I’m quite new to Android development. My understanding is that you can create several versions of the same image with different sizes and put them into the folders drawable-ldpi, drawable-mdpi, drawable-hdpi.
It seems obvious to me that you can handle this problem “the lazy way” by just resizing one image depending on the device’s pixel density. For this I programmatically find out what density the device has, like ldpi. The implementation itself is not the problem. I’m just afraid of any drawbacks (that prevent me later from running the app on different devices).
So, are there any (major) drawbacks of scaling images automatically ?
In which of the three folders do I put the image so that the compiler can find it?
You would put the image in your regular drawable folder. That way any phone can find it.
While you can programatically shrink images, shrinking usually has the effect of reducing image detail and causing jaggies.
Adding in smaller assets will also reduce memory usage on smaller phones. Keep in mind that some Android phones are notoriously bad with memory (see: HTC Status), so any and all savings help.
I would recommend just photoshop scaling images down large images yourself. For smaller images, it is not as big a deal.