[microlight]

Thanks, M2 - it was the post that you linked to that turned me on to this method a while back, and have used it since.

[emalvick]

I am going to be in the bin of less is more. Now I don't the details of the noise you are getting since you are scanning. Is it noise because of grain in the original film or is it from the scanning process? I would probably work at removing the noise due to scanning and worry less about that from grain.

That being said, it is important to think a bit about what you will be doing with the images. The resolution for images is quite high these days, and noise removal is easiest to perform when you are looking at the image at a 1:1 scale on your monitor. However, unless that 1:1 image represents your usage of the image (I would guess this is rarely the case), then it is probably overboard to try and remove all the noise at the level. Doing so is how you can end up with a lot of blurring and unnatural aspects of an image.

These days I actually perform noise reduction at a different scale and usually just go with a default amount. I got to this point because if I am using an image for the web or screen viewing, I am mostly looking at an image at such a wide scale that even noisy images might not appear noisy. Or, if I am printing, I rarely see the noise at the scale I print at (and I actually did some tests using a noisy image and various scales of noise removal).

All that aside, I also find noise removal is most necessary in parts of an image where things tend to be more uniform (i.e. away from areas of detail). At least that is where our eyes go to. For instance the sky, a smooth wall, a macro type photo where you have a large leaf or flower petal, skin in a portrait.

So when I need more noise removal, I take whatever software I am using and work with Affinity Photo (Photoshop, GIMP, or any other layer enabled software could be used):

1. duplicate the image onto multiple layers (usually 2 is enough)
2. apply noise removal on the top layer with an emphasis on the smooth areas
3. User layer blending and/or masking to reveal/hide the areas of the image so that you don't lose the details you want to keep but you get the most noise removal out of the areas where detail is less important. This is kind of what the mask part of the LR noise reduction does with a bit more science to it. The method here is for more extreme cases such as when I really want to clean up the sky.

Software itself, most noise software these days will work ok. I like the Topaz denoise, and it is what I use when I've gone to Photoshop for the effort. Admittedly, I moved over to DxO Photolab Elite, and their Prime noise reduction is my favorite, but that is part of a RAW converter that usually costs closer to $150 (and I wouldn't pay for it and LR).