czwartek, 3 sierpnia 2023 o 19:17:11 UTC+2 Bonita Montero napisał(a):
> Am 19.07.2023 um 08:38 schrieb fir:
> > i need tos et asteroids in game (about 100 or 200) but not in square
> > area with two rands one for x and second for y but need it to
> > be sphericel symetric around central point and in more densitu near it and less desity with bigger radius until some radius when it is zero
> >
> > something like
> >
> > SetRandomPointSphericalSymetricAndFadingWithDistance(
> > float x, float y, float r1, float r2);
> >
> > this r1 is radius of circle below which density is max and r2 is radius
> > where above density is zero - in beetweem r1 and r2 density should
> > be from max (at r1) to zero at r2
> >
> > how to do that?
> In C++ I'd do that this way:
>
> #include <random>
> #include <cmath>
>
> using namespace std;
>
> double nearDistance( double d )
> {
> thread_local double exp = 2.0;
> thread_local mt19937_64 mt;
> thread_local uniform_real_distribution<double> rnd1( 0.0, 1.0 );
> if( d < 0.0 ) [[unlikely]]
> {
> double beforeExp = exp;
> exp = -d;
> return beforeExp;
> }
> return pow( rnd1( mt ), exp ) * d;
> }
>
> This gives values from 0.0 to d which are more likely to be near
> to 0.0 than with a linear exponent. The exponent can be changed
> and should be > 1.0.

dont know what it is, howewer thinkong of what i wrote
if
x = rand2(0,a); gives 'rectangle' density

//rand2(a,b) gives random from a to b)

then there i should probably transform helf of the points (as rectangle is two triangles) into triangle
but how? should i maybe take x/a which is from 0 to 1 and take it as a probablility of 'flipping'
the value from right-up triangl;e to left-down?
int GiveRandomVauleWithLinearlyDecayingDensity(int A, int B)
{ x = rand2(A,B);
float flip_probability = 1.*(x-A)/(B-A);
if(chance(flip_probability)) x = A + (B - x);
}

this just flips the x values from more right area of A to B span based on how right vaule this is

chance would be a function returning tru with a chance of flip_probability

dont know but maybe this is what i need

czwartek, 3 sierpnia 2023 o 19:55:44 UTC+2 fir napisał(a):
> czwartek, 3 sierpnia 2023 o 19:17:11 UTC+2 Bonita Montero napisał(a):
> > Am 19.07.2023 um 08:38 schrieb fir:
> > > i need tos et asteroids in game (about 100 or 200) but not in square
> > > area with two rands one for x and second for y but need it to
> > > be sphericel symetric around central point and in more densitu near it and less desity with bigger radius until some radius when it is zero
> > >
> > > something like
> > >
> > > SetRandomPointSphericalSymetricAndFadingWithDistance(
> > > float x, float y, float r1, float r2);
> > >
> > > this r1 is radius of circle below which density is max and r2 is radius
> > > where above density is zero - in beetweem r1 and r2 density should
> > > be from max (at r1) to zero at r2
> > >
> > > how to do that?
> > In C++ I'd do that this way:
> >
> > #include <random>
> > #include <cmath>
> >
> > using namespace std;
> >
> > double nearDistance( double d )
> > {
> > thread_local double exp = 2.0;
> > thread_local mt19937_64 mt;
> > thread_local uniform_real_distribution<double> rnd1( 0.0, 1.0 );
> > if( d < 0.0 ) [[unlikely]]
> > {
> > double beforeExp = exp;
> > exp = -d;
> > return beforeExp;
> > }
> > return pow( rnd1( mt ), exp ) * d;
> > }
> >
> > This gives values from 0.0 to d which are more likely to be near
> > to 0.0 than with a linear exponent. The exponent can be changed
> > and should be > 1.0.
> dont know what it is, howewer thinkong of what i wrote
> if
> x = rand2(0,a); gives 'rectangle' density
>
> //rand2(a,b) gives random from a to b)
>
> then there i should probably transform helf of the points (as rectangle is two triangles) into triangle
> but how? should i maybe take x/a which is from 0 to 1 and take it as a probablility of 'flipping'
> the value from right-up triangl;e to left-down?
>
> int GiveRandomVauleWithLinearlyDecayingDensity(int A, int B)
> {
> x = rand2(A,B);
> float flip_probability = 1.*(x-A)/(B-A);
> if(chance(flip_probability)) x = A + (B - x);
> }
>
> this just flips the x values from more right area of A to B span based on how right vaule this is
>
> chance would be a function returning tru with a chance of flip_probability
>
> dont know but maybe this is what i need

btw how thuis chance function code shoule be?

int chance(float p)
{ return p>= float(rand()&0x7fff)/0x7fff;;
}

?

