The following discussion is based on looking at how Blender/PyPRP handles texture conversion for URU; I have not yet had the time to confirm that Plasma converts in exactly the same way. But I will double-check this and confirm in upcoming months!
___________________________________________________________
Plasma uses textures in the .dds file format. Now, I understand this has efficiencies in its compression methods and how well it integrates with Direct X. But some other things about it have always perplexed me.
Here's why:
in normal RGB displays a pixel's Red, Green, and Blue colors are represented by eight bits of information for each color. Red, Green, and Blue together adds up to 24 bits of data. The addition of an alpha channel for transparency (RGBA) adds another eight bits, for a total of 32 bits per pixel (see RGBA color space).
It doesn't matter whether you've saved your original texture image file as a non-alphaed .tif,, .jpg, or .png image. It doesn't matter how nicely you've compressed the original image, or how effiiciently you may have optimized the image palette (a la Photoshop's "Save For Web & Devices..."). It doesn't matter if you only have one pixel of black in the middle of a solid 1024X1024 white (or transparent) background. The DDS conversion saves in the same file sizes, no matter what. This is very inefficient!
As with the standard RGB pixel data above, the addition of an alpha channel (for transparency) should simply add another 1/3 - another eight bits - to the size of a 24-bit color image.
But the Plasma .dds alpha conversion instead DOUBLES the size of each resulting texture file!
Note the 8X difference between the size of a 512X non-alphaed and a 1024X alphaed image!
The other thing to note is that Plasma only uses images in multiple-of-two sizes: 16X16, 32X32, 64X64, 128X128, etc.
If your image is NOT formatted in a multiple-of-two size, the conversion process will scale your image UP to the next suitable multiple-of-two texture size.
In other words, if your original texture is 256X256 pixels, it will convert to that same size. But if it is a 257X257 image, conversion will scale it up to a 512X512 .dds file. You will have effectively lost 3/4 of the image resolution you could have had if your file was originally made as a 512X512 image!
So: build your original images in multiple-of-two sizes in the first place!
That said, Plasma does handle 1X2, 1X3, 2X1 and 3X1 variations of the above sizes just fine. (so something like a 128X384-ratioed texture should process correctly)
___________________________________________________________
Plasma uses textures in the .dds file format. Now, I understand this has efficiencies in its compression methods and how well it integrates with Direct X. But some other things about it have always perplexed me.
Here's why:
in normal RGB displays a pixel's Red, Green, and Blue colors are represented by eight bits of information for each color. Red, Green, and Blue together adds up to 24 bits of data. The addition of an alpha channel for transparency (RGBA) adds another eight bits, for a total of 32 bits per pixel (see RGBA color space).
It doesn't matter whether you've saved your original texture image file as a non-alphaed .tif,, .jpg, or .png image. It doesn't matter how nicely you've compressed the original image, or how effiiciently you may have optimized the image palette (a la Photoshop's "Save For Web & Devices..."). It doesn't matter if you only have one pixel of black in the middle of a solid 1024X1024 white (or transparent) background. The DDS conversion saves in the same file sizes, no matter what. This is very inefficient!
As with the standard RGB pixel data above, the addition of an alpha channel (for transparency) should simply add another 1/3 - another eight bits - to the size of a 24-bit color image.
But the Plasma .dds alpha conversion instead DOUBLES the size of each resulting texture file!
Note the 8X difference between the size of a 512X non-alphaed and a 1024X alphaed image!
The other thing to note is that Plasma only uses images in multiple-of-two sizes: 16X16, 32X32, 64X64, 128X128, etc.
If your image is NOT formatted in a multiple-of-two size, the conversion process will scale your image UP to the next suitable multiple-of-two texture size.
In other words, if your original texture is 256X256 pixels, it will convert to that same size. But if it is a 257X257 image, conversion will scale it up to a 512X512 .dds file. You will have effectively lost 3/4 of the image resolution you could have had if your file was originally made as a 512X512 image!
So: build your original images in multiple-of-two sizes in the first place!
That said, Plasma does handle 1X2, 1X3, 2X1 and 3X1 variations of the above sizes just fine. (so something like a 128X384-ratioed texture should process correctly)