There’s a long-standing debate in the icon design field as to how best to create icon graphics. Some people, including your’s truly, prefer creating icons with vector graphics. However, there are a sizable number of people who prefer using pixel-based tools like Photoshop to create icons. Some people argue that one technique will get you better results than the other, but if you understand the strengths and weaknesses of both you can find what works for you.
I really don’t believe there is a right answer, and I don’t think one option is more work than the other. They both have their advantages, but they also both have some challenges that take a little elbow grease to overcome.
A Little Background
So for many years icons were drawn pixel-by-pixel, painstakingly crafted by graphic artists. In the very early days these icons were usually done on graph paper before being handed off and programmed in by a developer. As time went on though, tools became available that allowed artists to place pixels directly on the screen.
This sounds very tedious, especially when you consider that icons today can be as large as 1024 pixels square, if not larger. However, for many years icons were merely a fraction of that size, and didn’t contain any colors except black and white (not even grey!). As time went on and display and graphics technology improved, icons became more complex. First they added color, then more colors, and then grew and grew until we arrived to the relatively gargantuan icons we use today.
Although some designers still work directly with pixels, they’re no longer drawing each one individually. Instead, tools like Photoshop allow the designer to simply draw shapes, color them in, and add any number of effects. However, designers can manipulate the individual pixels when needed and that can make it much easier to attain pixel-perfection. Working in a format computer screens can understand natively, if you put the work in, can result in much crisper icons than vector.
However, a major downside to working with pixel-based graphics tools is that you cannot easily resize your work. A pixel-based image is essentially just a grid of colored squares, and when you increase the size of that image you’re going to just increase the size of the colored squares. Your enlarged icon is going to just be a mess of blocks, and that’s no good (unless that’s what you’re going for I suppose). Modern graphics tools have functions that can interpolate and smooth out blown up images, but these are always fuzzy and not an acceptable solution. On the other end, shrinking pixel grapics down can also lead to some graphical artifacts that can affect the quality of your work.
To solve the resizing issue you will need to redraw your icon at each size the system you’re designing for requires. For example, modern versions of OS X require 7 sizes for app icons ranging from 16 pixels square to 1024 pixels square.
You’re going to find some very similar tools in vector graphics software for drawing shapes and adding color and styles. However, since vector is math-based it mostly solves the resizing issue in the fact that vector graphics are infinitly resizable without loss of quality. Rather than drawing a grid of squares, vector images are built with lines and points.
However, vectors do have their own set of issues by virtue of the fact that they have to be translated into pixel-based graphics to be used as icons on most modern operating systems. This is because vector points and lines will not always correspond strictly to the pixel grid when converted. Upscaling is generally easier since you’re only doubling the pixel size, but if you’re trying to shrink a vector icon you’re going to run into very similar issues as pixel design.
You could simply draw the icon for export at the smallest size and then output the same image at larger sizes but then since you can’t pack as much detail into a smaller canvas you run into having an icon that is overly simplistic. It’s best to draw your icon at 2 or 3 different sizes and upscale those accordingly. The more the better since it’s a little harder to achieve pixel-perfection with vector art, especially when resizing as edges may not always line up exactly with the pixel grid.
I don’t believe that there is a clear winner here. Neither technique is going to make the work easier or more efficient. Both have shortcomings, it’s just a matter of which you can find more enjoyable. I personally love tweaking vector paths to be exactly perfect, but you may enjoy recreating your icon and figuring out how to make it look great at different sizes.