by Meysam                

Submit your Photo
Hall of Fame

Please participate in Meta
and help us grow.
Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Can I get the same results by copying a RAW (same exposure) file many times and stacking to reduce noise as I would if I used many separate exposures?

share|improve this question
It's unclear if you mean one single "click" (a literal single exposure) or if you mean multiple photographs taken at the same exposure level. Can you clarify? – mattdm Mar 13 at 18:15

Can I get the same result by coping a raw ( same exposure ) file many times and stacking to reduce noise as I would if I used many separate exposures?

No. If you stack copies of the same image, you'll amplify the noise right along with the signal.

Stacking images to reduce noise is an averaging process. It's like a science experiment: you don't measure something just once, you measure it multiple times and average the measurements to reduce the effects of error in the measurement process. Doing what you propose would be like taking just one measurement and writing it down in your lab notebook several times.

When you take an image, you're measuring light intensity at every point on the sensor, and noise is essentially the error in each of those measurements. Because the error is random, it changes from one exposure to the next. For two adjacent photosites, the data recorded across 5 exposures on a 0-255 scale might be:

150, 152
148, 146
144, 145
151, 148
147, 149

Adding those up and dividing by 5, the average value for both photosites is 148. Averaging the values over several exposures smooths out the data, eliminating the noise in each photosite. Even though the individual measurements for the two photosites are a little different, they both have similar averages because they're both recording the same part of the scene.

Now consider what would happen if you stacked multiple copies of the first exposure:

150, 152
150, 152
150, 152
150, 152
150, 152

Now you're not removing any noise at all -- the averages are 150 and 152 respectively. You get exactly the same degree of noise that's present in the image that you used.

share|improve this answer
"taking one measurement and writing it down several times" is a great analogy. – immibis Mar 14 at 8:09
It's worth noting that sensor noise also has a static, non-changing (or more slowly-changing) component - in scientific terms there's both random error and systematic error. The latter can be counteracted by subtracting dark frames and bias frames from the image. – JohannesD Mar 14 at 14:01
@JohannesD That's true, but the OP is asking specifically about stacking images to reduce noise, and of course that won't help with the static component for the same reason that stacking copies of the same image won't help at all. – Caleb Mar 14 at 14:51
Excellent thanks – Chad Williams Mar 14 at 16:41

You cannot. Removing noise via photo stacking works on the principle that the noise in your images is random, and appears in different places of the image between exposures. When you stack multiple exposures on top of each other, a common method to remove this random noise is called median blending. During this process, software will evaluate the same pixel in each version of the image, find the median value, and assume that is the true color value of the pixel since the random noise values should not be near the median of the pixel's color distribution.

When you stack multiple copies of the same exposure, your noise is going to be in the same exact location in each copy, and thus cannot be removed via this method. There are however noise reduction tools that will attempt to remove noise from a single exposure of an image.

share|improve this answer
What do we do when there's only two stacked images? Use local information to choose which pixel to take as "median"? I suppose that has its own problems... – Mateen Ulhaq Mar 14 at 16:42
There are stacking options besides median. However, if you used a median stacking method for two images you should get the same result as if you had used a mean stacking method (the median of an even set of values is the mean of the middle two. In the special case of a set of two values, the median is simply the average of the two). If you're interested, here is a bit more information about the standard stacking methods in Photoshop: – tittaenälg Mar 15 at 2:26

One possibility might work and be what you say: the same RAW exposure processed different ways using RAW-processing tools, to bring out different parts of the image. E.g. tge background may have its exposure boosted ans be run through a plug-in advanced noise reducer, while the foreground has different white balance, lower standard NR, but a touch of smoothing.

If the RAW-processing tool doesn't have regional editing or layers, or you're actually using different programs for different parts of the image, then you can bring them in as layers in the photo-editing program for final compositing.

Or course that's not the same as the layer-averaging tool you were implying.

share|improve this answer
This technique would only help circumvent shortcomings of the editing software but not change the noise characteristic in general. I never understood why people would feed their HDR tools with processed copies of a single raw file with different exposure. – Chris Mar 14 at 9:54
Downvoter: which category are you claiming: "egregiously sloppy, no-effort-expended post, or an answer that is clearly and perhaps dangerously incorrect"? I don't think one typo is "egregiously sloppy" and thinking up a possible reason for this is definitely effort. It's possibly useful as stated, not dangerous by any stretch. – JDługosz Mar 15 at 0:10

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.