R magick: compose images without vectorization

Question

The description of the image_composite function in the magick R package states:

The image_composite function is vectorized over both image arguments: if the first image has n frames and the second m frames, the output image will contain n * m frames.

Here an ugly example:

banana <- image_read("https://jeroen.github.io/images/banana.gif")
image_composite(banana, banana, offset = "+70+00", operator = "Add")

Is there a way to avoid vectorisation so that the bananas can dance together? Alternatively, is there another function (also from other packages) that allows for this?

Answer 1

In the case where it's the same image, you can just modify each frame, turning the vectorization inside-out, sort of.

image_apply(
  banana, FUN = function(img) {
    image_composite(img, img, offset = "+70+00", operator = "Add")
  }
)

If you have two images of the same number of frames, that's harder using only the exposed functions in magick. An easier way is to just separate the magick-image objects into lists, and use functional programming tools like purrr::map2, then rejoin them afterwards:

purrr::map2(
  as.list(banana), as.list(image_negate(banana)),
  ~image_composite(.x, .y, operator = "Add", offset = "+70+00")
  ) %>% 
  image_join()