try to make a silly little song
My efforts are all wasted 'cause machinery goes wrong*

Computer Song - Jim Noir - HD quality

@Chris C# is great

Hi! Good day, you guys familiar with bezier curve??

Yes

But don't ask to ask, please. (Just ask your question instead)

Lol, do you mind?

Nope :P

Kevin van der Velden

2:04 PM

No, but it saves time =p

But others may

Kevin van der Velden

Plus, people sometimes read but don't say anything. However if your question is interesting they'll respond (source: I've done this many times)

^ (I only responded because the timing of everything was really lucky)

++

Kevin van der Velden

Also there's a log so someone can come in in a bit and see the question and answer it, while otherwise you'd be stuck at asking to ask

2:06 PM

user image

Oh.. I got shy, haha nevermind

@jane Bezier curves are linear interpolations between linear interpolations. What do you want to know? I'll have to take a 5 minute break in about 20 seconds

...

@Pip Great won't cut it. It's excellent!

Kevin van der Velden

@jane do ask! Bezier curves are cool

Hey guys is windows 10 really this bad?

Everyone seems to be having so many problems

Kevin van der Velden

2:07 PM

Jul 18 at 13:15, by Kevin van der Velden

user image

wait, I'm reading that Microsoft bought Pando Networks in 2013: techtimes.com/articles/39960/20150316/…

Kevin van der Velden

@TheMuffinCoder it's the first day of a new operating system, wait another month

@jane bezier curve image :D

@KevinvanderVelden yea i was planning on doing that :P i like windows 8.1

I do have the preview on a older computer and had no problems other than no audio worked

I'm just really having a hard time on how I would calculate it on my code :P I'm very stupid when it comes to math.

Kevin van der Velden

We've coincidentally talked about this a few months ago so I'm just gonna quote myself:

Jul 18 at 13:21, by Kevin van der Velden

Curve3( p0, p1, p2, p3, t ) = lerp( Curve2( p0, p1, p2, t ), Curve2( p1, p2, p3, t ), t )
Curve2( p0, p1, p2, t ) = lerp( Curve1( p0, p1, t ), Curve1( p1, p2, t ), t )
Curve1( p0, p1, t ) = lerp( p0, p1, t )

2:09 PM

Lol yeah the lerp pyramid is gonna be the lightest load on your math brain

What is lerp pyramid??

Kevin van der Velden

@jane not a specific term

Haha, see stupid questions lol

Kevin van der Velden

Lerp is a linear interpolation (from point A at t=0 to point B at t=1)

user92578

@KevinvanderVelden :)

Kevin van der Velden

2:10 PM

And a pyramid cause, well, it's a pyramid shape kind of :D

Everyone!!! Gangplank is dead in league

Kevin van der Velden

lerp(
	lerp(
		lerp( a, b, t ),
		lerp( b, c, t ),
		t
	),
	lerp(
		lerp( b, c, t ),
		lerp( c, d, t ),
		t
	),
	t
)

user92578

BTW I never got the solving for y working

Lol it's kindof a trip.... I want to post what makes it clear to me, but I bet it will confuse the situation worse

Alright, so if I have p1 and p2 as a line segment... how would I be able to calculate the tangent? I have searched some answer... but really, my brain is not getting anything at all.

2:13 PM

@jane You want the derivative of your bezier curve?

I'm interpreting "tangent" as the tangent line at a given point on the curve

Yeah, exactly!

Making things with maths - Steven Wittens

So it's called derivative?? Is that the term I'm looking for? Lol

@jane Well that's a big part of it, you actually want the first order taylor series

@jane I applaud you for trying to figure it out. I don't like math either

Kevin van der Velden

2:15 PM

@Joe I just like the gif I posted because I was looking at making bezier curves and couldn't wrap my head around it... and then I saw that gif on wikipedia and it clicked

@jane Which is just y0 + (x - x0) * Derivative

Kevin van der Velden

@Joe Also I will watch this in a bit

this whole talk is amazing

@jane [x0, y0] would be the point you want the tangent line to cross the bezier curve

this gave me a grasp on alot of concepts i i hadnt considered before

2:17 PM

I will download it and watch it on my way to work tomorrow hehe thanks @Joe

I think i actually got it from a post here maybe a year ago.. i dont remember who linked it

Kevin van der Velden

That exact link isn't found in the search, just claim it for yourself

What is that equation for?

@jane The equation I listed is your tangent line, in y = m * x + b format

Kevin van der Velden

Which one? (if you hover over a message there's an arrow on the right side of it which you can click to reply to a specific one)

2:21 PM

@MickLH so that equation is derived from y = m * x + b?

