A Brief Summary of Image File Formats (WIP)

This page will explain some of the Image file types you can use in your Ogre projects. This list was gathered directly from DevIL, the image library historically used in Ogre. Ogre now uses the FreeImage image library which supports a superset of the image formats listed below.

Now, I'm going to omit some of the lesser known or used formats, and just give a brief description of what they are, how they are typically used, how to create them, and any special requirements for Ogre applications. (Could use help on that part).

The .BMP

Bitmap File
File Type: Raster Image
File Extension: .bmp

Bitmap files have some confusion associated with them. Some refer to any pixel-based image as a bitmap file. However, a true bitmap image file refers to the standard Windows image format. This type of file is mostly used on DOS- and Windows-based machines.
A bitmap file is a raster- (or pixel-) based format that only supports the RGB colour space and bit depths of 1, 4, 8, or 24 bits per channel. The BMP (bitmap) file type is used by many Windows applications. A Bitmap image is a representation of a graphic image, displayed by setting individual pixels to white, black, or a colour.

The bitmap is very accurate in recreating and holding your image (appearance-wise). But the drawback is, this can be the largest file type you can use. It stores all the colour information pixel by pixel, and some would say it's outdated. But this is a good format if you're working on a game that you wish to run on a broad range of computers.

The .JPG

JPGs do not limit colours, making this format the preferred for photographs. Always save your original image in a format other than JPG so you can go back and make changes later. If you save a JPG as a JPG after altering it, you are compressing the image again, causing lots of icky (that's a technical term :-) ) things to happen. Your JPGs can be highly compressed, resulting in really small files, or just a little compressed for larger files. You should experiment and save the image with the most compression, while still maintaining an acceptable image. If you save a JPG in Photoshop, you will not see the effects of the compression until you close the image and reopen it.

The JPG is a savable option in most real (As in NOT MS Paint) image editing packages. Photoshop even has more options that you can use to compress the filetype to your heart's content. So this format is created after you create or manipulate the image. Note: the compression effects of the JPG format will 'stack'; if you save your original image as a JPG, then reopen it and save it again, you have just compresed it twice, and it will look the worse for it.

The .GIF

GIF was created by Compuserve as a machine independent file format. GIF files can contain one or more images with 256 colours or less. The images are compressed using LZW compression, reducing the file size by 30-50% or more. GIFs are flexible, and have capabilities that JPGs don't, such as animation (GIF89a) and transparency. There are interlaced GIFs that load gradually, allowing the user to see parts of the image while it's loading. Interlaced files also take longer to decompress, and may actually take more time to download, although it may seem faster because the user has something to look at. It isn't worth the decompression time for small files. GIFs can also have transparent backgrounds to show the underlying background color or image. GIFs are generally smaller than JPGs, but larger than PNGs. Unless you are doing a large image, there is little to no difference in picture quality, but your palette is limited to 256 colours.

The GIF Is a savable option in most real image editing packages. Photoshop even has more options That you can use to compress the filetype to your heart's content. So this format is created after you create or manuplate the image, but it's recommended that you keep an eye on the colors you use while creating the image. When you convert the image to the GIF format, if you have more than 256 colours the program you use will try to "Match Closest Colour". This may create unintended effects when you save it as a GIF.

There is no reason to use GIF as an image format for games. For paletted textures, use PNG; for true colour textures use JPG or PNG.

The .RAW

The RAW file format is digital photography's equivalent of a negative in film photography: it contains untouched, "raw" pixel information straight from the digital camera's sensor. The RAW file format has yet to undergo demosaicking, and so it contains just one red, green, or blue value at each pixel location. RAW also has the advantage of being smaller (generally) than an uncompressed TIF file. If you want to avoid the lossy compression you get with JPGs, but don't want to put up with the huge file sizes of TIFs, this is a super way to go. Plus your RAW image is kind of like a photographic negative - if you mess up your processed image somehow, you can always use the RAW file to make a new one.

Note that different brands of digital cameras have different sensor mosaics (both in size and topology) and hence different RAW formats. To add to the confusion, various imaging and 3D packages employ their own RAW format. This makes RAW unsuitable as general image exchange or texture format.

In Ogre, this fileformat is used in Greyscale 8, 24 and 32 bit formats for representing -Height Maps. The values must be little endian encoded.

The .PNG

PNG is of principal value in two applications

  1. Large palette. If you have an image with large areas of exactly uniform colour, but containing more than 256 colours, PNG is your choice. PNG's strategy is similar to that of GIF, but it supports 16 million colours, not just 256.
  2. Lossless. If you want to display a photograph without loss on the web, PNG is your choice.


