An efficient workflow to scan Xpan negatives

newtorf

Established
Local time
5:08 AM
Joined
Aug 26, 2011
Messages
180
Being an Xpan user for a long time, scanning Xpan negatives efficiently without sacrificing quality has always remained a pain point to me.

In the past, I used a Nikon Coolscan 9000 ED scanner with a glass holder to scan three Xpan negatives in a batch. The negatives were carefully cut into short strips of three pictures, and positioned parallel into the glass holder. For the first negative, I previewed in Vuescan, and selected an area slightly larger than the image. Hit the Scan button to scan the first picture, keep the Frame number to 1 with the Frame offset 66 for the second picture, and change Frame number to 2 with Frame offset 49 for the third picture. Scanning with Nikon 9000ED was painfully slow. And often I had to deal with Newton rings caused by the glass filter. The process always made me think twice before taking pictures with Xpan.

On the other side, scanning regular 35mm negatives with Nikon Coolscan 5000ED has always been a joy. 5000ED produces high-quality scans with up to 4000dpi and ICE. With an SA-30 holder (actually modified SA-21 holder), I was able to scan the entire roll in one batch, with a single press of the Scan button in Vuescan. Of course it would take an hour to finish the task, but it did not require my attention at all.

Unfortunately, Nikon Coolscan 5000ED does not support Xpan format. Many creative minds have come up with various approaches to scan Xpan negatives in two passes (e.g., [1], and [2]). However, often such approaches require a lot of manual intervention during scan, or was not entirely clear about their details.

So here is a workflow I figured out recently. It exploits the SA-30 holder to scan a whole roll of Xpan negatives in a batch, with minimal user intervention. It does two passes to scan the Xpan negatives, and requires software stitch after the scan. But it is a huge improvement compared to Coolscan 8000ED/9000ED workflow.

1. Feed Xpan negative to Nikon SA-30 film holder on a Nikon Coolscan 5000ED scanner.

2. Launch Vuescan, click Input tab, and turn off Frame alignment.

3. Click Preview button to get the preview of the first picture. Set appropriate Frame offset so that the beginning edge of the first picture is close to the edge of the scanning area.

4. Click Crop tab, set manual crop size, making sure that the crop area covers the full length of the scanning area.

5. Adjust all other settings as your wish.

6. In Input tab, check Lock exposure.

7. In input tab, set Frame spacing to 66.500. This is arguably the most important parameter to set!

8. Turn on Batch scan, and press Scan button to scan the whole roll for the first pass.

9. Increase Frame offset from its old value by 29.5. For example, if the old Frame offset is 3, you should change it to 32.5 (= 29.5 + 3).

10. Set Frame number to 1, and press Scan button to scan the whole roll for the second pass.

That's it. You will get two batches of images that are ready to be stitched in software. After saving the settings to a Vuescan profile, the only thing you need to do is to adjust the Frame offset after the preview.

SA-30 holder is hard to find on the market. Fortunately, there is an easy hack to turn a regular SA-21 holder into an SA-30 [3].

I hope this will not further push up the now already hefty price of Nikon Coolscan 5000ED and Xpan cameras. But who knows?

[1] https://www.stockholmviews.com/xpan-scanning/index.html
[2] https://www.l-camera-forum.com/topic...coolscan-5000/
[3] http://www.helmut-stoepfgeshoff.de/sa21-sa30e.html
 
This is an interesting approach to scanning the overlap. I have always considered it a bug that vuescan does not let you adjust the frame size and spacing so that frames overlap. It is not a limitation of the scanner. When I scan panoramic 120 negatives in a Nikon 8000, I adjust the frame offset between the scan passes. This step should not be needed.
I am surprised that you find stitching less work than loading the glass holder. Either way there is work involved.
 
For the glass holder, I was fighting the losing battle with Newton rings all the time for those curvy Kodak films. Lately I applied a trick by sticking a narrow strip of photo paper along the long edge of the glass holder, effectively creating some space between the top and bottom glass. But it introduced other problems like uneven lighting alone the edge of the strip. In the past I always thought it's some kind of light leak of the camera, till I scanned the films with 5000ED!

And as said, scanning with 8000ED/9000ED requires adjustment of parameters in vuescan for every picture, not to mention switching films for every three pictures. It was just too much.

The new workflow requires minimal attention. I just press the Scan button and walk away. After 30 minutes, come back and press the Scan button again. Stitching in software is much faster and is just a one time thing.

But of course, Vuescan should be able to combine the two pass scan into one automatically in a special mode. I will e-mail the author of vuescan, see if he is willing to do it in software.
 
The Nikon glass holder was designed to be used with masks for each film format. There probably never was an x-pan option. I make my own masks which I designed and laser print onto slightly heavy weight Letter/A4 size paper and then cut out with an x-acto knife. With 35mm the mask goes to the edge of the frame, and covers the sprocket holes. The LS8000 notoriously doesn't like the flare caused by sprocket holes. The mask holds the film nicely away from the shiny side of the glass, and the AN glass in the top lid presses it flat.
I also use a LS5000 with an original SA-30 holder for batch scanning. I do find that for short strips, or for the first/last frame in a strip it doesn't prevent the curl near the cut edge, resulting in unsharpness. For select images that have this issue, I rescan them with the glass holder on the LS8000.
Scanning always feels a bit like babysitting to me, there are some shortcuts, but it is inevitably time consuming.
 
I understand the mask solution. Had thought about similar approach but gave up when thinking about all the positioning work I need to do to make the film strip perfectly align with the masks. Just too much work.

Right now I am negotiating with Ed Hamrick, trying to persuade him to add a new mode into Vuescan for scanning Xpan. So far he is not convinced. :(
 
Back
Top