Theory 3/3 > The most usual image file formats

As said, digital optical information is stored in digital image files. This happens in various ways, and as a result there are various image file formats. We will see here the most important ones, because later we will want to save our work -- in what file format will this be done?

XCF

This is the GIMP file format. GIMP will be the basic software with which we will do our digital image processing, and this software has an inherent file format to save our work - the XCF file format.

• It supports layers - of course we will discuss layers in detail later in the school.
• It supports a lossless compression, which gives us a smaller file size but still without loss in optical quality, and about 30% better compared to PSD (which is the inherent file format of Photoshop).
• Its relative disadvantage is that some other programs do not open it, however GIMP both opens and saves PSD.
• It supports color depth to 24bit.
• It supports transparency at the layer level.
• Together with the layers saved, we can save valuable information with our work like channels, selections, and such, as we will see.

PSD

The file format of the known Photoshop is widely supported in many applications that import image data. Color depth is 24bit. It supports transparency, layers, and other important information. GIMP both opens and saves PSD, apart from some minor issues.

PNG

Another well known format, widely supported in the web and multimedia applications. It features transparency to 24bit, and offers a compression algorithm which is theoretically lossless and selectable.
PSD and PNG are good for DTP and prepress applications, and in multimedia when we want transparency at the file level. They are very useful and widely supported. Their difference lies in the fact that a PSD file can have layers but PNG not.

JPG

A well known format, appropriate for photos and photorealistic images. It supports 24bit color depth. It supports compression, which is selectable - this effectively being its strong point. Its only disadvantage is that it is lossy. The degree of loss is related with the compression selected. The compression algorithm allows for a huge range of compression, and of course greater compression results in lower quality. In the web we are concerned about small file sizes, and JPG is widely used because of that by offering a selectable compromise between compression and file size. Due to the compression and the consequent loss, it should not be used for archiving. For our photo archive we should prefer other file formats that are lossless.
For the same reason, a JPG file may not be saved twice. From a lossless file i may create a JPG anytime for a web page or an email. If this file, which has been made as lossy, is opened in an image editing application and saved again as a JPG, new loss is summed up, so we have a further loss of quality. If i want to have a JPG with precicely controlled compression, i should better open the original lossless file and have a JPG from that one, so the loss happens only once.
If we need to save transparency, we prefer PNG, since JPG does not support transparency. So we have color depth of 24bit, a lossless and selectable compression, and it is used mainly on the internet.

GIF

Here a color palette comes in, so we go down to 256 colors, as explained for the indexed files. This excludes photography, due to important loss of quality. Instead for a logo, for example, if we do not have tonal continuity as in a photograph but only just clear borders between shapes, it may be useful theoretically, since a logo may have only geometric shapes, if we can go with a pallette of 256 colors. Its compression is lossless, so if we do not need tonal continuity we may use it. The only reason to use it actually seems to be the small file size, being it indexed. It practically has special uses, by web designers, or animated graphics on the web. Such a graphic consists of frames made one by one, toggling in time to give motion. If that was done with full-tone images, we would have huge file sizes. So here we have a color depth of 8bit, lossless compression on those few colors, transparency, and still one color can be defined as transparent, and the possibility of animation.

TIF

It is an old standard, but it remains very useful. It is supported widely on applications and on all platforms, which is very important. It does not support layers though. If we want the absolute quality, it is appropriate even for the demanding photographer. It allows saving in CMYK, and thus was widely used in classic typography. Color depth is 24bit. Some versions support transparency. Its compression (if selected) seems to be a bit controversial. It allows saving in 48bit for special uses. It supports layers only in Photoshop.