@2x PNGs and 8bit compression - The truth

Hey guys,

I am getting ready to put out some new content, and have a few questions in regards to PNGs and corona.  My app is built to use the @2x suffix and I then have images that are half the size without that suffix.

First question:  On a lot of my half sized PNG images, their actual file size is sometimes the same as the @2x or even bigger!  How does this happen?  Also, with the file size staying about the same, am I losing the benefit for older devices seeing as the file size is no smaller?

Second question: I just came across what seems to be an awesome website called Tinypng.org.  Many of you may already use this.  On most of my PNG files it was able to compress them to be at least 50% smaller!  It says that it converts the image to a smaller 8-bit indexed image, is this still fully compatible with Corona? Anyone see any downsides to using this?

Hope I can get these questions answered and hope it can benefit many of us!

if you take an image and save it with the wrong algortihm, it can in theory get ‘bigger’ if you reduce the dimensions of the image.

for instance, I haven’t tested this in a while, but in photoshop - if you “save as” and pick png, vs. if you “save for web” and do png, the “save for web” option will be much smaller (I always save this way and use the 24-bit png if its a transparent image)

The method I used was OSX’s Automator.  I told it to resize the images to 50%. Is the benefit of smaller dimension images lost if the file size isn’t any smaller I wonder?  

if you take an image and save it with the wrong algortihm, it can in theory get ‘bigger’ if you reduce the dimensions of the image.

for instance, I haven’t tested this in a while, but in photoshop - if you “save as” and pick png, vs. if you “save for web” and do png, the “save for web” option will be much smaller (I always save this way and use the 24-bit png if its a transparent image)

The method I used was OSX’s Automator.  I told it to resize the images to 50%. Is the benefit of smaller dimension images lost if the file size isn’t any smaller I wonder?