It is y=mx+b, where b = y0, and m = Derivative

It's shifted on the x-axis also, to be centered on your point

@Chris thanks! :)

I'm not sure how to explain this without resorting to calculus

m=(y2-y1)/(x2-x1) For linear curves

You're fundamentally asking for a calculus operation

Kevin van der Velden

2:24 PM

Damnit now I want to write a raymarcher or something

Hmmph

@KevinvanderVelden Prototype my "Shadow Map Bending" idea :P

It uses ray marching to re-project shadow maps for small movements of the light source

Kevin van der Velden

Maybe, I'll need more info =p

Am I giving you a hard time? @MickLH sorry XD

@jane No it's all good :) I am just trying to not blast you with math since you asked specifically for not that :P

Kevin van der Velden

Though... hmmm, I could create a prototype that dynamically shoots rays and updates the texture in idle time

2:26 PM

@KevinvanderVelden If you actually want to work on it with me, I can get my ass in gear and get a first draft of a whitepaper finished up

Kevin van der Velden

Sure, don't know how much time I'll have

Given that I have about 50 bazillion projects I want to work on =p

lol I understand... all too well

user image

@MickLH the equations you gave to me are hints to my treasure.. haha, thanks for the effort!! I'm just gonna dig deeper in internet
hehe

@jane I'm sorry I must do this, but I need to break the stagnation in our conversation.

d/dx lerp(f(t), g(t), t) = t*g'(t)+(1−t)*f'(t)+g(t)−f(t)

So, @jane, how far into that do you get before calling "fuck it" ?

2:30 PM

So this is calculus?

Haha

Yeah that's the analytic derivative of your function in closed form. (Aka, after the game has been won)

I like art, and my brain considered that as abstract. Haha

What is that lerp?

Linear interpolate

If I have to derive a formula from that equation....

lerp(a, b, t) has t as a percentage (from 0 to 1)

lerp(a, b, 0) = a

lerp(a, b, 1/2) = (a + b) / 2

lerp(a, b, 1) = b

2:33 PM

So a is the starting point and b is the end point?

Yes

And that one in the middle is....

Midpoint?

Please be very specific, which one in the middle?

:23110705

Oh

ah, that's just to show you, that if you give it 1/2, it calculates the middle for you

2:35 PM

That with 1/2 at lerp

And if you give it 1/4th, it gives you the point 1/4th away from a, 3/4ths away from b

So why do i need the middle point?

What middle point? In your bezier curve?

Ahh... yeah?

Ooohh

Lerp from start to middle, lets call that "Start Middle"
Lerp from middle to end, lets call that "End Middle"
Now your Bezier is the lerp from "Start Middle" to "End Middle"

2:37 PM

oohhh!!!

That would be a 2nd order bezier curve.

Now I get it!

Kind of

You can do it even a level deeper, that's a 3rd order curve.

Here's the picture for that:

30 mins ago, by Kevin van der Velden

Jul 18 at 13:15, by Kevin van der Velden

user image

You could do as many levels as you want

I think I'm getting it :)

I called it a pyramid earlier, because the math goes like this:

$user image$

2:39 PM

You understand that?

Lol yes

Applied mathematics is the art of chaining together lots of tiny functions, each of them doing almost nothing, in just the right way.

so, programming

Just pretend it's a B for bezier @jane :P

user92578

user image

2:41 PM

@Chris Only if you're programming in side-effect free languages, or avoid functions and methods with side effects (like everything involving IO).

true

Alright haha, now I have an idea. :) thanks a lot @MickLH you're my math hero!

but really, like anything, just breaking things down and understanding how every little thing does its job

Dat mutable state doe

There should be a nerd comic about this.

Something something, Shows protagonist pages of hard math...
> No problem!

Alternate reality, Same exact hard math, but one function saves and loads with a variable...
> Nuclear bombs going off, shows galaxy with exploding earth bits flying by

user92578

FUCK THE RENDERING ARTIFACT LINES ARE BACK

user92578

2:50 PM

MY FUCKING ARCH ENEMIES

@MickLH Got a gif of a 2nd Brezier?

2nd order that os

user image

Normally the 3rd order is used, because it feels intuitive

And is flexible enough to make a passable approximation of a circle by hand

Cool

That's really cool actually.. Why don't we learn this in school?

With the 3rd order curve, P1 is like a "softened" but mostly direct control over the derivative at P0

Yeah

2:58 PM

You can also approximate Euler spirals pretty damn well with them.

