image archive VS image strip

Question

Hi, i've noticed that plenty of games / applications (very common on mobile builds) pack numerous images into an image strip. I figured that the advantages in this are making the program more tidy (file system - wise) and reducing (un)installation time. During the runtime of the application, the entire image strip is allocated and copied from FS to RAM. On the contrary, images can be stored in an image archive and unpacked during runtime to a number of image structures in RAM.

The way I see it, the image strip approach is less efficient because of worse caching performance and because that even if the optimal rectangle packing algorithm is used, there will be empty spaces between the stored images in the strip, causing a waste of RAM.

What are the advantages in using an image strip over using an image archive file?

Answer 1

Caching will not be as much of an issue as you think, if the images in the strip are likely to appear at the same time - i.e. are all part of an animation. As far as packing algorithms are concerned, most generic algorithms produce similar levels of efficiency - to get huge increases in packing efficiency you normally have to write an algorithm that is optimised for the dataset involved.

As is usually the case with optimization of any kind - caching or packing - don't worry about it until it's been proved to be an issue, and then you need to approach the problem with a good understanding of the context.