I’ve been using NLP for a few years now, and I did come across a few photos that gave me a hard time to get them to look as desired, but never like this photo… I developed this roll in a Lab, cinestill 400D @ speed box and I’m scanning them like I usually do, 5DMKIV mounted into a copy stand, Canon 100L Macro F2.8 and I’m using a Neewer Ultra-Slim LED Video Light, 40W 3200K-5600K CRI95.
The whole roll basically has this strong color cast, I scan another roll of same film stock I dropped at the lab same day and it seems to be fine. The negative looks fine (for me) I compared with other rolls I have here I can not see any difference in between them. PLEASE HELP
There is almost no information in the shadows and the image consists of greys and yellow/red tones. Not much for NLP to go by. I tried several settings but none would do any real good, so I came up with this:
Convert with NLP (using Basic and pre-saturation = 1)
Added masks for the sky and the tarmac and used WB to tune them to taste.
Occasionally, NLP can’t handle difficult (thin, not much colour variety) negatives and then, some post-processing is needed. In this case, the two masks would be all it took. Note that the sky is a bit too blue and bright for the time of day … but the whole exercise is just a proof of concept.
Hey mate, thanks heaps for taking the time to look into this and for the edit, really appreciate it.
I’m still a bit confused why this roll came out like that though, the negatives look fine to me, so I’m wondering if it could possibly be a development issue?
I’m thinking in getting them scan at my local lab here to see if they can get them look better with a proper scanning set up.
@Valientefer , if you have your film lab scanned and printed, bear in mind that what you get might have some manual tweaks in order to „improve“ the results.
NLP can deliver more balanced results with images that contain r, g and b color components with wider dynamic ranges each. Images with narrow histograms tend to get casts which get stronger when not all color components are similarly present. Such cases often call for manual interventions and that might not change in lab scans, unless they are tweaked by the operator (who then does the manual intervention)
Nevertheless, a lab scan will be different and this difference might be just enough for NLP to deliver more consistent conversions. Please post your findings.
Completely agree with @Digitizer . The original simply lacks all details in large piece of shadows. on top of that the frame has not been scanned properly with obvious light un-uniformity on lower left side. So edit it to taste but this is not NLP fault - overall its most likely underexposure or bad development . Sorry for breaking bad news