(Those are used as transition curves in highway and railroad engineering).

user92578

Sorry, I know that you guys just got off helping someone with beziers

user92578

But

lol, still getting off @Tyyppi_77

user92578

I have implemented the nested lerp calls for tween curves to my framework

user92578

And I plan on using the beziers so, that they always start at 0,0 and end at 1,1

user92578

3:00 PM

See cubic-bezier.com/#.17,.67,.83,.67

I predicted the link accurately

user92578

Now given the x value (the Time in the link), I would have to solve the y value (Progress) from the curve

user92578

So basically that's a sort of limited bezier I guess

I tackled this problem a while back for my bezier renderer

user92578

I have no idea how I would do something like that, but I guess it shouldn't be as hard as solving a "real bezier"

3:02 PM

It was ugly lol, yes you will definitely want to take advantage of the special case

user92578

Yeah I researched a bit on solving beziers a while back, did not look nice

Basically, solve for t from x, then plug that into your equation for y

Solving for t from x is the ugly part, because it's 3rd order polynomial

user92578

Ah, see I was confused with the t & x, probably assuming that in the special case they were the same

user92578

So is that for the special case?

Well lol, your special case is why it's possible to "just plug it in"

user92578

3:05 PM

Oh, okay

Solving for t will give you 3 answers

But in your special case, 2 of them will be complex

So just take the real valued t and you're set

user92578

It always takes me a while to get this stuff because I first have to realize what these things mean in Finnish

user92578

Aight, third degree polynomials, I think I might be able to do that

user92578

Do you do this stuff on a paper Mick or do you use some software?

maxima

sometimes it chokes on hard integrals and you have to bust out mathematica but for the most part it's good enough

user92578

3:10 PM

Three versions in sourceforge.net/projects/maxima/files/Maxima-Windows/…, which one do I want?

Kevin van der Velden

Biggest one

It's the same program with different runtimes bundled, I'd pick the fastest one

user92578

The one with most dowloads

The program is really not any kind of high performance anything :P

user92578

So first I want to get the equation for t using x, right?

3:11 PM

It needs any help it can get

@Tyyppi_77 Well, that's a way about the problem, I'm not sure if it's the only way.

user92578

and x in my case is the "Time", so a value from 0-1

user92578

Now if I just knew how to use this shit

Define lerp: lerp(a,b,t):=a*(1-t) + b*t;

Define cubic:
cubic(a,b,c,d,t):=lerp(
    lerp(
        lerp(a,b,t),
        lerp(b,c,t),
        t
    ),
    lerp(
        lerp(b,c,t),
        lerp(c,d,t),
        t
    ),
    t
);

user92578

Yeah that what I was looking into

user92578

wait why do I implement lerp like that?

3:16 PM

because it's easiest for me to type it that way without thinking

user92578

But a + (b - a) * t is okay too?

whatever you like!

Kevin van der Velden

Probably more calculations for higher dimensional values

user92578

you dont want to tell me why my version is worse than yours? :D

the program can convert between them really

It doesn't matter

Type it my way, then type ratsimp(lerp(a,b,t))

Kevin van der Velden

3:17 PM

Though, hmm

I wonder which is faster

Meh

returns: (b−a)*t+a

user92578

Oh, cool

user92578

fuck i forgot the ":"

it wont go back to my way with factor() lol

it goes to b*t−a*t+a instead

which gives a tiny insight I hadn't noticed before

user92578

@MickLH you really typed all this in a surprisingly short period of time?

3:20 PM

Lol the ridiculous part is that I should have copied what @KevinvanderVelden typed earlier

Kevin van der Velden

@MickLH yes

I think it's letter for letter the same, except whitespace

user92578

LOL

Kevin van der Velden

1 hour ago, by Kevin van der Velden

lerp(
	lerp(
		lerp( a, b, t ),
		lerp( b, c, t ),
		t
	),
	lerp(
		lerp( b, c, t ),
		lerp( c, d, t ),
		t
	),
	t
)

@Tyyppi_77 wanna see something pretty? type cubic(a,b,c,d,1/2)

user92578

3:22 PM

all the fractions

lol wanna what happens if you try to copy paste it here? ((d/2+c/2)/2+(c/2+b/2)/2)/2+((c/2+b/2)/2+(b/2+a/2)/2)/2

user92578

(((d-c)/2-(c-b)/2+c-b)/2-((c-b)/2-(b-a)/2+b-a)/2+(c-b)/2-(b-a)/2+b-a)/2+((c-b)/‌2-(b-a)/2+b-a)/2+(b-a)/2+a

user92578

Why is mine so much longer???

that's what he said

user92578

Because I used the other Lerp implementation?

Kevin van der Velden

3:23 PM

Probably

probably

lol

Bernard the singing Poo

Kevin van der Velden

For 1/4, simplified: (d+9*c+27*b+27*a)/64

user92578

Okay, and then I need to solve the t

user92578

So it's Cubic(0, a, b, 1, t)?

user92578

3:27 PM

(-3*b+3*a+1)*t^3+(3*b-6*a)*t^2+3*a*t

lolokok srstim agn

Kevin van der Velden

(d-3*c+3*b-a)*t^3+(3*c-6*b+3*a)*t^2+(3*b-3*a)*t+a

That's for generic t, huh

user92578

And, but my case simplifies it just a tad

Kevin van der Velden

Yeah

Essentially removing one variable

user92578

And then solve(Cubic(0, a, b, 1, t), t)?

