Post-processing q: removing colour noise from scanned image

[tgray]

Advertisement (gone after registration)

And if you aren't making large prints, then it's fine. Plus you can always try scanning at a lower spi to help reduce visual grain if you don't need the bigger file size.

 

On my Coolscan, I've found that doing that makes film look MUCH grainier. Much better off scanning at 4000 dpi and reducing the file afterwards.

[CalArts 99]

On my Coolscan, I've found that doing that makes film look MUCH grainier. Much better off scanning at 4000 dpi and reducing the file afterwards.

 

You're quite right.

 

I mentioned scanning in a high spi and then downsampling in my earlier post, for better results. And for larger printing that means going to 6k or 8k spi (on a PMT scanner) and then down to 4k to retain file size for a larger print.

 

What I meant about scanning at a lower spi on a consumer Coolscan was based on needing non-printing files, i.e., web, etc., where the image is so small anyway that you may very likely not see the issues on a display. And a piece of small 35mm film at 2k will be a pretty small print and ink dots are not exactly comparable to pixels (plus there's the way the dithering and screening works, etc., etc.) But no one has ever really mentioned what they are doing with their files specifically.

 

Anyway, sorry for the confusion.

 

I think people should maybe try working backwards: what is the final output, visualize what it should look like for your needs/desires, and then 'walk backwards' to reach that goal (starting at the point of initial image capture.) That's what I do. Does that make sense?

 

I've personally never had a file (or film format and type) that can do everything at once. There are just so many variables (pigment ink, dye ink, carbon pigments, chemical RA prints, etc..) Even different printers have different issues, e.g., the Lambda has a better algorithm for interpolation than you can do in PS, etc., etc.. But I do realize the temptation (and often the need) to get a file to where it can become a sort of 'do everything file.'

 

Anyway, my head can sometimes spin chasing these things. I usually find it's better for me to use my time and effort on the project itself and capturing the images, and just let a professional studio do the rest.

[tgray]

Got you.

[plasticman]

Love the images by CalArts, incidentally - the scan of the storefront is incredible, both technically and aesthetically.

 

Anyway, about the color noise in negative scans. I haven't printed any film images really big yet, (I'm waiting right now to see whether Epson has a new printer on the way, to replace the 3880), but it seems to me that a lot of this worry about pixel-level color noise is due to the usual 100% screen-view obsession we all carry over from digital files.

 

In a tutorial I read somewhere (forgotten where I'm afraid) about using Noise Ninja correctly, I remember it stated that the user should zoom out to 50% in order to assess the effect of any noise reduction. Zooming-in closer than this was both counter-productive, and led to more aggressive processing than was necessary or desirable.

 

As I said before, I really like the textured grain of film, and I'm willing to bet that even some of these pixel-level color artifacts will either be invisible, or possibly even enhance the texture of most images when printed large.