Hey everyone, looking for some feedback if anyone else has gone through the same process as me. I use a Nikon Coolscan V to scan my 35mm negatives, with Nikon Scan 4. My question pertains to scanning black and white negatives. So far I’ve been scanning black and white negatives as grayscale, which creates a 16bit grayscale tiff file (40mb or so). I noticed however in the conversion page of NLP that it mentions taking “16bit RGB files”.
I got curious and thus tried scanning my black and white negatives as “calibrated RGB” which creates a 48bit RBG tiff file (130mb). Yes I understand that there is a difference in the file size, but to me thats not super important. It seems as if on some images, the RBG file produces a noticeable increase in detail versus the grayscale file. On others it seems imperceptible.
I am wondering if anyone else has gone through this experimentation and what their conclusion has been. Is NLP meant to use RGB tiff files as its basis, and is it possible that using grayscale tiff files might be reducing some tonal detail as well?
Any advice on this would be great. Thank you everyone!
B&W film responds slightly differently to different wavelengths, even though manufacturers try to make the response differences as minimal as possible. Depending on where details can be found along the wavelength axis of the response curve, details are attenuated or amplified as you have noticed.
From a more general point of view, all transformations incurr some loss and it is therefore good practice to start with as much info as possible (rgb) and convert to monochromatic in one of the last steps of customizing the image, just before the “last polishing”.
Thanks for the response and the welcome :). Thats what ive been noticing. Its definitely me nit picking, the detail for example in the grain can only really be seen when zooming in past 100%. But since I’m still experimenting with scanning on my own, it’s been kind of difficult to notice those differences, or if its just me going mad (“like is there really a difference??”).
Anyways I think your argument makes sense, starting with the most data from the start! I think my thinking was wondering if NLP was better at taking RGB tiff files or grayscale tiff files in its conversion process.
Basically, you don’t need NLP to convert B&W negatives, no matter if they are single or triple channel files. Just flip the tone curve and draw the white and black points to where you want them…but NLP does that automatically, which is a great relief for batch conversions.
I’m pretty sure files from cameras DSLR or Mirrorless are RGB even if you set the camera to B&W. I used to change them to greyscale to save disk space, but don’t now. The main reason is when sent to a printing company they print a lot better sent in RGB. The printing software changes them to CMYK to print, not just K.
I include B&W with colour photos in “Keynote” presentations and found that it is most important to calibrate your computer screen to assess the final colour of your colour photos. B&W RGB files don’t matter, you just set colour saturation to zero. The projector/screen that you show them on should also be calibrated.
What used to happen to me was either the B&W images had a noticeable colour cast, or the colour photos did because my computer screen had a green cast which I didn’t realise, my eyes adjusted to it! But when showing mixed colour and B&W photos on another screen the difference was most annoying.
I appreciate the feedback Graham! Indeed I did think about the fact that digital cameras take images that are RGB, meaning no matter if you convert a black and white negative or a color one, the input file will be the same. Thats most likely why NLP asks for a RGB tiff file as its basis.
About color casts, what’s interesting is that no matter if I convert from a grayscale tiff or a RGB tiff, the resulting conversion in terms of color cast is identical. I tend to notice however small differences in detail present in the grain for example.
I’m still honestly not sure what is best. It seems as if the difference is certainly minute.
…abd this is completely okay, no one knows what is best, but there are a lot of opinions about it.
From a technical point of view, the results need to fulfill certain requirements and as long as one hasn’t establish these, we cannot say if results are good or bad. Requirements differ greatly if you compare web output to 6 ft prints that need to be seen from as close as 2 inches away.
