I realised it has been nearly 2 months since I sent feedback to DxO about PL9’s new Scanned Film Optimisation (SFO) panel. I guess they may have been busy fixing bugs and adding new modules.
The problem I reported was that if none of the provided preset curves get close, then manually achieving good colour is a real chore.
I’ve done little modern photography lately, so I decided to have another play and I think I got this one looking pretty decent without too much effort.
Set the white balance using the eyedropper on a part of the film surround.
Set SFO on.
In the SFO panel, for each of the Red, Green, and Blue curves, drag the bottom left towards the right and drag the top right towards the left — leaving both in their original vertical position (bottom and top respectively). This is done to get an initial “eyeball”. It may require revisiting each colour several times.
“Dial in” the top curve points by hovering the mouse over a neutral highlight to see whether any colour is dominant or subservient and adjusting the corresponding point appropriately. In this image I had the white of the roundel, plus the clouds.
“Dial in” the bottom curve points in the same way using a deep shadow. In this case, the propellor blade.
Use the master RGB curve (in SFO) to create a subtle “S” to see if it will improve the colours. It did in this case.
For each of the R, G, B curves, grab the midpoint and bow it up or down to see if either direction improves the colour.
That got me to the colours you see here.
I also played around to see what I could do with the very obvious grain and discovered I could tame it significantly. (Something I had not previously figured out.)
Set Microcontrast to -25 (or thereabouts). This had a huge effect on the grain.
Turn on Standard noise reduction
Set both Chrominance and Low freq. sliders to their max.
Play with the Luminance slider to taste. Too high makes it look like a painting. To low leaves the grain.
I also discovered ClearView Plus quickly brings the grain back into focus. It could stand a value of 5 but any higher was detrimental. (On native digital photos, I will occasionally go as high as 15-20.)
In theory I should be able to copy most of those settings to every other frame from that film, but previously when I have tried this, the results were mixed. I’ve not yet tried it with this one.
I tried this inversion but left it quickly. I have used VueScan for more than 25 years and it has always given IMO better inversions and also colour fading compensation. Once I’ve saved tiff formats with it I then use PL at taste, with some of the latter settings that you mention.
VueScan has its own issues. When I finally figured it out, I wrote a blog post about how to successfully set up multi-frame scanning. I know I was not the only one who had trouble figuring it out because that post has been, by far, the most visited post on my blog since the day I published it. Very few people visit my site… except for that page.
The other thing about flatbed scanning is it’s sloooow compared to the “DSLR scanning” approach. I wrote a blog post about this, too. Add in, in my case, scanner software that is touchy when doing high resolution scans and enormous TIFF files and I find that the inversion is the only part of the process that is easier that way. I also found VueScan’s success on inversion wasn’t guaranteed. Sometimes it really messed up if shots were particulaly under- or over-exposed. That, I think, is one of the downsides of automation.
Now, if you have a proper film scanner, then speed isn’t going to be an issue, but then you’ve demonstrated a commitment to spending money I cannot bring myself to do. I spent a few hundred dollars on some gear that I can (and do) use for other tasks, and use my DSLR which, of course, I also use for regular digital photography.
If I really don’t want to fiddle with inversion in PhotoLab I can always use NegativeLab Pro in Lightroom Classic, but then I am back to dealing with enormous TIFF files. The beauty of an all-PhotoLab approach is that it is 100% non-destructive, non-constructive (no extra files, tools etc needed), and the only extra you end up with is the final output.
As I have many negative positive films I bought a Canon scanner before 2000 (replaced by a reflecta RPS 10M) and a flatbed scanner with the ability to scan large negative and positive films and plates, an Epson one. That’s not cheap but it’s so satisfying to get good results from old films… The oldest negatives I processed are from the 1900s and the first slide I scanned is from 1924 (autochrome).
Erasing the scratches automatically is also an advantage of VueScan (or other similar softwares).
