sprite_collision_mask(ind, sepmasks, bboxmode, bbleft, bbtop, bbright, bbbottom, kind, tolerance);
Argument Description ind The index of the sprite to set the bounding box of. sepmasks Whether to create collision masks for each sub-image of the sprite (true), or one mask for all (false). bboxmode What kind of bounding box to use. 0 = automatic, 1 = full image, 2 = user defined. bbleft The maximum left position of the bounding box. bbtop The maximum top position of the bounding box. bbright The maximum right position of the bounding box. bbbottom The maximum bottom position of the bounding box. kind The kind of mask, a constant (see the table in the description). tolerance Indicates the tolerance in the transparency value (0=no tolerance, 255=full tolerance).
With this function you can set the properties of the collision mask that a sprite should have. If you select either automatic (0) or full image (1) as the bounding box mode then the individual bounding box values can be set to 0. However for a user defined mask (2) you will have to set these values. The different bounding box values are always relative to the top left corner of the sprite (irrespective of the x and y origin) which is considered position (0, 0).
Setting the kind of mask sets the general shape for the mask itself, but note that anything other than a rectangular mask will require more processing power when resolving collisions, with a subsequent drop in performance. In general, you should only use mask types other than rectangular when absolutely necessary.NOTE: This function does not permit the rotated rectangle collision mask kind.
The kind of mask that can be set will be one of the following constants:
Constant Description bboxkind_rectangular A rectangular (non-rotating) rectangle collision mask shape bboxkind_ellipse An elliptical collision mask shape bboxkind_diamond A diamond collision mask shape bboxkind_precise A precise collision mask, where the mask will conform to the non-transparent pixels of the sprite, based on the tolerance value given (see below))
Finally, tolerance is used to define how precise the precise the mask is (when used with a "full image" mask, this will have no effect), with a tolerance of 0 meaning that the mask will follow every single pixel that has a transparency over 0, while other values will shift the collision mask perimeter depending on the transparency of the pixels.NOTE: This function is only useful for bitmap sprites and will not work with SWF or JSON (Spine) sprites.NOTE: This function will only work on duplicated sprites and not directly on pre-made resources. you can duplicate sprites using the function sprite_duplicate.
spr = sprite_add("player_5.png", 16, true, true, 0,
sprite_collision_mask(spr, true, 1, 0, 0, 0, 0, 0, 0);
The above code loads a sprite from an external source and stores the new index in the variable "spr". The code then sets the new sprite to have a precise collision mask for each of its sub-images.