I was wondering, as the title mention, in an android application, what's the point to create graphics for multiple densities, while you could only have the highest density (xxxhdpi), and then the system would resize the image to the right size on the fly. It would be resource consuming if you have a lot of images, but the size of the package would be significantly smaller, and on recent phones, and by recent I mean less than 3 years, I don't think that could be an issue at all. Am I missing something here ?
Sometimes reducing a 512px icon or application graphc resource to have the same proportions on a 320px screen result in a very blurry and non defined image, specially with texts. Defining a low res Icon allow you to exactly show what you want, even not only reducing size, but cutting non relevant parts of an image.