|Al's Photoshop Tutorial||Advanced Concepts|
Noise is, very simply, an effect where every pixel gets a random value added to it. This may sound simple, but it can be the first step towards just about any surface that has that "natural chaos" look to it. In this first step, I explain some of the math behind noise, but don't be frightened! You don't really need to understand the math to use the techniques.
To the right, you see the Add Noise window. This is activated with Filter » Noise » Add Noise. The Amount slider works in an 8-bit scale, which means that it works with 256 shades of gray between black and white. The slider goes far beyond 256, however. When the slider is set to maximum, your noise will be almost all pure black and pure white with no grays. In this example, I'm starting with a piece of 50% gray (that's 128 to the computer, or 0.5 (128/256) to us), to which random values between -56/256 and 56/256 are being added. I also have Monochromatic turned on, so all values are applied identically to R, G, and B, so that I'm still working with grays only.
Photoshop gives you two choices for noise distribution, Uniform and Gaussian. These are
shown below, along with a snapshot of their resulting histogram (the chart from the
Above, we see Uniform noise. Values here are purely random and unbiased, so that any value is (in theory) just as likely as any other value. This was done with a value of 56 in the Amount slider, so all values fall randomly between (128-56)/256 and (128+56)/256.
This is Gaussian noise. Here, the randomness falls into more of a curve, where midtones are most likely, and darks and lights are rare. You'll notice that values here are found beyond the +/-56 border that the Uniform noise fell into, even though this was also made with an Amount of 56. This value, in this case, is acting more as a midpoint, and the curve continues to approach 0% probability as it reaches positive and negative infinity, in theory. In reality, it bumps into solid white and solid black and bunches up there. You can see the spikes at the very edges of the graph.
Well, that just about does it for explaining what noise is, and how it works. In the next step, I'll start explaining practical uses for it, and I'll try to give some insight into why what I do works. If you understand the functionality behind it all, you'll be able to make your own unique effects.
|Back: Advanced Concepts||Home||Next: Common Filters|