For a test, just a test, I set my new Fuji X-T3 ISO to 12,800 the other night. I was surprised by the results – the pictures were usable. Yes, grainy and there is a loss of resolution but still, usable – more on that later. While working on the images, I discovered some differences in the noise reduction capability of the different raw convertors. So, I decided to research this area a little. And, away we go….
I selected the image above for my ‘study’ – it features lots of noise in the background and the resolution has obviously been reduced. First, I wanted someway to compare images. While this is a less than perfect method, I converted a raw image using the camera. The only option is to convert to a jpeg but at least that gives me a ‘benchmark.’ Even with that, now I have to compare screen captures – rendered in jpeg – and present those for comparisons. Again, less than perfect.
To start, I processed the image above in Lightroom, then exported it as a jpeg. That gives me an idea of how Fuji’s conversion compares with Lightroom. Left is the LR version. (remember – click on an image to see a larger version) While at this resolution it’s difficult to see a difference, on my monitor the LR version is slightly better in resolution but has a little more grain.
Next up, reduce noise and sharpen in Photoshop. Here’s the result – 100% crop( the in-camera conversion on the left in the following comparsions):
The Photoshop version has more grain but slightly more detail. Lightroom and Photoshop yield close to the same results.
As you may recall, in my comparison of raw convertors, ON1 came out ahead. So, let’s take a look at ON1 and see how it does with some noise. Here’s Fuji’s conversion vs ON1: Uh, not so good. In fact, the smaller image and the screen capture make ON1 look better than it does on my monitor. While I have not used ON1 a lot, I was not able to really reduce the noise – dragging the sliders around didn’t seem to change much at all.
OK, what about Luminar then? Here’s that comparison: In a surprise, Luminar does better with noise reduction and is OK sharpening this particular image. It would appear you should select a raw convertor on a per image basis. Hmm.
So, where to go with this? One thing to consider is a statement from MattK – I’m paraphrasing him – …”noise isn’t a problem when printing.” ( One reason I’m paraphrasing here is I can’t find the place he said that. Hope I’m not making this up.) So,I printed a 12×8 print of the Lightroom version. Viewed at a “proper distance*” he’s correct. If you play pixel peeper, yes you can see the noise. So, I conclude if your goal is printing, some noise is acceptable. Even better, if you just want to put the image on-line, almost anything would be ‘acceptable.’ Meanwhile, so far as choosing raw convertors, if you subscribe to the Adobe apps, use those. If you really don’t want a subscription, well, maybe ON1 for now with Luminar a close second. I see Luminar closing in on ON1 but they are not there yet.
* Proper Distance means standing 2 to 4 feet for a print that size. Or not. What it doesn’t mean is putting your nose on the print. That’s an easy way to pick out the photographers viewing an exhibit.