Am 03.08.2023 um 19:55 schrieb fir:

> dont know what it is, ...

Im taking a random double between 0.0 and 1.0, exponentiate it with
an exponent > 1.0, thereby having more small output values than high
and then I multiply that with the maximum value supplied. This makes
the desired distribution where lower values are more likely as you
increase the exponent.

piątek, 4 sierpnia 2023 o 17:22:30 UTC+2 Bonita Montero napisał(a):
> Am 03.08.2023 um 19:55 schrieb fir:
>
> > dont know what it is, ...
>
> Im taking a random double between 0.0 and 1.0, exponentiate it with
> an exponent > 1.0, thereby having more small output values than high
> and then I multiply that with the maximum value supplied. This makes
> the desired distribution where lower values are more likely as you
> increase the exponent.

probably should work but i wrote this laternative with fliping...

on the right side the probablility of flipping to left (like 0.8 to 0.2, or 0.9 to 0.1) is
rising to 100% where on left it is rising form 0% so this probably could turn "rectangle" distribution into triangle... but i dont need this code for hurry so now im not plan to test it - will test it when i will need it

fir <profesor.fir@gmail.com> wrote:
> i need to set asteroids in game (about 100 or 200) but not in square
> area with two rands one for x and second for y but need it to
> be sphericel symetric around central point and in more density near it
> and less desity with bigger radius until some radius when it is zero
>
> something like
>
> SetRandomPointSphericalSymetricAndFadingWithDistance(
> float x, float y, float r1, float r2);
>
> this r1 is radius of circle below which density is max and r2 is radius
> where above density is zero - in beetweem r1 and r2 density should
> be from max (at r1) to zero at r2
>
> how to do that?

Haven't read the followups. You mean something like...
https://www.deviantart.com/eigenheit22/art/Starfield-v1-860092004
....without the animation? That's done in x,y (and z). And you can then
just calculate r=sqrt(x*x+y*y) and discard the ones you don't like.
--
John Forkosh ( mailto: j@f.com where j=john and f=forkosh )

sobota, 5 sierpnia 2023 o 12:57:46 UTC+2 John Forkosh napisał(a):
> fir <profes...@gmail.com> wrote:
> > i need to set asteroids in game (about 100 or 200) but not in square
> > area with two rands one for x and second for y but need it to
> > be sphericel symetric around central point and in more density near it
> > and less desity with bigger radius until some radius when it is zero
> >
> > something like
> >
> > SetRandomPointSphericalSymetricAndFadingWithDistance(
> > float x, float y, float r1, float r2);
> >
> > this r1 is radius of circle below which density is max and r2 is radius
> > where above density is zero - in beetweem r1 and r2 density should
> > be from max (at r1) to zero at r2
> >
> > how to do that?
> Haven't read the followups. You mean something like...
> https://www.deviantart.com/eigenheit22/art/Starfield-v1-860092004
> ...without the animation? That's done in x,y (and z). And you can then
> just calculate r=sqrt(x*x+y*y) and discard the ones you don't like.

possibly - now i think you just can imagine field of discarding (from 0.0 to 1.0 chance of discarding in given point) then just calculate chnce of discarding in that point according to a field and use it

thsi is a bit wastefull to count something than discard it but probably could be enough ok

On 7/18/2023 11:38 PM, fir wrote:
> i need tos et asteroids in game (about 100 or 200) but not in square
> area with two rands one for x and second for y but need it to
> be sphericel symetric around central point and in more densitu near it and less desity with bigger radius until some radius when it is zero
>
> something like
>
> SetRandomPointSphericalSymetricAndFadingWithDistance(
> float x, float y, float r1, float r2);
>
> this r1 is radius of circle below which density is max and r2 is radius
> where above density is zero - in beetweem r1 and r2 density should
> be from max (at r1) to zero at r2
>
> how to do that?

