Provided by a kind reddit reader. Helped me understand it. Archived for future use.
Vector graphics is another way for a computer to store and explain an image. A photograph from a digital camera is stored and “explained” by the computer as a grid of pixels of different values (colour) attached to them, this is a raster. This raster can then be compressed using different methods like JPEG GIF or PNG witch looks an the image and for example it sees that there are 500 fully white pixels in a row. And rearranges it so that the computer in stead of saying: “White pixel, white pixel, white pixel” 500 times, says “500 white pixels” which saves space.
A vector image on the other hand is made up out of mathematical curves in a coordinate system that have more… Well mathematical properties applied to them. Say that you have 3 points on a canvas. These 3 points makes up a triangle. This triangle is now a shape in your workspace. This shape can now have a list of properties, it can have a colour applied to it, it can have an opacity, it can have a blurred edge with the amount of blur proportional to the size of the triangle…
Then we can add more complex shapes to the workspace by adding more points to them. we can even make the lines that go between the points bend in smooth curves by adding tangent handles (more information basically) to each point. Choosing colours and other properties for each shape and step by step building up an image.
The great reward that sets this apart from raster images is that these mathematically built up images don’t “care” about resolution. If we want to make out image 2 times bigger then the computer just says “Sure ill just multiply the relevant values by 2″ and it will all be nice and crisp no matter how big you want to go.
Now if you tell a computer to make an image bigger… Say you have an image that’s 1000×1000 pixels with different colours, your gandmother and her cats or something… If you tell the computer “Make this 2 times bigger” the computer is going to have to add 4 000 000 pixels to the image, pixels which it really has no idea about how they are “supposed to be”… Now applications are smart enough to look at the surrounding pixels and determine “fictional” values for the new pixels. But no information is actually added to the image and thus the images resolution stays the same. This is why images appear blurry when blown up.
The weakness for vector images is though that it becomes very difficult, nigh impossible to represent complex images like photos this way. Therefore its suitable for graphics, logos and such but not pretty sunsets.