Digital SLR photography — What is noise and what can you do about it?
When photographers talk about noise, they aren’t usually talking about the sound their shutters make while firing (although that can cause problems too). So what do you need to know about noise?
With the right settings, even the shadows in a photo can be virtually noise-free.
First of all, what is it?
Noise in a photograph is the speckling of incorrectly-tinted pixels, like in the examples lower down this page. Taken to extreme, noise (and efforts to remove it) can result in a big loss of detail, colour accuracy and overall image quality.
So, noise is a bad thing, right?
Only if it starts to bother you during the ‘normal’ ways of looking at a picture. By normal, I mean looking at a print of the photo, or looking at the whole image on a screen. If the noise still bothers you at those times, then you will want to do something about it. But because almost every digital photo will have some amount of noise in it, you don’t have to worry about it all the time.
Why do we hear so much about it?
These days, it’s just too easy to blow an image right up big on a screen, zooming in on just a small part of the image so you’re seeing the individual image pixels. When you do that, it’s possible to see the tiniest imperfections, including noise.
Because we can see it, some of us are worried about it.
But when you think about it, filling a big monitor with only a small part of an image to see noise is not really the way to appreciate a photo. It’s like going over someone’s garden inch by inch with a magnifying glass to determine if it’s an attractive garden.
So noise is really only a problem if it gets so bad that it starts to detract from the quality of the overall image.
Why does noise happen?
The surface of a camera’s sensor — that bit which takes the place of the film — is made up of millions of tiny pixels. Each one of those pixels is crazy-small, and that’s part of the reason why noise happens. You see, the tinier those sensor pixels get, the less surface area they have to capture photons. Which means that each pixel will have less information to work with. That results in mistakes. Some of the pixels might register the wrong colour (colour noise), or the wrong light/dark amount (luminance noise).
You’ll find that, as the amount of light diminishes, the amount of noise in your photos is likely to increase. Which makes sense when you think about it. Because if there’s less light around, then there are even fewer photons landing on those already-struggling sensor pixels. So even more mistakes happen.
I talk a whole lot more, including a mention about why large megapixel counts in digital cameras can contribute to noise, in this article about camera megapixel counts. But to sum up the relevant bit here: the larger the number of megapixels in your camera’s sensor, the tinier those little pixels are going to have to be to fit them all in. And as we now know, the smaller they are, the bigger the chance that they will make some mistakes. That means noise. Sometimes giant megapixel counts are not such a great thing.
Increasing your camera’s ISO setting is like using faster film in a film camera. It allows you to work with faster shutter speeds or smaller apertures in the same amount of light. It does this by increasing the sensitivity of your camera’s sensor. The problem with that is, if you increase your sensor’s sensitivity too much then you will be sacrificing accuracy in how well each sensor pixel works. Which means more noise. For this reason, a lower ISO figure is usually best if you want to get less noise in your photos.
Getting it wrong: The packaging in this photo should be an even, flat red colour. Instead the photo is carrying a lot of noise.
It should make sense that, if the pixels in your sensor aren’t getting enough photons to register the correct colour, then they will be receiving even fewer photons from the shadow areas in your photos. That’s why the shadows are the first areas you will start seeing noise.
Yes folks, the people who come up with ideas like low-calorie mineral water are also contributing to noise. You see, some camera manufacturers are advertised with unrealistically high maximum ISO settings and equally ridiculous high megapixel counts. While the megapixels counts are arguably better handled by digital SLRs, due to their larger sensors, some of the very high megapixel counts on the tiny compact cameras are probably too high for what their technology can handle in anything but the best lighting conditions. So there are now heaps of cameras — especially the smaller compacts with zillions of pixels squeezed onto tiny sensors — producing noisy pictures due to some marketing need to promote ISO and megapixel features beyond what they can properly handle. Personally, I don’t want a camera which can record a zillion megapixels if none of the pixels are the right colour, but big-business marketing takes advantage of the fact that most customers don’t know what is reasonable to expect from a camera according to current levels of technology.
This photo of a Superb Fairy-wren (above) was badly affected by noise
Same kind of bird, same camera and same lens. But this photo was taken in better light.
How to get rid of noise.
You can use software to remove noise from your photos. Programs like Adobe Lightroom allow you to view your Camera Raw files and drag a slider to adjust the amount of noise, and providing that the noise levels aren’t too bad then you can end up with some great results that way. In fact, I give most of my shots a very slight software adjustment to their noise levels. But there are limits to what even the best software can do. In the photo below I’m starting out with some really, really bad noise levels. See how, in the next version, the software fixed the noise problem, but the resulting image was still pretty horrible. That’s why you’re better off avoiding lots of noise in the first place.
You can use software to fix noise but if the noise levels are extreme, like they were in this shot of a Green Catbird, then you will still pay the price of poor image quality.
If you’re getting a lot of noise in your photos then what is happening is that your sensor’s pixels are struggling to capture sufficient numbers of photons to tell them what to do. You can’t make your sensor pixels bigger/better unless you buy a different camera. But you can help your existing sensor pixels do their job by giving them more information to work with.
That means, you need to feed more photons onto those struggling sensor pixels and so it’s not surprising that most of these suggestions involve getting more light into the picture. We’ve already found that increasing your ISO settings further will just result in more noise, so instead, you’re going to have to consider some of the things mentioned here:
- Move your subject out of the shadows
- Use a flash
- Reduce your ISO settings
- Open up your aperture more (if your ISO is set to auto then this will reduce your ISO settings).
- Slow down your shutter speed (this too, will reduce your ISO settings if your ISO is set to auto, although it’s not much good if your subject is moving quickly)
- If you can’t get more light into your scene and so your camera keeps automatically maxing out its ISO settings, lock the ISO at something you know will produce acceptable levels of noise. Chances are that doing that will then cause shutter speeds to slow down a lot as your camera struggles to get a correct exposure, so just be careful you don’t end up replacing noise with camera-shake blur.
If you’re already set on your widest-possible aperture and the slowest shutter speed you can manage hand-held, then consider using a tripod or any other sturdy object/surface to hold your camera still. However, now that I mention long exposures, you might also want to know that really long exposures, like ones lasting minutes, can also cause image noise.
One more trick!
If you reduce the resolution (image size) of your image then you will be throwing away a lot of pixel information. While that can sound bad, it can actually work to your advantage. Because when your image editing software throws away those unused pixels, it has to average out the tonal values of the ones it has left. This can result in the noise in your images diminishing. You might not always have the option of shrinking your image sizes down but for web work and email, small image sizes are perfectly acceptable. In fact they are preferable!