So JT and I spent today making a new testing file that is specific to scalability testing. First we ditched JPGs, and 32 bit BMPs because they were having problems and were just overall not viable formats. Then we made versions of every remaining image format at 4 different resolutions, 1 set at the original resolution, 1 set at a resolution that was 2 times as large, 1 set at a resolution that was 4 times as large, and finally 1 set at a resolution that was 1.73 times as large so that if the functions we were using to scale the images had built in optimization techniques, that image would still give it a "run for its money", so to speak. After we had all of the new images we created a new version of the testing program that allowed us to scale all of the images to any size we wanted. We then ran the test and scaled every image to the original image size.

The goal with this test was to determine if scaling images through the code is a viable option. Ideally, if it turned out that scaling the images through the code did not cause a major slowdown, software creators would have the ability to create one image set at the highest resolution they needed, and then just scale the images down to accommodate lower resolutions. This would be helpful because it would drastically decrease hard-drive capacity needed for the game. The only negative that would come about from this situation would be that the images would lose detail when being scaled down. In theory though, this would not be that big of a deal and would be a viable sacrifice considering the upsides.

We are running the tests now and I will update the blog when we have our results.