[soulsinsepia]

I read about Neat Image on the forum and downloaded a trial and subsequently purchased it. It is software for removing noise. So far, I am very impressed with the product. I am wondering though if those who use it can tell me when during the process of restoration that you use Neat Image. Do you start out with it or run it after setting contrast?

Thank you
Karen

[Kraellin]

generally speaking...very general, you want to start out 'finding the image'. noise removal software usually doesnt fit that definition. so, if finding the image is the first step, then things like curves, brightness/contrast, levels and so on are your first steps. once you've found the image then you can fix it, and that's where noise removal software comes in. exactly where in the 'fix it' stage probably depends as much on the restorer as anything else. and, of course, the image itself may also determine this. so, i dont know that there is any fixed stage where you would apply noise removal, but do recommend 'finding' the image first.

and here's why, noise removal software relies on blurring to some extent. that is exactly opposite of what you want in restoration, at least in the first stages. images are often smudged, blurred, dirty, stained and so on. your first task is to bring out what's really there. i mean, how can you restore if you dont know what's there? so, using a tool with the side effect of blurring is usually not the right way to proceed. and mind you, i did say usually. but generally, on a badly damaged image, noise removal is going to come later.

on lightly damaged images, noise removal may be all you need. the image is already 'found', so, no need for that step.

make sense?

[philbach]

I usually use Neat Image towards the end of the process. After the contrast and color have been restored and the spots removed. Using Neat Image the preview feature is helpful to decide what setting to use. After using Neat Image many times I will copy that layer and apply a layer mask to that layer and then sharpen that layer. Using the mask I will selectively sharpen areas that are improved by that sharpening like the eyes, eyebrows and things like that.

[soulsinsepia]

Craig,
Yes, you explained it beautifully. I understand completely and it was what I thought myself but couldn't have put it better. "Finding the image", brilliant!

Phil,
Why do you run the Neat Image several times and do you ever use the sharpen feature in it? If not, what do you think is the best sharpening process for old photos? In addition, NI seems to eliminate alot of the spotting, are you doing double work by doing the spotting first and then running NI?

Thank you both!

Karen

[philbach]

Well I only run Neat Image once towards the end of the retouch. I don't use the sharpening settings in Neat Image just because I haven't. Neat Image does not work well on big spots. So first I clear those up with various healing brushes and cloning. Neat Image works well on small more particulate noise like blue channel noise or other tiny artifacts. At any rate I've found that after using neat image coming back and sharpening selected portions of the image seems to work. To Sharpen I copy the layer and use overlay blending mode on the copied layer and Photoshop's High Pass filter. To that layer I add a mask.

Many times its just seems weird that you blur and image with neat image and then you sharpen again and the end result looks better.

[Kraellin]

soulsinsepia,

good, and thanks

[MatthewMarshall]

I use it before, then final clean up. the point of neat image is to get rid of jpg compression and digital artifacts, scanner defects and so on, its not a restoration tool, it wont get rid of tears or major problems the photo already has. Neat Image looks at all 3 colors layers and fixes the noise based off of that. Its best to do your Neat Image first then final clean up the image. Neat image should be a tool to get rid of "digital noise" and be only used as a small part of your restoration project.

Neat image is best used to get rid of noise that is the same through out the photo. See example below for what I considered is noise.

http://www.neatimage.com/examples.html