Fwiw, check out this older simulation I made around 9 years ago:

https://youtu.be/-EaHLaZbMFs

It tries to diffuse itself...

fir <profesor.fir@gmail.com> wrote:
> John Forkosh wrote(a):
>> fir <profes...@gmail.com> wrote:
>> > i need to set asteroids in game (about 100 or 200) but not in square
>> > area with two rands one for x and second for y but need it to
>> > be sphericel symetric around central point and in more density near it
>> > and less desity with bigger radius until some radius when it is zero
>> >
>> > something like
>> > SetRandomPointSphericalSymetricAndFadingWithDistance(
>> > float x, float y, float r1, float r2);
>> > this r1 is radius of circle below which density is max and r2 is radius
>> > where above density is zero - in beetweem r1 and r2 density should
>> > be from max (at r1) to zero at r2
>> > how to do that?
>>
>> Haven't read the followups. You mean something like...
>> https://www.deviantart.com/eigenheit22/art/Starfield-v1-860092004
>> ...without the animation? That's done in x,y (and z). And you can then
>> just calculate r=sqrt(x*x+y*y) and discard the ones you don't like.
>
> possibly - now i think you just can imagine field of discarding
> (from 0.0 to 1.0 chance of discarding in given point) then just
> calculate chnce of discarding in that point according to a field
> and use it
>
> this is a bit wasteful to count something than discard it but
> probably could be enough ok

You said 100-200 asteroids, so "wasteful" is hardly a concern.
But if you're concerned anyway, suppose you wanted asteroids
at radius r1 <= r <= r2 (you wanted "max density" at r<=r1,
whereas this will give 0 at r<r1 and r>r2; your exercise is to
figure out the necessary modification). So go back to r,theta
polar coords and just do something along the following lines...
#include <stdlib.h>
#include <math.h>
void SetRandomPoint( double r1, double r2, double *x, double *y ) {
double r = r1 + (r2-r1)*drand48(), /* random r1<=r<r2 */
theta = 2.0*3.14159*drand48(); /* random 0<=theta<2pi */
*x = r*cos(theta); *y = r*sin(theta); /* return random x,y coords */
return; }
And the density will naturally decrease with increasing r.
No work required.
--
John Forkosh ( mailto: j@f.com where j=john and f=forkosh )

poniedziałek, 7 sierpnia 2023 o 15:13:37 UTC+2 John Forkosh napisał(a):
> fir <profes...@gmail.com> wrote:
> > John Forkosh wrote(a):
> >> fir <profes...@gmail.com> wrote:
> >> > i need to set asteroids in game (about 100 or 200) but not in square
> >> > area with two rands one for x and second for y but need it to
> >> > be sphericel symetric around central point and in more density near it
> >> > and less desity with bigger radius until some radius when it is zero
> >> >
> >> > something like
> >> > SetRandomPointSphericalSymetricAndFadingWithDistance(
> >> > float x, float y, float r1, float r2);
> >> > this r1 is radius of circle below which density is max and r2 is radius
> >> > where above density is zero - in beetweem r1 and r2 density should
> >> > be from max (at r1) to zero at r2
> >> > how to do that?
> >>
> >> Haven't read the followups. You mean something like...
> >> https://www.deviantart.com/eigenheit22/art/Starfield-v1-860092004
> >> ...without the animation? That's done in x,y (and z). And you can then
> >> just calculate r=sqrt(x*x+y*y) and discard the ones you don't like.
> >
> > possibly - now i think you just can imagine field of discarding
> > (from 0.0 to 1.0 chance of discarding in given point) then just
> > calculate chnce of discarding in that point according to a field
> > and use it
> >
> > this is a bit wasteful to count something than discard it but
> > probably could be enough ok
> You said 100-200 asteroids, so "wasteful" is hardly a concern.
> But if you're concerned anyway, suppose you wanted asteroids
> at radius r1 <= r <= r2 (you wanted "max density" at r<=r1,
> whereas this will give 0 at r<r1 and r>r2; your exercise is to
> figure out the necessary modification). So go back to r,theta
> polar coords and just do something along the following lines...
> #include <stdlib.h>
> #include <math.h>
> void SetRandomPoint( double r1, double r2, double *x, double *y ) {
> double r = r1 + (r2-r1)*drand48(), /* random r1<=r<r2 */
> theta = 2.0*3.14159*drand48(); /* random 0<=theta<2pi */
> *x = r*cos(theta); *y = r*sin(theta); /* return random x,y coords */
> return; }
> And the density will naturally decrease with increasing r.
> No work required.

