Hi Kennant,
I think I can answer your second question.
I don't know what 'setp' is but I can explain about roi usage.
Basically, you can give an IPP function a pointer to the start of your data. When there is no ROI, this would usually be the (0,0)-origin of your image, or the address of the first element of your data array.
Since, IPP assumes that the data is contigous it must be told how to treat this data as a 2D array (i.e. image), hence it must get the width of the image, so it can distinguish between the image rows.
The width is given through the step arguments. This argument says how many array elements there are from the current one to the one right belowit on the next row.
Using this scheme, it is easy to define ROIs, by setting the source pointer to the ROI starting pixel. The image step width, will always bring you to the pixel right below it on the next row.
The roiSize argument tells the IPP function the actual size of the roi.
Note: In my description I assumed a top-left image.
The step width argument is a very useful and flexible representation of the width, since it can naturally allow for image right-padding (e.g.for better MMX performance or display) or for working of interlaced fields.
I hope this helps,
Adi
