#99381 - Goosey - Mon Aug 21, 2006 5:29 am
I am trying to convert Patater's 'orange ship' tutorial ( http://www.patatersoft.info/manual.html ) to use fixed point math as a training exercise for myself.
I finished the conversion except for the Accelleration function which had the sin/cos math functions. Attempting to convert these to use the libnds luts has been a headache. I think I am missing something simple, but can't seem to find it.
For reference:
- Angles are being stored as RADIANS in signed 20.12 fixed format
- AFAIK the SIN/COS luts are 4.12 fixed format and 512 degree increments
Here is what I have tried for indexing based on the suggestions from people in #dsdev. Below the macros is the function where they are being applied.
Any help would be greatly appreciated. At a frusterating point where I just simply don't understand where I am erring so I am randomly trying stuff in hope it will work. Not a fun place to be.
I finished the conversion except for the Accelleration function which had the sin/cos math functions. Attempting to convert these to use the libnds luts has been a headache. I think I am missing something simple, but can't seem to find it.
For reference:
- Angles are being stored as RADIANS in signed 20.12 fixed format
- AFAIK the SIN/COS luts are 4.12 fixed format and 512 degree increments
Here is what I have tried for indexing based on the suggestions from people in #dsdev. Below the macros is the function where they are being applied.
| Code: |
|
#define f32slowsin( _rad ) ( floattof32(sin(f32tofloat(_rad))) ) #define f32slowcos( _rad ) ( floattof32(cos(f32tofloat(_rad))) ) #define f32slowtan( _rad ) ( floattof32(tan(f32tofloat(_rad))) ) #define PI ( 3.14159265358979323846264338327f ) #define F32PI ( floattof32(PI) ) #define RadToDeg512Mul ( divf32(inttof32(256), F32PI) ) #define Deg512ToRadMul ( divf32(F32PI, inttof32(256)) ) #define lutIndexMul ( divf32( inttof32(512), mulf32(inttof32(2), F32PI) ) ) #define lutIndex( _rad ) ( mulf32( _rad, lutIndexMul ) ) #define radToDeg512PROPER( _rad ) ( mulf32( _rad, RadToDeg512Mul) ) #define deg512ToRadPROPER( _deg ) ( mulf32( _deg, Deg512ToRadMul) ) #define radToDeg512( _rad ) ( f32(f32tofloat(_rad) * (256/PI)) ) #define deg512ToRad( _deg ) ( f32(f32tofloat(_deg) * (PI/256)) ) // This causes the accel to be perpendicular (looking N and accel = E, deccel = W) #define f32sin( _rad ) ( COS[radToDeg512PROPER(_rad) & 0x1FF] ) #define f32cos( _rad ) ( SIN[radToDeg512PROPER(_rad) & 0x1FF] ) #define f32tan( _rad ) ( TAN[radToDeg512PROPER(_rad) & 0x1FF] ) // This functions the same as above (I don't understand why, since the non-PROPER doesn't seem.. well, proper) //#define f32sin( _rad ) ( COS[radToDeg512(_rad) & 0x1FF] ) //#define f32cos( _rad ) ( SIN[radToDeg512(_rad) & 0x1FF] ) //#define f32tan( _rad ) ( TAN[radToDeg512(_rad) & 0x1FF] ) // This also functions the same as the above two (OK NOW I AM REALLY CONFUSED! the above at least feed 512 DEGs, this feeds RADIANS!) //#define f32sin( _rad ) ( COS[_rad & 0x1FF] ) //#define f32cos( _rad ) ( SIN[_rad & 0x1FF] ) //#define f32tan( _rad ) ( TAN[_rad & 0x1FF] ) // For the love of crap.. This also seems to behave the same as the 3 above. What am I missing? -_- // #define f32sin( _rad ) ( COS[lutIndex(_rad) & 0x1FF] ) // #define f32cos( _rad ) ( SIN[lutIndex(_rad) & 0x1FF] ) // #define f32tan( _rad ) ( TAN[lutIndex(_rad) & 0x1FF] ) // These work exactly as intended, except for them being slow of course // #define f32sin( _rad ) ( f32slowsin( _rad ) ) // #define f32cos( _rad ) ( f32slowcos( _rad ) ) // #define f32tan( _rad ) ( f32slowtan( _rad ) ) void Ship::Accel(float amount) { f32 f32Amount = floattof32(amount); f32 incX = mulf32(f32Amount, f32sin(m_Angle)); f32 incY = -(mulf32(f32Amount, f32cos(m_Angle))); m_Velocity.x += incX; m_Velocity.y += incY; if (m_Velocity.x > m_MaxSpeed) m_Velocity.x = m_MaxSpeed; if (m_Velocity.x < -m_MaxSpeed) m_Velocity.x = -m_MaxSpeed; if (m_Velocity.y > m_MaxSpeed) m_Velocity.y = m_MaxSpeed; if (m_Velocity.y < -m_MaxSpeed) m_Velocity.y = -m_MaxSpeed; } |
Any help would be greatly appreciated. At a frusterating point where I just simply don't understand where I am erring so I am randomly trying stuff in hope it will work. Not a fun place to be.