Grainy versus Smooth: How to Reduce Noise in Digital Photos
Noise is Horror
[image courtesy of Poltergeist, a Spielberg Production, 1982]
Poor Carol Anne, the little girl sucked into the static of a demon-possessed tv... it is the worst of image noise, the kind we digital photographers would like to avoid.
A Brief Personal History of Static
Many years ago, before cell-phones and before iPods, while driving across Wyoming I stopped at a wayside for a much-needed break. And promptly locked my keys in the car. This wouldn’t’ve been such a horrible problem if Wyoming weren’t so darned big! I called the state police: they were many hours away and had a policy of not helping locked-out drivers (‘Too bad you had your windows closed’ said the operator, and ‘well, here in Wyoming you might just have to break in.’) I called the nearest town service station – 130 miles away. The earliest they could arrive would be four hours. I walked around and waited, and waited, and waited. I tried listening to my Walkman’s radio, but I was so far away from the civilized world that it too was out of the range of radio stations. This meant I could listen to a symphony of static.
Static is caused by many things – loss of radio signal, occasional solar flares, or just poor electrical equipment. If you ever have the chance to listen to some static, about ten percent of the noise is caused by the world around us, and another percent is caused by the radio itself. A tiny portion of radio static is in fact one of the great scientific discoveries of the 20th Century. Radio astronomers Arno Penzias and Robert Wilson at first thought the noise they experienced in their radio telescope was residual interference from nearby New York City. It wasn’t. They checked their equipment, and after much testing, figured out that a small slice of the noise was the universe itself – what today is known as ‘cosmic microwave background radiation’ (CMBR). Its existence is evidence that the universe is billions of years old and resulted from the Big Bang – a tiny bit of the static is waves and ripples streaming through space, caused by a great and distant explosion. Penzias and Wilson won the Nobel Prize for figuring out how to precisely measure CMBR, figuring out exactly what was the universe making noise, what was their radio equipment. Since then, radio telescopes have been greatly refined and have extremely fine sensitivity and very high signal-to-noise ratios, so that they can discern the most distant sources of radio waves in the universe. So, sitting in Wyoming, all was not lost and I could try to imagine what portion of my Walkman’s static was not merely static, but instead a faint signal from the universe 13.7 billion years ago. Finally the tow truck arrived, the service man jimmied my car open, I paid him a hefty fee, and have forever kept a spare car key in my wallet.
It turns out that just like a radio signal received by a Walkman, all electronic equipment is subject to some noise. Digital cameras are no exception to this rule. When you search for a radio station that you can listen to, you are searching for a strong signal – strong enough that the small amount of noise caused by your radio equipment is overwhelmed and thus, not noticeable. If you take a photo in low-light settings, when you have a hard time getting a clear visual signal, then you’re camera will record a noisy, grainy image. It’s static at it’s finest. Noise is the opposite of Signal. What’s Signal? It is the information, the data, that your camera’s light sensor records – the light that moves through the lens and hits the camera’s sensor. In any electrical device we talk about the “Signal-to-Noise Ratio.” In other words, to get a digital photo that is less noisy (and thus sharper, smoother, and less grainy), we need to figure out how to increase the Signal.
Noise is increased whenever you lose the Signal. In digital photography, you can lose the signal in many ways, such as leaving the lens cap on (no light), to selecting the wrong exposure for a low-light setting. A low-signal digital photo will look noisy like this:
This is a photo of Luna, who was being very patient with me. And this is a crummy photo full of colorful static – noise. My camera tried really hard to make a good image, my model was very forgiving and willing to sit still, but there just wasn’t enough signal. There’s nothing that can be done to rescue this dismal photo.
Here’s a much, much, much better photo that was made from a very clear signal:
The lighting was good, the camera settings were correct, and thus sharpness and focus were good. And as always Luna got a little tired of me fiddling around with camera settings – she fell asleep. The fact remains that this is a good photo, and with image processing software, you can do just about anything with it.
Some artists use graininess on purpose. I think you’re better off shooting a smooth, sharp photo and then using an ‘add noise’ filter if you would like a grainy look. The problem is that it’s darned difficult to remove noise from an image, in the same way that it’s impossible to sharpen a poorly focused image and make it focused. So, we start with low noise and perfect focus, and then alter the noisiness and blurring in an image editing program if needed.
How to Reduce the Noise in Your Images
- Start with a good Signal.
o Great photos begin with good lighting matched to your camera’s settings and capabilities. Some cameras are just more noisy than others – most cameras have a couple of settings where everything works well to produce low-noise, high-signal images. Experiment to find out what kinds of images produce good signal for your camera.
- Check Your ISO Settings.
o an ISO setting of 100 is usually the best. For digital photography, ISO is a ratio of how sensitive your camera is to light – set it higher and higher and the ratio changes. The higher the ISO, the more sensitive to light, and the more overwhelming the noise becomes. Even the best digital SLR cameras at ISO settings about 1600 produce a lot of noise.
o Some scenes require higher ISO settings. Aim for 100, but alter it up to 400, 800, or 1000 if you know you need to because you are working in lower-light settings, etc.
- Know Your Lighting
o Where is the light source? Is there a lot of light or are you in a low-light setting?
o If you’re in a low-light setting, can you add some light? (use a reflector, flashlight, or flash?)
- Use the ‘reduce noise’ filter in an image processing program. Try it out and see what happens – but be warned that too much of the filter will make your photo blurry.
o In Photoshop, it’s under menu Filter:Noise:Reduce Noise
o In GIMP, it’s under menu Filters: Enhance: Despeckle
o In Photoscape, Filter: Noise Reduction (Clear Skin): select High, medium, low.
o Consider trying ‘Noise Ninja’ – this is not a free program – but it is awesomely useful. You could get it as a standalone program or a Photoshop plug-in at http://www.picturecode.com/index.htm You can download a free trial version that will show you what it can do, but won’t let you save images. At the very least you can read the company’s description of noise in digital imaging at http://www.picturecode.com/noise.htm
- Reshoot. It’s possible to take a variety of photos of the same scene, and then combine them to reduce noise. This is an advanced topic that we’ll look at later.
For more information, read this link: http://digital-photography-school.com/photography-1017-iso
Bonus: Artists Tom Moody and Ray Rapp have made photos of static into fine art: http://www.digitalmediatree.com/tommoody/photoinst/statphotos/
