We've been warned against it. Don't pixel peep. It's likely to disappoint and there are good reasons why an image can't always be razor sharp or without noise when zoomed in to 400%. Yet, the temptation is there. Especially when you've just bought a new camera that is supposed to deliver sharper images with less noise.
As mentioned in a previous blog post, I was fortunate enough to be able to upgrade to a new camera recently. It has a 'next generation' sensor promising more dynamic range, less noise, and due to improved in-body image stabilisation and various other enhancements it should be possible to get sharper shots under certain conditions. So of course, I was always going to zoom in on the images from my new camera and pixel peep. I needed to know if the marketing speak was to be believed or not.
The first thing I noticed was that the noise was still there. Maybe somewhat less compared to my old camera, but it was there. I obviously knew there was going to be noise - all digital images have noise - but my expectations had been a little too high perhaps. It got me thinking though about ways to reduce the noise. Apart from managing noise by the way you take pictures (low ISO setting, expose to the right, etc.), in my experience Adobe Lightroom will only get you so far. Apply too much noise reduction and the detail in your pictures begins to suffer.
DxO PureRaw Trials
Call me ignorant, but I had never experimented much with software such as DxO PureRAW or Topaz DeNoise AI. I have always relied on Adobe Lightroom for sharpening and noise reduction. I ended up downloading trial versions of both products and in the end decided to purchase DxO PureRAW. Not because I did an in-depth comparison and liked it better than Topaz, but just because DxO happened to have a great offer when I looked on their website. As simple as that. I'm sure that Topaz and similar other products do a great job as well.
The results of DxO PureRaw truly surprised me in the most positive way possible.
Disclaimer: DxO are not sponsoring me, and I'm not affiliated with them.
I used a shot of a local Australian magpie in our front yard taken at f4, 1/2000s, and ISO 1600 for my tests. As you will see in the picture on the left - that has basic sharpening and noise reduction applied in Lightroom - there is definitely noise detectable. It's not too bad when looking at this small picture, but if you click on it and enlarge it you will clearly see the noise. It can't be unseen. The picture on the right has been processed in DxO PureRaw. Again, when looking at this small picture you probably won't notice much difference with the one on the left. It becomes a different story though when you enlarge it again. In my opinion DxO PureRaw has done an amazing job eliminating the noise and sharpening the image. I was absolutely stoked seeing the results.
Left: basic sharpening and noise reduction applied in Lightroom. Right: DxO CamerRaw output.
The jury is still out on whether or not the sharpening applied by DxO PureRaw is a bit too heavy-handed. I suppose it's up to personal taste, but it's good to know that the sharpening can be turned off if desired.
Letting Go of Fear of High-ISO Shots
My conclusion so far is that I need to let go of my fear of high-ISO shots. Previously I was getting nervous shooting at ISO 1600, let alone ISO 3200 or higher. That would pretty much guarantee a nervous breakdown. Keep in mind that I'm using a micro four thirds camera. Given the relatively small sensor size of this system, there is typically more noise in the images than those shot with bigger sensors. Looking at the results of DxO though, I am fairly confident there are not going to be any issues shooting up to ISO 6400 or even higher. I need to do more testing to convince myself, but my initial test results are very encouraging, and I have started increasing the maximum ISO value of the Auto ISO setting on my camera.
Other Lessons Learned
Apart from letting go of my fear of high-ISO shots, here are some other observations and lessons learned so far:
Pictures taken at a low ISO don't need to be processed in DxO. This is a personal choice I'm making. If there's no noticeable noise in an image than what's the point of the additional processing in DxO? I can apply basic noise reduction in Lightroom and do sharpening there as well. I haven't worked out where the cut-off point is going to be, but anything shot at ISO 800 or lower should easily be good enough for processing in Lightroom only. Time will tell where I end up on this.
As mentioned earlier, I'm getting the impression that the sharpening applied by DxO is perhaps a little too much for my liking. I need to conduct more testing and compare sharpening in DxO versus sharpening in Lightroom. The holidays are coming up, so I should have time for that.
It's important to keep in mind where the final image is going to end up. Is it going to be a small thumbnail on social media, or is it going to become a large format print that's going to be framed and hung on the wall? I have never had anyone complain about too much noise in my Facebook photos, so I know what my priorities are going to be.
Finally, processing my images in DxO takes time. A lot of time. It takes approximately 2 minutes per image, so DxO is exposing how old and slow my computer is. I am extremely reluctant to upgrade my computer and am fighting Gear Acquisition Syndrome with all I can throw at it, but when the time comes, I know that I need to consider the demands of DxO as well.
So, this concludes the roundup of my limited testing of noise reduction and image sharpening software. Plenty of caveats apply, and I'm nowhere done with my testing. However, it has been an interesting journey so far and I like the results I'm getting. I think DxO PureRaw will stay on my computer for a long time...
Comments