I’ll start by talking about noise in photos… I think the easiest way to visualize this noise is to think of the static you’d sometimes see on an old analog TV. Film negatives have some noise, because they work by having chemical crystals stuck on their surface, & after developing those crystals form film grain. Higher ISO / ASA film has more visible grain. I *think* the optical enlargers that used to be common were kinder, minimizing visible grain / noise compared to a high-resolution scan nowadays. The photo paper they used for prints works similarly. Digital scanners and cameras also add noise – how much depends on the make / model – which actually could be called static, as in unwanted stray electrical signals / interference. With a digital camera you can increase the ISO, which increases the sensitivity to light by amplifying the electrical signals from the photo sensors, and that adds noise – the higher the ISO the more noise, and again how much depends on the make / model.
Depending on what you intend to do with a photo or photos, you might be able to ignore any noise entirely – if there’s an extreme amount of noise it can make a photo look terrible, but most of the time if all you want to do is look at the photo, you’ll never notice it. [I have a fairly high-end Kodak camera from years past, a Z915 – while OOS, Walmart lists it for $240 – and it’s Only capable of pretty noisy photos] Enough noise can make an image look not quite as sharp, but that’s usually about it. Where it becomes a nightmare is in editing those photos. For one thing, all those noisy pixels can be unpredictable, standing out when you adjust the whole photograph. For another, it can make blending a fix or edit impossible – you can’t duplicate that random noise so that it fits in with the surrounding area. It can also make selections, or anything depending on a selection or selections, hit or miss… take a post-It note & tear it in half [no creasing 1st – that’s cheating]. That irregular edge down the middle, where you tore it, is something like what the edges of objects [people etc.] look like in a noisy photo, and you can spend all the hours you can spare trying to perfect that selection and still not get a good result.
So, until recently all you could do was read reviews about whatever camera or scanner before you bought it – good camera reviews also cover the noise you get at various ISO settings with example images. Software could do a little bit – IMHO Franzis Denoise was one of the best of a pretty poorly performing lot – and you always risked blurring, loss of detail, and sometimes artifacts. Compared to that, AI photo denoising software is Revolutionary. It really is that good! Ideally you want to use it as a plugin for a good image editing app with layers – that way you can use parts of more than one copy of a photo if/when there’s a glitch [which can happen], or more likely, you have to turn the strength Way up for one part, but that turns out to be too much for the rest of the photo.
I was introduced to AI noise removal with the DVDFab Photo Enhancer AI that was on GOTD. And I gave it a pretty good workout, using it on a couple few hundred photos. And now I can honestly say that if that’s all you can afford, go ahead and use it – otherwise buy a competing product from ON1 or Topaz, uninstalling the DVDFab app to reclaim a few GB. When it works, the DVDFab app works well, but it’s unfortunately prone to leave smaller areas of the photo either untouched or distorted, so you get rid of one problem, noise, and create other problems you’ll have to deal with. There were also about a dozen photos -- parts of different batches I’d previously scanned – that crashed the app. There was nothing unusual about them, and photos scanned before and after worked just fine. The offending photos work just fine in other apps too.