Are there any downsides, or drawbacks to this? I understand that higher order would be difficult to control, but maybe in terms of performance or compute, is this approach at all worse? Or is the only real drawback that it's more user-intensive to have to piece together cubics manually as opposed to having 1 function?
marilynjoyce
Wouldn't this approach take more computation time/power? Because imagine a really high-order Bezier curve, then you would have to split it up so many times into the lower-order curves. Is there a more efficient approach? And how are those lower-order curves even decided? Is it a bound of order 4?
srikartalluri
@marilynjoyce That is an interesting point to make. Similar to how KD trees work by splitting a large problem into multiple subproblems, there is usually a stop condition at which the curve will not be split anymore. For example, it might continue splitting until it meets the flatness criterion or it reaches a maximum recursion depth
Are there any downsides, or drawbacks to this? I understand that higher order would be difficult to control, but maybe in terms of performance or compute, is this approach at all worse? Or is the only real drawback that it's more user-intensive to have to piece together cubics manually as opposed to having 1 function?
Wouldn't this approach take more computation time/power? Because imagine a really high-order Bezier curve, then you would have to split it up so many times into the lower-order curves. Is there a more efficient approach? And how are those lower-order curves even decided? Is it a bound of order 4?
@marilynjoyce That is an interesting point to make. Similar to how KD trees work by splitting a large problem into multiple subproblems, there is usually a stop condition at which the curve will not be split anymore. For example, it might continue splitting until it meets the flatness criterion or it reaches a maximum recursion depth