Despite the rapid improvement of camera technology, especially in terms of resolution, dynamic range and noise performance at high ISO's, there are still many circumstances where the combination of multiple images of a scene can be combined to create a result which the camera cannot render on it's own. Examples include:
1. HDR - multiple (different) exposures are combined to create a composite image with dramatically increased dynamic range.
2. Panoramas - multiple images are taken of a scene at different viewpoints (all at the same exposure). The images are then stitched together to create a panorama.
3. Focus stacking - multiple images are taken of the same scene but at different focus points. The images are then combined to produce a composite image with increased depth of field. This is especially useful for macro photography.
4. Reduce noise and increase dynamic range - mujltiple images are taken of a scene with identical exposure, focus and white balance. These images are combined and averaged to produce a composite image which has reduced noise and increase dynamic range. This is especially useful when images must be taken in low light with high ISO resulting in high noise for a single image but reduced noise when multiple images are averaged.
5. Super-resolution - multiple images are taken of the same scene at the same settings. These images are combined using an algorithm which can extract sub-pixel information resulting in a final image which has higher resolution than each single image. This technique relies on small variations between each image to interpolate between pixels. Another, more recent tequnique is to up-scale the image using “AI” techniques. Topaz Gigiapixel AI has pioneered this technique.
Most of these techniques can be executed in Photoshop/Lightroom or with dedicated apps which I will review later. However, a common requirement is that multiple images need to be accurately aligned in order to ensure the final image is of the highest quality, without artifacts. There are multiple ways in which to chieve this alignment - within each of the apps. I will attempt to review each option with a view to judging the best method for accurate alignment. Of course, the best starting point for alignment is to shoot the images on a stable tripod platform thus eliminating most of the mis-alignment of the static elements of the image. However, in many cases this is simply not practical (left my tripod at home) and so it becomes important that the alignment algorithms can handle normal image variations when shooting handheld. There are other factors which can effect accurate alignment:
1. Parallax - even small shifts in the position of a sequence of images can create issues if they are subject to parallax, that is, the relative position of near and far objects changes as the camera is moved, even slightly. Parallax effects can cause "ghosting" in the combined images which can only be elimininated manually. For panoramas, in particular, parallax can be a significant problem as the camera is swept across the scene. There are speciallize tripod adapters which can eliminate this error by allowing the camera to be positioned at the nodal point of the sweep thus minimizing parallax errors.
2. Subject movement - if there are elements of the scene which change during the duration of the multiple exposures then these elements will not be able to be aligned and will also produce ghosting of the final image. Some apps attempt to eliminate these errors but, again, this generally results in a manual process.
Let's look at what apps do what . .
HDR.
Lightroom
Photoshop
Aurora HDR-2107
Bracketeer (front end GUI to Enfuse)
Photomatix Pro (hdrsoft)
Panoramas:
Lightroom
Photoshop
PTGUI Pro
Focus Stacking:
Lightroom
Photoshop
Zerene Stacker
Helicon Focus (don't have this)
Reduce Noise, increase dynamic range:
Photohop
Super-resolution:
https://petapixel.com/2015/02/21/a-practical-guide-to-creating-superresolution-photos-with-photoshop/
Let’s look at #4, stacking and averaging images to reduce noise and increase dynamic range. Tom Hogan, several years ago, gave some suggestions on how to accomplish this:
http://www.sansmirror.com/articles/technique/reducing-noise-the-mean-way.html
Tom refers to Russel Browns “Stack-A-Matic” script, found here: www.russellbrown.com/scripts.html
Follow his instructions to download and install this script into Adobe Bridge. Also, refer toe here: http://uberplugins.cc/help/how-to-install-photoshop-script/ for instructions on installing scripts into Photoshop. For now, however, Stack-A-Matic only seems to work from Bridge CC (2019). From the Bridge menu, select the image stack then go to “Tools” -> Dr Brown’s Services -> Stack-A-Matic. This will open Photoshop from where you will see the Stack-A-Matic dialog box. Select Smart Objetc and Mean. check Auto align and 16 bit. Then go.
Below is an example of stacking and averaging. This is not an obvious one since it’s intention is to reduce, or eliminate, the atmospheric effects of capturing images over long distances on a hot day. In this case I took a series of shots of a harbor in San Francisco with my 500mm PF lens. Distance to subject was at least a mile and the weather was quite warm. The image below is a single image from the stack (of 10, handheld). This is a100% crop. You can see the atmospheric effects on the boat “Imperial”. Notice the black striping on the boat hull. It looks quite “wobbly”. If you look closely, there are othe signs of atmospherics in vertical lines, railings etc. Despite the warm weather, I was surprised that the effects were not worse given the distance captured.
The image below is the output from Stack-A-Matic, Sharpened with Topaz Sharpen AI (using the Stabilize option since there was slight movement of the boats which could not be corrected from image alignment alone). Notice that the “wobbly-ness” has completely disappeared. So, for static objects this can be an effective way to reduce atmospheric effects when shooting at distance with long lenses.
If you can’t be bothered with Dr Brown’s script, here’s, basically, what it does so you can do it manually:
1. In Lightroom, select the images you want to align and average together.
2. Go to Photo -> Edit in -> Open as layers in Photoshop.
Photoshop will open the images will be imported as a series of Layers.
3. In Photoshop, select all the layers. Go to Edit -> Auto Align Layers. Select Automatic (default)
Photoshop will then align the layers.
4. With all the layers still selected, go to Layers -> Smart Objects -> Convert To Smart Object. This will take a while.
5. When completed, select the Smart Object, go to Layers -> Smart Objects -> Stack mode -> Mean. Photoshop will then render the stack mean.
6. You can then flatten the image and save it. The image will automatically be re-imported to Lightroom.
Here’s another example. Again over 1 mile away using the 500PF lens. These before-after images are cropped to show the detail:
Original:
After:
Note that while the original had only the default LR sharpening, the “after” image required considerable sharpening - first with Topaz Sharpen AI, then also within LR using Diglloyd’s standard setting.