Well look this basically comes down to opinion and artistic choice. It’s ok to have a different opinion than mine I assumed when I said it there would be other opinions and neither of us are wrong or right.
There are just a couple of technical points here though.
My advise to use 2 but no more than 3 photos for HDR is a technical matter not an artistic matter. DJIs AEB only allows .7 stop increments so with a 5 photo AEB you get -1.4, -0.7, 0,+0.7, and +1.4 exposures. Adobe specifically says this is pointless and subjects your photos to unwanted artifacts and ghosting.
This is from Adobe’s Manual on Lightroom.
Learn how to easily merge photos into an HDR, panorama, or HDR panorama. Also, learn the requirements to create an HDR panorama.
helpx.adobe.com
“Create HDRs
How many photos are required to process a quality HDR merge?
Read these guidance notes and tips by Rikk Flohr (Software Quality Engineer for Lightroom ecosystem of apps, Adobe).
HDR photos are used to capture scenes having a large dynamic range. However, using more number of photos can lead to unwanted artifacts from poor alignment or ghosting. For optimal HDR merge, the aim is to capture photos in a manner that each part of the scene is well-exposed, that is neither blown-out nor under-exposed in at least one of the photos.
Use the following guidelines to identify how many photos work best for your case:
- If your HDR bracketing is less than 3.0 stops in total separation (-1.5, 0, +1.5), use only the darkest and brightest exposures to generate an HDR. Capturing the middle exposure, or zero exposure, not necessary for generating a quality exposure blend in such cases. If you exceed the 3-stop separation between the darkest and the brightest exposures, an additional exposure offset becomes necessary to process a good quality HDR photo.
See the chart below to quickly determine the number of photos required for processing a quality HDR merge
Camera Bracket settings | Optimum number of exposures for merging photos to HDR |
---|
-1.5 to +1.5 | 2 |
-3.0 to +3.0 | 3 |
-4.5 to +4.5 | 4 |
-6.0 to +6.0 | 5 |
End quote”
So a photo with an exposure that is less than 3 stops from another photo is only introducing unwanted artifacts and does more harm than good. Don’t take my word for it...take Adobe’s.
The technology is such that yes you can still get pretty good images from these 5 exposures but at best Adobe is smart enough to discard that extra information and you’ve only wasted time having Adobe figure out what photos to discard and at worst you introduce unwanted artifacts into the photo.
I believe DJI’s metadata will only register -3 to +3 stops so at most you could manually take three exposures -3,0,+3 and merge those in Lightroom hence where the at most three photos comes from. Lightroom looks at the exposure value in the metadata to merge photos and the exposure values have to be equidistant from each other or Lightroom will not merge them. To be clear you couldn’t use the AEB function to do this.
Next the faster your shutter speed the sharper your image. An ND filter will restrict the amount of light to your camera forcing a slower shutter or increased ISO both of which reduce your sharpness on the entire image not just the clouds. What you really want is a graduated filter that is specifically for reducing the exposure of one part of the image without reducing the shutter speed. There are lots of ways to easily soften, blur, or reduce the sharpness of areas of the image in post but getting a soft photo to become sharp is an arduous process if it can be done at all.
You know your way around Lightroom and photoshop enough that you could even replace the sky completely with a different photo.