|
| 18 Oct 2013 09:10 AM |
| I only have 27 kB of RAM, and each number costs 18B of RAM, my question is, should I use Rotational Matrix or Euler Angles? |
|
|
| Report Abuse |
|
|
|
| 18 Oct 2013 09:22 AM |
| 2D or 3D and what is this 18B you speak of |
|
|
| Report Abuse |
|
|
| |
|
RaidenJPN
|
  |
| Joined: 22 May 2013 |
| Total Posts: 6920 |
|
|
| 18 Oct 2013 09:30 AM |
How in the hell do you only have 27kB of remaining RAM?
Do u evn shutdown? |
|
|
| Report Abuse |
|
|
Oysi
|
  |
| Joined: 06 Jul 2009 |
| Total Posts: 9058 |
|
| |
|
|
| 18 Oct 2013 09:39 AM |
1 number = 18 bytes
wat
Anyways: -Euler angles for compact storage (cant do any useful operations on them really) -Matrices for converting between coordinate spaces (which is needed in a lot of stuff) -Quaternions for interpolating between 2 rotations and also for compact storage. I dont know if quaternions can move stuff between coordinate spaces (when it involves translation and not only rotation)
I would use quaternions over euler angles unless too lazy to deal with the math. But youll need matrices most likely.
If you do a reasonable amount of operations on the rotations it might make sense to just store matrices since if you continuously convert from quaternion to matrix it might cost too much processing power. |
|
|
| Report Abuse |
|
|
|
| 18 Oct 2013 09:40 AM |
| "27 kB of RAM," not "27 kB of remaining RAM." 27 kB is the most the system has. |
|
|
| Report Abuse |
|
|
|
| 18 Oct 2013 09:41 AM |
And you can always use some self made format as in; { yDir=(0,1,0) xDir=(1,0,0) }
and use cross product to get Z every time.
:P |
|
|
| Report Abuse |
|
|
|
| 18 Oct 2013 09:47 AM |
| Radio, I agree, aside from that, Rotational Matrix is redundant. |
|
|
| Report Abuse |
|
|
|
| 18 Oct 2013 02:45 PM |
| I'm just impressed you can script on a potato battery hooked to a monitor. |
|
|
| Report Abuse |
|
|
|
| 18 Oct 2013 02:52 PM |
MasterDesigner, it is probably a programmable SOC called an arduino :p
I would suggest using 32 bit floating bit floating point rather than 64 bit. 16 bit isn't very accurate, but it might work. I don't think arduino's hardware supports it, you would probably have to make soft-floats for 16 bit rather than hard-floats. I would suggest using roll, pitch and yaw rather than Euler Angles. It might be a bit easier to make and manipulate. |
|
|
| Report Abuse |
|
|