[jmarshall]

Hi all,

Those following the SVN logs will have noticed that we now support compressed DXT1, 3 and 5 textures in the skin engine. This opens the door for the possibility of storing fanart and thumbnails in these compressed formats rather than in JPG.

The main advantage here is memory usage and loading time. DXT1 is a 1:8 compression ratio (4 bits per pixel) in terms of the amount of data kept in GPU memory, as the GPU decodes the image as it's rendering it. Thus, no CPU decode is needed at all, compared with JPG which needs CPU decode, then a full 32 bits per pixel texture. For a 1080p backdrop, this is 8MB versus 1MB - a large saving. It means more textures can be kept in memory longer further reducing load time.

The disadvantage is quality: DXT1 may not be as good as JPG, and some images may suffer from this more than others. This is what this thread is for - determining whether or not the quality degradation is:
1. Noticeable.
2. Acceptable.

To do this, I've added support for loading .dds versions of fanart in SVN via the <useddsfanart> advancedsettings.xml tag. You need SVN r23421 or higher for this

Code:

<advancedsettings>
  <useddsfanart>true</useddsfanart>
</advancedsettings>

Secondly, you need your fanart converted to .dds. If you're on windows, you can use the batch file and script in the attached .zip file. Simply place the MakeDDS.exe and Convert.bat script in your fanart folders (UserData/Thumbnails/Video/Fanart and UserData/Thumbnails/Music/Fanart) and run the convert.bat batch file. with convert.bat *.tbn It'll take a few seconds and 1-2MB per image. NOTE: You'll need the SDL libraries (SDL.dll, SDL_image.dll) in your path - you can place them in windows/system32 or next to MakeDDS. Grab them from the SDL website if you don't have them.

If you're on linux or OS X you'll have to wait someone to do you up a build - the source code is in tools/TexturePacker in SVN.

If you have a good eye for detail, please give it a go and see what you think. If a particular image is bad, please note down what the quality level was when it was compressed (just rerun MakeDDS on it), and perhaps post which image it is and what the quality number was.

Cheers,
Jonathan