how can I get nonlinear distance from an object with physics.

Author

Message

brunuu

5

Years of Service

User Offline

Joined: 4th Jun 2018

Location:

Posted: 4th Mar 2020 18:57

Link

Normal physical formulas will not work very well here due to the physics setup. because of the command SetPhysicsScale
setPhysicsGravity(0, 100)
SetPhysicsScale(0.01)

the objective is to obtain the distance covered by an object that has free movement.

like this

Back to top

Profile PM

Mesher

User Banned

Posted: 4th Mar 2020 19:17

Link

To solve for distance use the formula for distance d = st, or distance equals speed times time.

distance = speed x time

Rate and speed are similar since they both represent some distance per unit time like miles per hour or kilometers per hour. If rate r is the same as speed s, r = s = d/t. You can use the equivalent formula d = rt which means distance equals rate times time.

distance = rate x time

To solve for speed or rate use the formula for speed, s = d/t which means speed equals distance divided by time.

speed = distance/time

To solve for time use the formula for time, t = d/s which means time equals distance divided by speed.

time = distance/speed

Good luck

Back to top

Profile

Virtual Nomad

Moderator

18

Years of Service

User Offline

Joined: 14th Dec 2005

Location: SF Bay Area, USA

Posted: 4th Mar 2020 22:45 Edited at: 4th Mar 2020 22:52

Link

some pseudocode:

+ Code Snippet

do
...
   distance# = distance# + lastmove()
...
loop

Function lastmove()
   lastmove# = currentxy# - lastxy# `actual/normal 2d distance formula here
   lastxy# = currentxy#
EndFunction LastMove#

thats about the only way you can do it, i believe. and, you'll need some tweaking to find an acceptable "accuracy".

not sure that what mesher said applies to your inquiry but good reference, otherwise.

My Itch.io Home.

Back to top

Profile PM Website

nz0

AGK Developer

16

Years of Service

User Offline

Joined: 13th Jun 2007

Location: Cheshire,UK

Posted: 4th Mar 2020 22:56

Link

A simple way would be to do a direct distance calculation accumulator every frame between last position and current position.

(Just read Virtual Nomad's post and we appear to agree

)

You could optimise it slightly by checking if the new position=old position and skip the distance formula.

My Robotron remake
My I,Robot remake WIP
My Defender remake
AGK WadPacker - Protect your Media!

Back to top

Profile PM Email Website

Virtual Nomad

Moderator

18

Years of Service

User Offline

Joined: 14th Dec 2005

Location: SF Bay Area, USA

Posted: 4th Mar 2020 23:37 Edited at: 5th Mar 2020 00:09

Link

worth adding since i don't know what the project might also entail but if there is any collision involved, simple currentxy# - oldxy# will be misleading/potentially (and thoroughly) inaccurate.

consider a sprite with a velocity of 1000 pointed at a wall 500 units away. it could run into the wall and come back to where it started which would give you a LastMove# of 0.0 ; see this extreme example of what i mean.

if this kind of thing might affect you, you'd be looking at tracking each contact point, applying sprite offsets (or otherwise account for shapes involved), and the distance formula to each"step" of movement in a given frame.

btw, when you venture into the velocities anywhere near those in the vid, the "system" gets "iffy".

IE, i'm checking for contact between the ball and the paddle. if it hits the bottom wall, the ball goes red (which you see plenty of) while i've confirmed that it has actually hit the paddle (where the ball should go green, and doesn't) much of the time (i've got that on auto-pilot, obviously, which leaves only small gaps at the edges for the ball to hit the bottom).

so, it reacts properly (physics-wise) but the contact data retrieved can't be trusted.

add: make sure to review this since you've been focused on box2d lately, if you haven't already. i found it somewhat insightful where i went in completely blind, otherwise

My Itch.io Home.

Back to top

Profile PM Website

nz0

AGK Developer

16

Years of Service

User Offline

Joined: 13th Jun 2007

Location: Cheshire,UK

Posted: 5th Mar 2020 00:09

Link

Quote: "consider a sprite with a velocity of 1000 pointed at a wall 500 units away. it could run into the wall and come back to where it started which would give you a LastMove# of 0.0 ; see this extreme example of what i mean.
"

I see what you mean.
Depends what the OP exactly wants and if any collisions are involved.

I wouldn't trust using data you send to the physics engine to match what you expect to see on the screen. "Iffy" is not a strong enough word to describe half of the things I've seen happen with it

My Robotron remake
My I,Robot remake WIP
My Defender remake
AGK WadPacker - Protect your Media!

Back to top

Profile PM Email Website

brunuu

5

Years of Service

User Offline

Joined: 4th Jun 2018

Location:

Posted: 5th Mar 2020 00:25 Edited at: 5th Mar 2020 00:26

Link

yes there are collisions, that's why I was trying to use physics formulas

Back to top

Profile PM

Virtual Nomad

Moderator

18

Years of Service

User Offline

Joined: 14th Dec 2005

Location: SF Bay Area, USA

Posted: 5th Mar 2020 00:51

Link

then you're probably on the same road i am with the g-breaker project which is, given the AppGameKit implementation of box2d, venturing outside of its intended use.

in that project, i've abused what's available and getting to the point where i should write my own physics from scratch to account for what i'm trying to squeeze out of box2d (i'm adjusting velocities manually, each frame, barely hanging on to a "natural" feeling, for example).

but, if you're intent on pushing forward, consider some pre-calculations (my next step, if i continue with g-breaker/am up for the challenge) which would probably include time(s) of impact and 2d rays to help but, then, borders on the edge of the "do it manually/from scratch" notion.

if it's not more much that what you've depicted in the last image (balls vs walls), it shouldn't be too terribly hard. i just haven't done it personally.

regardless, good luck

My Itch.io Home.

