Nine years ago, I spent a night shooting wolverines in Erä-Eeros’s photography hide in Eastern Finland. I was using a Pentax K-5 and a Sigma 100-300mm F4. The Pentax was a APS-C camera, so that’s equivalent to 150-450mm in 35mm terms.

I just purchased a set of licenses for TopazLabs products (DeNoise AI, Sharpen AI and Gigapixel AI), and it came to me that the wolverine photos from nearly a decade (ok, over 9 years) back would be perfect test material for trying the TopazLabs tools.

Please note: Many will not read this article to the end, so I’ll mention an important fact up first: If you are convinced of the capability of the TopazLab apps and want to buy them, use the code FRIEND15 to save 15%. The code is not an affiliate code, just a promo code to save you money. See bottom of article for an affiliate link if you want to support this site.

Now let’s get to business.

Case 1: DeNoise and Sharpen and GigaPixel

Let’s take as a first case study this photo that has been shot at ISO1600, 280mm, f/4.0 and 1/500s shutter speed. I cropped from the full 16 megapixels down to about 7 megapixels. It was exported as a JPEG at 100% quality from Lightroom. Working with the RAW might have provided some benefit, maybe I’ll try it with some other case. If you click on the photo to view it full size, you can see the noise, you can see a bit of blur.

Click on the photo to view in full size

So here’s what we’ll do to improve it:

  1. Denoise it with DenoiseAI
  2. Sharpen it with SharpenAI
  3. Increase size with GigaPixel AI

Here’s the photo after DeNoise AI, in which the automatic model selected the Low Light algorithm for denoising. I must say the difference is slight, with most benefit in the smoother background. The Pentax K-5 was quite good at ISO1600, although the loss of detail is probably the worse problem with the ISO than the noise pattern.

After the SharpenAI, we get the result below. Now the effects are more apparent. I’m not quite sure, if the results are good, excellent or even too aggressive.

As a final step, let’s upscale it with the GigaPixelAI and compare it to the Adobe Super resolution from Lightroom. You can download the upscaled to 28 megapixel image here or see the crop below.

Case 1: Conclusions

For this photo, the biggest benefit was clearly the SharpenAI. GigaPixelAI seems to do a good job for scaling the photos, but more comparison with the Lightroom Enhance / Adobe Superresolution would be warranted to see the different in quality it produces. DeNoise AI was not very critical here, since the noise was already well under control because of the correctly exposed photo.

Case 2: DeNoise AI and Sharpen AI for a darker image

Let’s see the same set of post processing, but this time on a ISO12800 photo at 110mm, f/4.0, 1/125s. The photo has not been selected because of any merit as a photograph, but rather as a good technical example of a high ISO noise.

Here is the original 16 megapixel photo, cropped to 8.5 megapixels.

Here is the post processed photo after DeNoise AI and Sharpen AI

Let’s just add a 100% crop comparison to make it obvious

Case 2: Conclusions

I think this case shows a better use case for the DeNoise AI. Sharpen AI also works quite well. The image goes from borderline unusable, to well usable on social media and maybe even printable. Of course, no amount of apps can save the photographer from a badly composed photo. But sometimes, in a photography blind, you just have to take what’s given to you.

Update on 14.9.2021: Save 15% on the TopazLabs apps

After being blown away by the TopazLabs software, I checked if they have an affiliate program, and they do. Partnerning with TopazLabs is just the kind of affiliation I look for this site, where I love the product first, and after that partner with the company.

Therefore I am happy to let you know, you can save 15% AND support this site, if you buy the apps through the links on this site and use the code FRIEND15. Please note: the code is not an affiliate code, just a promo code to save you money. So you can use it without using the affiliate links if you want to.


Please enter your comment!
Please enter your name here