Coming Soon -- Ron's New eBook

Comprehensive Guide to

Mastering Digital Photography

Check It Out!


Follow Ron on Twitter

Transcript for NOise Averaging Video

Narration by Ron Bigelow

Photoshop CS4 Used in this Tutorial

Welcome to the Ron Bigelow Photography Noise Averaging tutorial. Today, we are going to look at a very interesting technique to reduce noise. Now, most of the digital cameras today have very low noise under most circumstances. However, even with the best equipment, we can sometimes end up with noise. When this happens, photographers usually resort to various noise reduction techniques. However, most of these techniques share a common problem. The techniques can not reduce the noise without causing some degradation of the image detail. In other words, the more you reduce the noise, the more the image detail is degraded.

However, this is not the case for noise averaging. The noise averaging technique can dramatically reduce noise without affecting the image detail. On the other hand, all digital techniques have their strengths and weaknesses, and noise averaging is no exception. The downsides of noise averaging are that it will not work with all images and the photographer has to plan ahead of the shot for noise averaging to work.

Notice that I said that the photographer needs to plan ahead of the shot in order to use noise averaging. This is because noise averaging requires the photographer to shoot identical multiple shots of the scene that he is photographing. In other words, the photographer puts the camera on a tripod and takes multiple shots of the subject. In order to make sure that all of the shots are identical, the photographer needs to lock down the focus, exposure, and white balance so that they cannot change. Consequently, that means that noise averaging works in situations where the photographer can use a tripod and where the parts of the subject that are going to be noise averaged cannot move during the exposures.

Now, let’s take a look at how we carry out the noise averaging technique.

Here we have an image of a doll. When we fit the entire image on the screen like this, it doesn’t look too bad. However, when we move in to see the actual pixels by pressing Contol+Alt+0 on a PC or Command+Option+0 on a Mac and use the navigator so that we see the face of the doll, we see that this image has a very significant noise problem. Clearly, I would not want to print the image when it looks like this. On the other hand, noise averaging can save the day.

If I click on the Arrange Documents icon and select Tile All in a Grid, you can see that I actually have several images of the doll. All of these images are separate but identical exposures of the doll – just like we discussed a moment ago. We now need to move all of these images into a single file. I think that I will put all of the images in the Doll 01 file. We can do that by first selecting the Move tool in the Tools panel and dragging the images to the Doll 01 file while holding down the Shift Key. Holding down the Shift key ensures that the images are placed directly on top of each other without any shifting of the images.

We now want to focus on just the Doll 01 image. To do this, we make sure that the Doll 01 image is selected. At this point, you need to make sure that all of your images are aligned. In my case, the images are all aligned, so we can proceed. We now click on the Arrange Documents icon again and select Consolidate All. If you look at the Layers panel, you can see that we now have a layer in this file for each of the images that we moved over.

Now, before we proceed to carry out the noise averaging technique, let’s talk a little about how the noise averaging technique works. Each of the image layers in the Layers panel has two items: image detail and noise. The important point here is that the image detail in all of the layers is essentially identical. Conversely, the noise in the images is random. The noise pattern in image one is different than the noise pattern in image two which is different than the noise pattern in image 3 and so on. To demonstrate this, watch as I one-by-one click off all of the layers except the Background layer. Did you see how much the noise changed?

This provides us with a great opportunity. What if we were to adjust the layers so that each layer contributed an equal amount of information to the overall image? Basically, we would be averaging the information from all the image layers.

Now, this is where the noise averaging magic works. Since the image detail in the layers is identical, averaging the layers has no effect on the detail – it remains the same. However, the noise is a completely different story. Since the noise is different in each layer, averaging the layers averages the noise. In a sense, it smoothes out the noise. As a result, the noise is substantially diminished.

Let’s apply the technique to this image. To make things easier, I am going to rename the layers so that it is easier to follow the technique. Rather than make you wait while I rename the layers, let’s fast forward to the point where I am done with the renaming.

I have just finished renaming the layers. Next, we need to average the layers. The way that we do that is by adjusting the opacity of each layer. So, the next question becomes, “How do we determine the opacity to use for each layer?” That’s easy; we use a very simple formula. The opacity equals one divided by the number of the layer. We then convert the number into a percent. Thus, the opacity of Layer 1 is one divided by one which is 1.0 or 100%. Since Layer 1 already has an opacity of 100%, we do not have to do anything. The opacity of Layer 2 is one divided by two which is 0.5 which converts to 50%. So, we turn on Layer 2 and adjust the opacity of the layer by clicking next to the Opacity setting and typing in 50. Notice that the noise has already improved. Watch what happens as I turn the eye icon off and on. The noise is much better by adding just Layer 2. Adding more layers will improve the noise even more.

I will now turn on Layer 3. The opacity of layer three is one divided by three which is 0.33 which converts to 33%. Thus, we adjust the opacity of layer three to 33%. We continue with this process until the opacity of all of the layers has been adjusted. Let’s fast forward to the point where I have completed adjusting the opacity for each layer.

All of the Opacities have now been adjusted. Let’s use the History panel to see the difference in the noise. This is the image before noise averaging. This is the image after noise averaging. The difference is significant. Most of the noise is gone, but the image detail is a good as ever. However, if you wanted even lower noise, you could simply use more images.

So, that concludes the noise averaging technique. At this point, I am sure some of you are thinking, “Okay, that works great if you are photographing static subjects like your doll, but it will not work for my images. In my images, the wind blows the trees around, water flows through rivers, and birds fly in the sky”. However, noise averaging can often be used even in these situations. Simply carry out the noise averaging technique. After flattening the image, you can drag the flattened image over to one of the original images and use a mask to mask out any areas that have movement. Thus, you get noise averaging only where you want it. Since noise is most prominent in the shadows, this frequently means that you would apply noise averaging to the shadow areas and leave the rest of the image alone.

That brings us to the end of this tutorial. However, before I sign off, I would like to take this opportunity to let you know that you can download the transcript for this video, view several other photography videos, and access over 100 photography articles on my website.