Kevin van der Velden

3:29 PM

Try it and see!

user92578

Which gives me three answers, and one of them is 0?

user92578

Can't be right

Think about what you're asking it to do

Hint: If you give it an expression instead of equation, it assumes you want the root

user92578

Ah

user92578

But instead I just want the equation

3:30 PM

solve(cubic(0,a,b,1,t) = x, t);

user92578

AH RIGHT

enjoy the horror!

[t=(−(sqrt(3)*%i)/2−1/2)*(sqrt((9*b^2+(−18*a−6)*b+9*a^2+6*a+1)*x^2+(−4*b^3+6*a*b^2+(6*a^2+6*a)*b−4*a^3−12*a^2)*x−3*a^2*b^2+4*a^3)/(2*(3*b−3*a−1)^2)−(a*(b*(18*x−3)−6*x−3*b^2)−9*b^2*x+6*b*x−x+a^2*(−9*x−3*b+6)+2*b^3+2*a^3)/(−54*b^3+a*(162*b^2−108*b+18)+54*b^2−18*b+a^2*(54−162*b)+54*a^3+2))^(1/3)+
(((sqrt(3)*%i)/2−1/2)*(b^2+a*(−b−1)+a^2))/((9*b^2−6*b+a*(6−18*b)+9*a^2+1)*(sqrt((9*b^2+(−18*a−6)*b+9*a^2+6*a+1)*x^2+(−4*b^3+6*a*b^2+(6*a^2+6*a)*b−4*a^3−12*a^2)*x−3*a^2*b^2+4*a^3)/(2*(3*b−3*a−1)^2)−(a*(b*(18*x−3)−6*x−3*b^2)−9*b^2*x+6*b*x−x+a^2*(−9*x−3*b+6)+2*b^3+2*a^3)/(−54*b^3+a*(162*b^2−108*b+18)+54*…

user92578

OH MY GOD

Kevin van der Velden

user image

It's better with colours

For some reason, I don't know if this is intended or just a side effect of the algorithms inside... but in cases like this where there is only one real answer, Maxima likes to put it at the end.

user92578

3:32 PM

So I want the last one?

Probably

also there is a last() function, to get the last item from an array, conveniently

Kevin van der Velden

I don't think he wants it at all

user92578

Is there a parameter that picks the last result line?

Kevin van der Velden

Do you want to take an output, the a and b values and then calculate the t from that?

If so: That's what the returned equation from solve does!

If not: then you use the simplify option =p

user92578

Like last(USETHEPREVIOUSRESULT)?

3:34 PM

@KevinvanderVelden He's trying to linearize a bezier curve, to use it's shape as a function of X, instead of a parameterized curve of T

@Tyyppi_77 the % character does that

last(%);

user92578

thanks

user92578

How can you tell so quickly that which one of them is real?

Well this is not actually a safe way, but it's the one with no imaginary i in it :P

user92578

Ah

user92578

Oh yeah, now I see it

3:37 PM

(It's not safe and it might turn out to not be that one though, because a negative inside a radical would become imaginary, and then could cancel the i outside the radical)

user92578

So I should be able to test this now, right?

user92578

Plugging in a few test variables, and feeding it to the Cubic() again

You could do that, I prefer to graph it

user92578

Graph the whole thing or just the equation I just got

looks like there's a latex extension for chrome, converts latex between [; and ;] into images https://chrome.google.com/webstore/detail/tex-the-world-for-chromiu/mbfninnbhfepghkkcgdnmfmhhbjmhggn
like so: [;e^{\pi i} + 1 = 0;]

user92578

3:41 PM

BTW thanks for suggesting maxima, this would have taken ages by hand

kinda screwed up on a dark background though, since it assumes a white background with dark text

@IcyDefiance there is also ChatJax, which uses $

I would love if people started using ChatJax here lol

math.ucla.edu/~robjohn/math/mathjax.html

look for this link:

start ChatJax
^^drag this^^ to your bookmark bar or right click on it to add it as a bookmark.

Then if you post a formula that maxima spits out, use tex(%);

and it will convert it to latex

Kevin van der Velden

$[\frac{d+3\cdot c+3\cdot b+a}{8}]$

ooh that renders to text instead of an image, too, which means I can style it with css

Kevin van der Velden

Yay!

3:45 PM

@Tyyppi_77 where were we, plotting right? I recommend always plotting as a sanity check

cubic([0,0],[0.17, 0.67], [0.83, 0.67], [1,1], t);
plot2d([parametric, %[1], %[2], [t,0,1]]);

« first day (1820 days earlier) ← previous day next day → last day (3197 days later) »

Transcript for

Jul '1530

Game Development

Game development and other polite discussion. Game development...

game-design game-industry rubber-ducks

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile