I have been a NLP user for the last few years and it truly is an impressive piece of software. I have processed thousands of negatives through it and got my process down pat. Enter trichromatic scanning… for me it’s impossible to unsee the benefits for my workflow. For the past few months I have been using both but have recently fully made the switch. Honestly diving into trichromes has made me appreciate NLP even more. What it was able to pull out of some problematic negatives is impressive. Hats off to Nate.
Here is an example. On the left is a 95 CRI white light capture processed with NLP and simply edited using my standard process. On the right is a combined sequential RGB capture inverted manually with contrast tweaks but no color correction. Seeing what my negatives ‘actually’ look like has been eye opening.
I will make more edits to the right one and it may even end up looking like the left but the fact that the look isn’t baked in is what has been an unexpected joy. I also know that for a lot of people the left image looks closer to what you would expect to get back from a lab and that’s what NLP does so well. For me I am looking for more flexibility these days and trichromatic scanning has provided it.
Does this mean that you’re not using Lightroom? I suspect that it does. If that is true then how easy would it be for you to use your method to ‘process’ all the exposures in a roll using your method?
For whole rolls of 35mm you can use something like this. I had Claude build me a little applet that is super lightweight and just does the simple parts of the inversions with no editing capabilities. I’ll happily share it with you if you are interested.
4x5 and 120 is usually in photoshop for the added control it provides.
Thanks, that looks like a good potential workflow to adopt for those with a suitable light source. Is it dedicated to just one RGB light source, have you made your own? Only a couple of months ago I see you were using the Cinestill CS-Lite+ Spectracolor but presumably now you have changed from that?
There was the suggestion that Nate would be looking at accomodating RGB light sources in NLP, from Nate himself I think. Whether or not the new Standalone NLP version will do this remains to be seen, as is how it will manage without the basic functionality built into Lightroom regarding the Develop module, setups, syncing etc. On the face of it it is looking as if RGB light sources have the potential to make the colour processing much simpler.
The conversion process for the trichromatic scans is so basic that it doesn’t require the digital ‘magic’ that NLP was so innovative with. I could see there being some real utility to having the RGB sequential capture side of a standalone be the color science side of the equation. Profiles for frontiers, noritsus, etc.
I am using the big scanlight from jackw01. He has this great web app that makes it easy to trim the RGB intensities and automatically capture the three sequential frames. It even has IR if you can take advantage of that.
Has Analogue Toolbox improved since the beta version a couple of months ago? I tried it out, and while it wasn’t bad, it was a bit hard to dial in and kind of slow and clunky. I couldn’t imagine processing a whole roll or batch of multiple rolls with it.
NegPy also now has a trichromatic feature, but I have not tried it yet. I have the same scanlight and have just been using it for single exposures in C1 with great results.
I have never tried Analog Toolbox so I can’t say. NegPy is an amazingly capable open source project and is actually what the maker of my light recommends. If I was just starting out with negative scanning this is definitely what I would use. It has some amazing features which are similar to the ethos of NLP which help you get your scans to look more like lab results pretty easily. For me, I basically have my editing down pat so the features are less enticing. If someone asked me to dev and scan a roll for them I would be tempted to run it through NegPy just for the time savings.
Its true, all the algorithms in the world can’t compare to a physical hardware level light source.
The greens and reds especially.
I still find that while Noritsu scanners use this tri chromatic method, Noritsu still add artistic license in adding their own color science to the rgb scans.
I have to do the same to my rgb scans too.
I guess it’s down to whatever you prefer and gets good results.
One thing I found too is that rgb scanning is far more consistent
Is there a way to reliably demonstrate this though? I think that’s the problem, with colour negative there is no definitive way to say how the resulting image should turn out and so next to impossible to determine which software is best, or (with camera ‘scanning’) which light source. Or is there? With colour transparencies things are much easier, you can see what you are aiming for, and you can do a colour profile with an IT-8 target if you wish to.
Wouldn’t it be true that with a Noritsu or Frontier scanner the ‘look’ is as much to do with the settings that the operator decides to dial in but from that point perhaps there is more consistency?
To me that would have to be the key to getting good results and would certainly push me towards using an RGB light source if it provides that, just as a dedicated film scanner should. In the darkroom you absolutely will get consistency across the images on a particular roll because it depends solely on the film itself, the colour filtration that you dial in and the exposure. I could add the type of RA-4 printing paper but in reality most people would I think stick to the same brand and there is very little choice now, though it’s a long time since I’ve done any colour printing in the darkroom,
With NLP perhaps the key to handling RGB light sources would be to disable a lot of the background colour processing between one image on a roll and the next.
I would say that the right way to think about RGB scanning is that you are reading the density of the RGB color information, rather than imaging the film. Having the tuned light source is like having RA-4 paper because that is what the designer of the light aimed for. None of it is perfect, as you might expect, and the ‘right’ way is to use a monochrome sensor so that you get zero crosstalk. But I agree that my results have become more consistent.
NLP is so good that images shot in similar light all look very close to one another. The issues I have had is the variation across a roll. RGB scanning will not have this variation because there is no ‘roll analysis’. Just like in a color lab, everything is made to work together so all you have to worry about is what ‘digital filters’ you are going to use to get the image where you want it.
If you have worked in a color lab and enjoyed it, you probably will like working in NegPy because the digital workflow was made to model a lab.
Last thing, yes the magic of the frontier and noritsu scanners is not the RGB scanning with a monochrome sensor. That just gives you the accurate baseline scan. It’s that interpretation they do that really matters. So I agree, there still is utility in a software with profiles for popular scanners and with the ability to make presets. Maybe NLP will be it, maybe not. I hope Nate gives it a shot because I have always loved the colors that come out of NLP.
Just wanted to add to this for those who are interested. I found a good example showing the difference in accuracy that I get from different capture and inversion techniques. Left is 95cri white light converted with NLP, center is a single capture narrowband RGB converted with NegPy .38, right is sequential trichromatic narrowband RGB combined with Negpy .38.
The inversion on the left looks pretty good until you look at the one on the right. The left has the signature red cast that NLP can sometimes get. The center image just looks overcooked to me, reducing the saturation may help but everything looks a bit heavy. To me the right looks open, airy and accurate to the way I remember the scene. It also has the least edits done to it after conversion. Only black & white point setting and less than 5 points of white balance adjustment.
Its point is to provide lighting with many more colours in order to accommodate the different colour processes found in old film. Going one step further would be possible with tuneable narrowband lighting (tuneable laser) but even the paper’s experimental setup looks like going beyond reasonable effort for “household” use imo.
One idea that might help deserves looking into though: Adding yellow lighting to fill the gap between red and green, A gap that most camera sensors bridge with a fair amount of overlap of sensitivity in red and green sensels respectively. Skin tones are in this area too, so adding some yellow might help with these too.
Scanning with CMY and converting that to RGB bight work too. Without the conversion, CMY is not helpful, I’ve tried it a while ago.
How do you get the trichromatic images out of NegPy? I just started playing with it, and do not get the color management at all. I cannot select any output profiles, and the colors are ridiculous in the exported tiff (opened in PS or C1). I’ve tried every setting in the export panel.
I find the interface of NegPy to be impenetrable. It already has so many features and they keep adding new ones at what seems to be a weekly pace… I wish there was more focus on refining the existing tools than adding new ones. Also, it is very slow on my M1 Mac, so speed improvements would be very welcome.
My method with NegPy so far has been to use as few of the tools as possible. Basically get in and out as quickly as possible. My steps are:
Import
Set Analysis Buffer so it is only on the image information (I ignore the crop tool in NegPy because it is so poorly executed so far, cropping is done later in Lightroom)
Copy ‘Process’ and across all images
I take a look at the conversions and make a decision to either export with the auto-processing NegPy has done
If I feel like the processing I go to the export tab and select ‘Print’, format is TIFF, size is Original, color space is sRGB
If I don’t like the processing I go to the export tab and click ‘Flat’ and export flat TIFF’s in sRGB
Import into Lightroom for cropping and final editing
I am using version 0.38 and I believe they are already on 0.40. For me the output color space is under ‘COLOUR’ on the export tab.
I am on 0.39 and can get the image exactly how I want on screen, but the color management is broken. I have no output profiles in the export icc dropdown and the exported tiffs are untagged and not remotely like the image inside NegPy if I assign Adobe or ProPhoto. I reinstalled and still no good.
Not only is the NegPy UI frustrating (in the way many linux centric UIs are, and I say this as someone who spent many years using Linux back in the day), but the rapid pace of change makes it impossible to dial in any sort of workflow because even if you get something you like today, in the next release it will probably turn out different. Active development is great and in their defense they have not made any claims that the current version is a remotely stable release, hence the sub-1.0 version number, but it makes it very difficult to use productively if you actually care about photography rather than playing with software IMO.