Back to top

Profile PM Website

blink0k

Moderator

11

Years of Service

Recently Online

Joined: 22nd Feb 2013

Location: the land of oz

Posted: 5th Mar 2020 01:15

Link

Quote: "i'm adjusting velocities manually, each frame"

Are you using linear velocity or impulse?

Back to top

Profile PM

Virtual Nomad

Moderator

18

Years of Service

User Offline

Joined: 14th Dec 2005

Location: SF Bay Area, USA

Posted: 5th Mar 2020 02:28

Link

don't want to hijack this thread. check ur PM

My Itch.io Home.

Back to top

Profile PM Website

brunuu

5

Years of Service

User Offline

Joined: 4th Jun 2018

Location:

Posted: 5th Mar 2020 02:41

Link

well, i did it, i think it works decent.

the distance traveled by the object starts to count when the mouse is released.

SetErrorMode(2)
SetWindowTitle("Distance")
SetWindowSize(1280, 720, 0)
SetWindowAllowResize(1) 
SetVirtualResolution(1280, 720) 
SetOrientationAllowed(1, 1, 1, 1) 
SetSyncRate(60, 0) 
SetScissor(0,0,0,0)
UseNewDefaultFonts(1)

setPhysicsGravity(0, 100)
SetPhysicsScale(0.01)
SetPhysicsDebugOn()

global spr as integer
spr = CreateSprite(0)
SetSpritePosition(spr, 150, 150)
SetSpriteSize(spr, 50, 50)
SetSpriteOffset(spr, GetSpriteWidth(spr)/2, GetSpriteHeight(spr)/2)
SetSpritePhysicsOn(spr, 2)
SetSpritePhysicsRestitution(spr, 0.60)

global distance as float
global px as float 
global py as float

do
	print(distance)
	print(RateApp
	moveSpr()
	Getdistance()
    Sync()
loop

function moveSpr()
	if GetPointerstate() = 1 
		obj = getspritehit(GetPointerX(),GetPointerY())
        if obj = spr and not getjointexists(1)
            CreateMouseJoint(1, spr, GetPointerX(), GetPointerY(), 100)
            distance = 0
        endif
        if getjointexists(1)
            SetJointMouseTarget(1, GetPointerX(), GetPointerY())
        endif
    else
        if getjointexists(1) 
			DeleteJoint(1)
			px = GetSpriteXByOffset(spr)
			py = GetSpriteYByOffset(spr)
			distance = 0
        endif
    endif
endfunction

function Getdistance()
	distance = distance + Sqrt(pow(GetSpriteXByOffset(spr) - px, 2) + pow(GetSpriteYByOffset(spr) - py, 2))
	px = GetSpriteXByOffset(spr)
	py = GetSpriteYByOffset(spr)
endfunction

+ Code Snippet

SetErrorMode(2)
SetWindowTitle("Distance")
SetWindowSize(1280, 720, 0)
SetWindowAllowResize(1) 
SetVirtualResolution(1280, 720) 
SetOrientationAllowed(1, 1, 1, 1) 
SetSyncRate(60, 0) 
SetScissor(0,0,0,0)
UseNewDefaultFonts(1) 

setPhysicsGravity(0, 100)
SetPhysicsScale(0.01)
SetPhysicsDebugOn()

global spr as integer
spr = CreateSprite(0)
SetSpritePosition(spr, 150, 150)
SetSpriteSize(spr, 50, 50)
SetSpriteOffset(spr, GetSpriteWidth(spr)/2, GetSpriteHeight(spr)/2)
SetSpritePhysicsOn(spr, 2)
SetSpritePhysicsRestitution(spr, 0.60)

global distance as float
global px as float 
global py as float 

do
	print(distance)
	print(RateApp
	moveSpr()
	Getdistance()
    Sync()
loop

function moveSpr()
	if GetPointerstate() = 1 
		obj = getspritehit(GetPointerX(),GetPointerY())
        if obj = spr and not getjointexists(1)
            CreateMouseJoint(1, spr, GetPointerX(), GetPointerY(), 100)
            distance = 0
        endif
        if getjointexists(1)
            SetJointMouseTarget(1, GetPointerX(), GetPointerY())
        endif
    else
        if getjointexists(1) 
			DeleteJoint(1)
			px = GetSpriteXByOffset(spr)
			py = GetSpriteYByOffset(spr)
			distance = 0
        endif
    endif
endfunction

function Getdistance()
	distance = distance + Sqrt(pow(GetSpriteXByOffset(spr) - px, 2) + pow(GetSpriteYByOffset(spr) - py, 2))
	px = GetSpriteXByOffset(spr)
	py = GetSpriteYByOffset(spr)
endfunction

Back to top

Profile PM

Mesher

User Banned

Posted: 5th Mar 2020 23:41

Link

Well done

Back to top

Profile

blink0k

Moderator

11

Years of Service

Recently Online

Joined: 22nd Feb 2013

Location: the land of oz

Posted: 6th Mar 2020 02:29

Link

An interesting exercise might be to save those points in an array and use DrawLine() to see how accurate it is

Back to top

Profile PM

Sorry your browser is not supported!

AppGameKit Classic Chat / how can I get nonlinear distance from an object with physics.