Jump to content

Recommended Posts

Posted (edited)

Is there a minimum image requirement to enable the Stacking feature, for processing noise reduction? I watched the Affinity tutorial where a burst of 30 were used. There was a mention of using up to 40/50 images but no mention of a minimum for the process to work. Would love to hear from folks who have used this feature. Thanks in advance.

addison

Edited by addisonk
Forgot to check notify for responses.

Share this post


Link to post
Share on other sites

To answer this, there is no minimum, but to explain it better you need to understand how it works. So it's working on the law of averages, whereby if you take a single pixel on the image and it had some noise (random value between 0 and 255 for example, although in reality it would be across three channels and unlikely to hit the extremeties of the value range). For the sake of simplicy let's imagine the orignal image without noise should be the value 127. So if the noise on the original image changes the value to 254. You now need another image with noise at a value of 0 to bring that back in line to where it should be ((254 + 0)/2 = 127). Of course noise isn't likely to be that severe, but it's very unlikely that with 2 images the two noise values would cancel each other out perfectly, but in theory the end result of two images should be closer in truth than either one of the images alone. Obviously, the more images used in the process the more they are likely to smooth each other out. In theory, each noisey value should be biased towards the original 'correct' value, so after adding them all up and dividing the result you should be pretty close to the true value. So yes you can do it with any number of images, but the more the merrier. Two images is unlikely to make a lot of difference, but you may see some improvement. Two should still be better than one. However, there must be a point at which adding more images also becomes fairly pointless. The difference in quality between using two and three images will be greater than the difference in quality between 45 and 46 images.

The next problem is capturing the images without capturing movement in the image, fine for studio type still shots but not so easy in other situations.

Share this post


Link to post
Share on other sites

At a qualitative level your explanation is quite good.  However, this is not what is actually happening at a statistical level.

Here is a copy of a paper written by John Paul Caponigeo (Use Multiple Exposures To Reduce Image Noise, December 13, 2017) in which he describes what goes on in Photoshop.  I assume AP behaves in a similar manner.  Occasionally I have used multiple exposures for star photography to reduce image noise in both PS and AP, it works.  

"So what is Photoshop doing? Photoshop first aligns a series of images as separate layers, converts them into a Smart Object, and blends them, reducing or amplifying the differences between the layers with a variety of rendering modes. You can choose one of eleven rendering modes; Entropy, Kurtosis, Maximum, Mean, Median, Minimum, Range, Skewness, Standard Deviation, Summation, and Variance. Few people will ever use all of them; most won’t use any of them; but I recommend you try two – Median and Mean. (Stacks were designed for analytical tasks in various scientific fields, like astrophotography or forensics and they’ve since been put to many other uses.)

Median and Mean select values in between the highest and lowest values, smoothing out the differences between aligned layers in a stack. Median works best for images with some motion, either subject or camera, to remove moving objects or noise. Mean works best for processing exposures without motion. (Astrophotographers typically make many exposures, sometimes dozens or more, of the same subject and use Mean to reduce noise.)

The more exposures you make and combine the better the noise reduction. Only practical limits apply. How many exposures can you make? How many exposures can Photoshop process on your computer? You can stack and process as few as two images. Three is my recommended minimum. Six is better. After that, you get diminishing returns. (Try using your camera in burst mode more frequently.) The most challenging part of this technique is identifying situations where it’s helpful and remembering to make multiple exposures. If you have the exposures you can take advantage of this great feature; if you don’t have the exposures you can’t."

Share this post


Link to post
Share on other sites
10 hours ago, Big_Stan said:

Should be

Post can be edited and hence corrected.


Affinity Store: Affinity Suite (ADe, APh, APu) 1.7.3.481.
Windows 10 Pro, Version 1909, Build 18363.476.
Dell Latitude E5570, i5-6440HQ 2.60 GHz, 8 GB, Intel HD Graphics 530, 1920 x 1080.
Dell OptiPlex 7060, i5-8500 3.00 GHz, 16 GB, Intel UHD Graphics 630, Dell P2417H 1920 x 1080.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×

Important Information

These are the Terms of Use you will be asked to agree to if you join the forum. | Privacy Policy | Guidelines | We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.