i wanted max constand density say in radius 100 then linerally decaying say
in radius from 300

it seems the rand only givesflat density, then probably proper way fould be to tranmsform/project this square cordinates into other shape as density is proportional to 'surface'if some transforms the surface stretching it or densing by some x,y -> x', y' then it will have les random density

but i dont remember how to do transformations ([probably they are normal functions of x,y into new x,y so some could be found) but there is also thsi way with making fields of percentage of discarding

this function is handy to me but its also not any way critical i need it jsu to set things in agme liak asteroid fields, space flowers or something liek that (but dont hope it would be look god i only write games wth symbolic graphics, something like kinda roguelikes)

poniedziałek, 7 sierpnia 2023 o 15:26:44 UTC+2 fir napisał(a):
> poniedziałek, 7 sierpnia 2023 o 15:13:37 UTC+2 John Forkosh napisał(a):
> > fir <profes...@gmail.com> wrote:
> > > John Forkosh wrote(a):
> > >> fir <profes...@gmail.com> wrote:
> > >> > i need to set asteroids in game (about 100 or 200) but not in square
> > >> > area with two rands one for x and second for y but need it to
> > >> > be sphericel symetric around central point and in more density near it
> > >> > and less desity with bigger radius until some radius when it is zero
> > >> >
> > >> > something like
> > >> > SetRandomPointSphericalSymetricAndFadingWithDistance(
> > >> > float x, float y, float r1, float r2);
> > >> > this r1 is radius of circle below which density is max and r2 is radius
> > >> > where above density is zero - in beetweem r1 and r2 density should
> > >> > be from max (at r1) to zero at r2
> > >> > how to do that?
> > >>
> > >> Haven't read the followups. You mean something like...
> > >> https://www.deviantart.com/eigenheit22/art/Starfield-v1-860092004
> > >> ...without the animation? That's done in x,y (and z). And you can then
> > >> just calculate r=sqrt(x*x+y*y) and discard the ones you don't like..
> > >
> > > possibly - now i think you just can imagine field of discarding
> > > (from 0.0 to 1.0 chance of discarding in given point) then just
> > > calculate chnce of discarding in that point according to a field
> > > and use it
> > >
> > > this is a bit wasteful to count something than discard it but
> > > probably could be enough ok
> > You said 100-200 asteroids, so "wasteful" is hardly a concern.
> > But if you're concerned anyway, suppose you wanted asteroids
> > at radius r1 <= r <= r2 (you wanted "max density" at r<=r1,
> > whereas this will give 0 at r<r1 and r>r2; your exercise is to
> > figure out the necessary modification). So go back to r,theta
> > polar coords and just do something along the following lines...
> > #include <stdlib.h>
> > #include <math.h>
> > void SetRandomPoint( double r1, double r2, double *x, double *y ) {
> > double r = r1 + (r2-r1)*drand48(), /* random r1<=r<r2 */
> > theta = 2.0*3.14159*drand48(); /* random 0<=theta<2pi */
> > *x = r*cos(theta); *y = r*sin(theta); /* return random x,y coords */
> > return; }
> > And the density will naturally decrease with increasing r.
> > No work required.
> i wanted max constand density say in radius 100 then linerally decaying say
> in radius from 300
>
> it seems the rand only givesflat density, then probably proper way fould be to tranmsform/project this square cordinates into other shape as density is proportional to 'surface'if some transforms the surface stretching it or densing by some x,y -> x', y' then it will have les random density
>

this waht you say is in fact such trabsformation x,x->x',y' buts its kinda wrong
becouse near radius zero it has no limited densityand in radius 0.1 you got 10 percent
of all asteroids which simply might be to high..so i gusee there is soem another transformation which coudl be found better..though still its probably a bit harder to write it nly by transformation without discarding

> but i dont remember how to do transformations ([probably they are normal functions of x,y into new x,y so some could be found) but there is also thsi way with making fields of percentage of discarding
>
> this function is handy to me but its also not any way critical i need it jsu to set things in agme liak asteroid fields, space flowers or something liek that (but dont hope it would be look god i only write games wth symbolic graphics, something like kinda roguelikes)