Answering my own question a bit maybe. Someone check my math on this? I may
have missed something...
If you go with straight default values, you have...
bias = 1.0;
shotBiasMin = -1.0;
shotBiasMax = 1.0;
Applying those values to the forumula below, you get
shotBias = ((1.0 - -1.0) * 1.0) + -1.0
= ((2.0) * 1.0) + -1.0
= 1.0
flatness = 1.0 * 0.5
= 0.5
Now - the random number generators use that flatness value, and essentially are
going to be cranking out values of X and Y that will vary from -1 to +1. And
this is where it really changes from GoldSrc. GoldSrc calculates a random
number from -0.5 to +0.5. Meaning GoldSrc has a shot pattern that's HALF THE
SIZE of Source, if I did my math right. Which means (again if I did this
right) you need to make your vecSpread be half the size in Source that it was
in GoldSrc to get the same accuracy.
The other difference is the while loop. The while look in the Source version
keeps all points inside of a circle, whereas the GoldSrc version ends up
creating a distribution of points in a square.
It's worth noting that an alternate way to calculate a random point in a circle
without the randomness would be as follows...
float x, y, r, theta;
r = RandomFloat(0, 1);
theta = RandomFloat (0, 2_PI);
x = sqrt(r) * cos(theta);
y = sqrt(r) * sin(theta);
More trig, less looping until you get a good number.
This is from http://mathworld.wolfram.com/DiskPointPicking.html
> Woops - hit send button while editing - here's the right one...
>
>
> I was looking at the base code that calculates bullet "cone of fire" or
> randomness in Source today, and comparing it with the calculation method in
> GoldSrc (HL1). They are different. I'm curious if anyone could help me
> grok the differences, and just what the effect would be for bullet spread.
>
>
> First, here's the code in shot_manipulator.h that Source uses...
>
> > if ( bias > 1.0 )
> > bias = 1.0;
> > else if ( bias < 0.0 )
> > bias = 0.0;
> >
> > float shotBiasMin = ai_shot_bias_min.GetFloat();
> > float shotBiasMax = ai_shot_bias_max.GetFloat();
> >
> > // 1.0 gaussian, 0.0 is flat, -1.0 is inverse gaussian
> > float shotBias = ( ( shotBiasMax - shotBiasMin ) * bias ) + shotBiasMin;
> >
> > float flatness = ( fabsf(shotBias) * 0.5 );
> >
> > do
> > {
> > x = random->RandomFloat(-1,1) * flatness
> > + random->RandomFloat(-1,1) * (1 - flatness);
> > y = random->RandomFloat(-1,1) * flatness
> > + random->RandomFloat(-1,1) * (1 - flatness);
> > if ( shotBias < 0 )
> > {
> > x = ( x >= 0 ) ? 1.0 - x : -1.0 - x;
> > y = ( y >= 0 ) ? 1.0 - y : -1.0 - y;
> > }
> > z = x*x+y*y;
> > } while (z > 1);
> >
> > m_vecResult = m_vecShotDirection
> > + x * vecSpread.x * m_vecRight
> > + y * vecSpread.y * m_vecUp;
>
>
> Here is the (I think) equivalent code that GoldSrc (HL1) uses...
>
> > x = UTIL_SharedRandomFloat( shared_rand + iShot, -0.5, 0.5 )
> > + UTIL_SharedRandomFloat( shared_rand + ( 1 + iShot ) , -0.5, 0.5 );
> > y = UTIL_SharedRandomFloat( shared_rand + ( 2 + iShot ), -0.5, 0.5 )
> > + UTIL_SharedRandomFloat( shared_rand + ( 3 + iShot ), -0.5, 0.5 );
> > z = x * x + y * y;
> >
> > Vector vecDir = vecDirShooting
> > + x * vecSpread.x * vecRight
> > + y * vecSpread.y * vecUp;
>
>
> So you can see how the X and Y are calculated differently in the two. Source
> has a "bias" value that influences how it works. Source also loops until you
> get an X and Y that lay inside of a circle (that while loop on Z) but GoldSrc
> doesn't.
>
> So basically, what will the difference be between these two systems? What
> will
> the bias value's influece be on the randomness?
>
> I actually started looking into this because I want to code a weighted
> randomness - with more points concentrated in the middle. Take a look at
> http://mathworld.wolfram.com/DiskPointPicking.html for an article about how
> NOT
> to do this (ironically).
>
>
> _______________________________________________
> To unsubscribe, edit your list preferences, or view the list archives, please
> visit:
> http://list.valvesoftware.com/mailman/listinfo/hlcoders
>
>
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please
visit:
http://list.valvesoftware.com/mailman/listinfo/hlcoders
