Issue with Photo Merge & DSLR scans

Hello!

I recently built a rig for taking multiple images of 4x5 sheets with my dslr and a macro lens. I've been experimenting with different levels of magnification and I'm running into some issues.

At first, I made a stitched scan using 9 exposures at 2:1 magnification, resulting in an image that was about 190 megepixels.

Then, just for funsies, I created a scan using 49 exposures at 1:1 magnification, resulting in an image that was 600+ megapixels. I realized immediately that this was overkill, but I was impressed with Photoshop's ability to successfully merge 49 photos.

Next, I aimed for a middle ground: 25 images at 1.5:1 magnification. I used a different sheet for this run, and for whatever reason, Photoshop is unable to make sense of it. Lightroom successfully merges the exposures but Lightroom doesn't have a "reposition" mode so the boundaries of the image are severely warped.

The rig I have is designed to capture each image with a 40% overlap between each row and column (as per Adobe's recommendation) so I really don't think overlap is the issue. The only thing I can think of is that most of this sheet is out of focus (selectively) and maybe it needs sharp content to compute the merge?

Here's what it ends up looking like:
Attachment 224565
Any ideas?

Does this 4X5 image have a lot of clear sky? That can often be a problem with photo merge in PS when you have a great many image files. What happens in this case is that many of the adjacent areas in an image with a lot of sky or other clear areas have nothing with detail to merge to.

Sandy

Just a guess, but maybe at 1:1 the resolution is enough to pick up the film grain, so the software can use that for alignment. Maybe at 1.5:1 the grain isn't quite large enough, and the fuzziness of the image itself is preventing alignment.

Are you doing manual focus with your macro lens, or autofocus? I've been doing experiments with a macro lens at 1:1, but I'm using autofocus for each shot. If the image itself is fuzzy, that's going to make both manual and autofocus of the camera difficult if it can't focus on the grain. I've heard about the big sky issue, as sanking pointed out, but I don't know how that can be resolved easily.

Quote:

---

Originally Posted by Schuster_Shots

Here's what it ends up looking like:
Attachment 224565
Any ideas?

---

I've had the same thing happen to me at times. I was overlapping by about 50%. For the heck of it, I reshot the images with only about 20% overlapping, and the photo merging worked fine.

Personally, a "middle ground" of shooting 25 images is an overkill for me. When I had to make 24x36 inch prints of some 4x5 Chromes, I found the sweet point to be shooting 4 overlapping images. Overlap was less than 20%. Biggest factor was the alignment of the sensor, lens, and the 4x5 Chromes. I used a Nikon D850 on a Nikon Multiphot which assured near perfect alignment. Lens was a 12cm f/6.3 Macro Nikkor (barrel lens) at its No. 3 aperture. Opening up or closing down by just one f/stop noticeably affected the image resolution. The 12cm f/6.3 Macro Nikkor was noticeably superior to Nikon's 60mm, 105mm, and 200mm macro Nikkors. Enlarging Nikkor lenses were inferior to any of the above optics to my surprise. My guess is because my enlarging Nikkor lenses date from back to the mid 1970s.

Quote:

---

Originally Posted by sanking

Does this 4X5 image have a lot of clear sky? That can often be a problem with photo merge in PS when you have a great many image files. What happens in this case is that many of the adjacent areas in an image with a lot of sky or other clear areas have nothing with detail to merge to.

Sandy

---

No sky at all. Pointed right at the ground.

This a good hypothesis.

I'm focusing manually using a a Manfrotto 454 rail. I use a piece of printer paper with a millimeter scale printed on it to focus since the the images I've taken so far haven't been quite sharp enough to tell when they're in critical focus when magnified. I feel like the paper is a similar thickness and grain can be seen in the resulting scans so I'll probably stick to this method.

Quote:

---

Originally Posted by sharktooth

Just a guess, but maybe at 1:1 the resolution is enough to pick up the film grain, so the software can use that for alignment. Maybe at 1.5:1 the grain isn't quite large enough, and the fuzziness of the image itself is preventing alignment.

Are you doing manual focus with your macro lens, or autofocus? I've been doing experiments with a macro lens at 1:1, but I'm using autofocus for each shot. If the image itself is fuzzy, that's going to make both manual and autofocus of the camera difficult if it can't focus on the grain. I've heard about the big sky issue, as sanking pointed out, but I don't know how that can be resolved easily.

---

This a good hypothesis.

I'm focusing manually using a a Manfrotto 454 rail. I use a piece of printer paper with a millimeter scale printed on it to focus since the the images I've taken so far haven't been quite sharp enough to tell when they're in critical focus when magnified. I feel like the paper is a similar thickness and grain can be seen in the resulting scans so I'll probably stick to this method.

Quote:

---

Originally Posted by Greg

I've had the same thing happen to me at times. I was overlapping by about 50%. For the heck of it, I reshot the images with only about 20% overlapping, and the photo merging worked fine.

Personally, a "middle ground" of shooting 25 images is an overkill for me. When I had to make 24x36 inch prints of some 4x5 Chromes, I found the sweet point to be shooting 4 overlapping images. Overlap was less than 20%. Biggest factor was the alignment of the sensor, lens, and the 4x5 Chromes. I used a Nikon D850 on a Nikon Multiphot which assured near perfect alignment. Lens was a 12cm f/6.3 Macro Nikkor (barrel lens) at its No. 3 aperture. Opening up or closing down by just one f/stop noticeably affected the image resolution. The 12cm f/6.3 Macro Nikkor was noticeably superior to Nikon's 60mm, 105mm, and 200mm macro Nikkors. Enlarging Nikkor lenses were inferior to any of the above optics to my surprise. My guess is because my enlarging Nikkor lenses date from back to the mid 1970s.

---

Hmm. Maybe I'll reduce the overlap. 40% might be overkill. I tried 10% at first and that didn't work at all. I'm also thinking that the 2:1 method is more than enough resolution. It's honestly hard to tell the difference.

For reference, I'm using a Nikon D810 and a Meike 85mm f/2.8 macro lens.