I got back for the lab a large set of negatives from my most recent trip to the
Llyn Peninsula. I was excited as I anticipated there may be a handful of decent images. I quickly set about scanning them. Since they were all taken with either my Fuji G617 or the Fuji GSW690iii I decided to scan them at 3200 dpi. (There is some debate as to the value of scanning beyond 2400 dpi on some scanners and higher resolution begets slower scanning in any case.) I had also recently purchase a BetterScanning medium format film holder and anti-newton glass insert as part of my drive to up my game in terms of scanning and image quality.
Focus Height
Before I could use my BetterScanning I decided I should make sure I had the film holder at the optimal height. There are a number of nylon screw feet that can be used to raid it from its default 1mm height. Finding a sharp image is the first challenge. I settled on my barrel cactus photo I used in the last round of experiments.
I took a series of scans at different heights in .2mm increments and tried to ascertain the best sharpness. It was initially easy but as I added height I had increasing trouble seeing a difference. I settled on 1.2mm height after going as far as 1.6mm.
Dust and Fingerprints
I have an Epson V500 scanner and I use Vuescan SW to run it. I also use ColorPerfect to import and color correct my negatives. On this trip I shot Portra 160 and Portra 400 film. I have a workflow on the scanner that is based on the ColorPerfect recommendations for raw 48-bit scans. It works well for me and better than Photoshop Elements.
Using the BetterScanning holder however I encountered lots of dust and blotches. The splotches I attribute to finger prints and so I carefully cleaned the anti-newton glass (ANG) with glass cleaner. This seemed to clear up that problem. The dust would require me to be more meticulous with wiping and blowing dust off the negative and ANG. The glass does add one more surface to keep dust free in exchange to flatter negatives. The film holder has another advantage in that it is longer and I can scan an entire 6x17 negative in one pass or two 6x9 in a single pass and singulate them in Photoshop.
 |
| Example Blotch Attributed to Fingerprint on ANG |
 |
| Rescanned with Cleaned ANG (dust still present) |
Scanner Color Channel Gain
Next I took another look at color channel gain as I had noticed using my Portra 800 settings (Red=1, Green=1.6, and Blue=2) seemed to restrict the histogram range for the Portra 160 I was scanning. For instance using my Portra 800 settings I get the following...
 |
| Red=1, Green=1.6, Blue=2 Histogram |
 |
| Red=1, Green=4, Blue=9 Histogram |
I did these scans at 400dpi to make them scan faster.
Scan Passes
Next I noticed noise in the images and had always wanted to experiment with added scanning passes. In making multiple passes the scanner averages the values for each pixel and so reduces noise for that pixel. Is it significant? Lets see.
I made 3 scans of the same image and tried to keep consistent color conversion as well. These were made at 1 pass, 4 passes, and 8 passes. This time at 800 dpi as the lower resolution was enhancing the noise effect. The noise is most apparent on the right in the water.
 |
| 1 Pass Scan at 800dpi. |
 |
| 4 Pass Scan at 800dpi. |
 |
| 8 Pass Scan at 800dpi. |
What is immediately apparent is that multi-pass scanning has a great impact on reducing pixel noise and smoothing out the graduated tones. It takes much longer though as well.
Also something I noted was the headland on the rock is full of blocked up shadows. I wasn't really worried about this as I thought I could solve that in Photoshop later.
However I then wanted to see what resolution did to the noise problem. I then made a series of scans (of a much smaller area of the film to save time) at 6400 dpi. A new series of observations emerged.
In setting up for the 6400 scans I ticked the multiple exposure box and got much less shadow blocking. So I went back to 800 dpi to see what happened on a single scan. Again no blocking. Further experimentation revealed the shadow blocking was caused by the Black Point Tails (BP Tails) setting in ColorNeg. I thought it was a scanner problem but if I turned off BP Tails for these negatives all of the shadow blocking disappeared..
 |
| Black Point Tails switched off in ColorNeg (1 pass scan) |
On to 6400 dpi... Here are the examples I have. Mind you these are at 660% zoom to get these results. The section I chose is a light section of sky where the mottling of the noise is more apparent.
 |
| 1 Pass Scan at 6400dpi 660% zoom. |
 |
| 4 Pass Scan at 6400dpi 660% zoom. |
 |
| 8 Pass Scan at 6400dpi 660% zoom. |
 |
| 12 Pass Scan at 6400dpi 660% zoom. |
Unfortunately there is an improvement at each level of more passes. It is also evident that scanning resolution makes the appearance of noise less which I suppose is because the eye will average surrounding pixels. Alas the scanner hangs when I tried 16 passes. Some limitation or bug in my setup.
One other thing I observed was the scanner made one last pass that did not reveal much of an image. Each previous pass shows the image as it is scanned. This is the multi-exposure pass.This should help with dynamic range and is different from multi-pass scanning. I may experiment more with this later.
Conclusions
- Make sure the ANG is very clean. Seemingly small smudges show up surprisingly well.
- A single pass with multi-exposure at 3200dpi will be my goto first scan flow for a new roll of film.
- I will use multiple passes for special images I need to scan for large prints.
Comments