PNG is superior to GIF. It produces smaller files and allows more colours. PNG also supports partial transparency. Partial transparency can be used for many useful purposes, such as fades and antialiasing of text, or grass that fades. The alpha channel can be regarded either as a mask that temporarily hides transparent parts of the image, or as a means for constructing a non-rectangular image. In the first case, the colour values of fully transparent pixels should be preserved for future use. There is no lossy compression in PNG. Existing formats such as JFIF (JPEG File Interchange Format) already handle lossy compression well.

The .TGA

The Targa TGA format was developed by Truevision for their Targa and Vista products.
The TGA format file is widely used by high-end paint programs and ray tracing packages. It is a lossless and uncompressed format, thus producing huge files. For production purposes, the PNG format is better suited as a lossless format.

The .DDS

The DirectDraw Surface graphics file format was established by Microsoft for use with the DirectX SDK. The format is specifically designed for use in real-time rendering applications, such as 3D games. It can be used to store textures, cube maps and mipmap levels, and allows for compression that can be carried onto the graphics hardware (S3TC / DXT) unlike other compression algorithms like JPG and PNG. Although hardware texture compression is not mandatory in DDS files, it remains the most common reason to use to format.

S3 Texture Compression (S3TC) (sometimes also called DXTn or DXTC) is a group of related image compression algorithms originally developed by Iourcha et al of S3 Graphics, Ltd. Its subsequent inclusion in Microsoft's DirectX 6.0 led to widespread adoption of the technology among hardware and software makers.

Wikipedia:
http://en.wikipedia.org/wiki/S3_Texture_Compression

benefits:

  • When using DXT/S3TC, saves video RAM, because the compressed data is directly transfered to vram
  • In Direct3D, there is no decoding since DDS is a raw surface dump. In GL (and also if you manually decode a DDS using the Image class), Ogre decodes the DDS header manually and then deals with the pixel data as per usual in a block
  • different kinds of Alpha Channels (partial alpha if you use DXT2-5)
  • precalculated mipmaps, cube map faces and volume slices can all be stored in the same file




drawbacks:

  • None for uncompressed data except that files are larger than some other formats
  • When using DXT/S3TC, compression is lossy (you lose data when encoding).
  • To use DXT/S3TC you need a GeForce 2 or later. If your video card don't support texture compression, Ogre3D will decompress the texture at runtime and use it like other fileformats (so its no real drawback).
  • When using compression, data is a little harder to manipulate since it is not stored directly per-texel, but as blocks of texels.




additions: Using S3 (dxt) Texture Compression also in OpenGL - NVIDIA document

There are several types of DXT compression offered by DirectX:

DXT1

* 0 or 1 bit alpha




DXT2

* Explicit 4-bit alpha, color data is premultiplied by alpha




DXT3

* Explicit 4-bit alpha, not premultiplied




DXT4

* Interpolated alpha, color data is premultiplied by alpha




DXT5

* Interpolated alpha, not premultiplied




Which DXT Compression to Use?
Obviously, there are some trade-offs between the different formats which make them better or worse for different types of images. Some general rules of thumb for good use of DXT textures are as follows:

1. If your image has no alpha, use DXT1 compression. Using DXT3/5 will double your image size over DXT1 and not gain anything.
2. If your image has 1-bit (on or off) alpha information, use DXT1 with one-bit alpha. If the DXT1 image quality is too low and you don't mind doubling image size, use DXT3 or DXT5 (which one doesn't matter, they'll give the same results).
3. If your image has smooth gradations of alpha (fading in/out slowly), DXT5 is almost certainly your best bet, as it will give you the most accurate transparency representation.
4. If your image has sharp transitions between multiple alpha levels (one pixel is 100%, the next one is 50%, and another neighbor is 12%), DXT3 is probably your best bet. You may want to compare the alpha results in DXT1, DXT3 and DXT5 compression, however, to make sure.


Specification:
http://oss.sgi.com/projects/ogl-sample/registry/EXT/texture_compression_s3tc.txt

Infos:
http://www.sjbrown.co.uk/?code=squish

NVidia has some Tools on their Homepage to produce DDS files from Commandshell or Photoshop.

Tools
To create DDS files in Photoshop you must download the Nvidia DDS plugin. DDS files may also be created using the DirectX utility DxTex, found in the Utilities/bin folder of the DirectX SDK.

http://developer.nvidia.com/object/nv_texture_tools.html

The GIMP also has a plugin which can handle DXT-compressed DDS images.