Adobe Photoshop has lately embraced the power of AI to a somewhat scary degree, but now it has something a little less scary than neural filters, a new feature called Super Res. The handy new tool, now available in Camera Raw XNUMX Thirty-two and soon landing in Lightroom and Lightroom Classic, it has a simple claim: to be able to double the resolution of your photos using machine learning, while improving their quality. This is essential for most photographers, especially those of us who remember the beginnings of the megapixel war. Back then (we're old, okay?), you couldn't even approach a camera launch without being thrown at a camera's megapixel count.
(Image credit: Adobe) 4 megapixels. 8. twelve. eighteen. We remember the arrival of the Canon EOS 5D Mark IV, with its XNUMX Mpx sensor, this is the end of the resolution wars. From then on, just about any camera you could buy would offer enough resolution to do just about anything with the images you wanted. This is doubly true in the age of Instagram. In fact, that's twenty-five times more true in the age of Instagram, because in most cases you only need to upload a XNUMX x XNUMX image, just XNUMX MP worth of photo. Adobe doesn't think the resolution issues have gone away, which is why its new super resolution has come to Adobe Camera Raw. It announces something simple: it's going to take an image (JPG and RAW are accepted) and double the number of pixels along the horizontal and vertical borders. Disbelievers will be seeking their “more is already” comments while we chat, but Super Resolution uses machine learning, which in theory allows it to generate higher-quality results. More does it? We did multiple tests to find out.
How to use Adobe's perfect super resolution mode
Using the tool is easy. Use Adobe Bridge and open an image in Adobe Camera Raw (ACR), then right-click your image and click Thrive, or use CMD-Shift-D if you're in a hurry.
(Image credit: Future) The super resolution tool itself is simple: a single checkbox marked super resolution and the ability to click and hold on the previous view of the image so you can see what the image will do. ACR.
(Image credit: Future) Click Thrive (it's good to see Adobe finally adopting the terminology of the traditional late-'XNUMXs Enemy of the State) and your computer will step back a bit. Super resolution is highly dependent on your system's GPU, so the more you spend on it, the faster your results will be.
How long does super resolution take and is it worth it?
Of course, Photoshop, Bridge, and Lightroom (there are other free photo editors) have offered image scaling for years, so what advantages does super resolution offer? After reading breathless exciting experiences, we excitedly put together some comparisons and…the results were mixed, to say the least. If all you want is to create a large version of a decent, uncropped image, you can save a lot of time, and even image quality, by using Photoshop's standard image resizing. We tested certain images, opening them, resizing them to 2% of their original linear dimensions, and then putting an "enhanced" resize on top of them. While there was always a difference between the two, the difference was rarely in favor of super resolution.
The results
1.) Robin's Test
A good example of the restrictions of Photoshop's Super Res mode is this image of a blackbird. On the left is the original image (which you can see here) resized to twice its original dimensions using Photoshop's basic image resizing, with resampling set to automatic. You can see the full size version here. And the proper version (which you can see in full size here) is what it looks like after being "enhanced". Can you tell the difference? We didn't either, although with the super resolution that a huge DNG file generates every time you use it, our hard drive could certainly do it. In addition to this, zooming to one hundred% showed a drawback. The difference is almost unbearable, but the basic resizing (on the left) shows more sharpness and fine detail than the considerably more CPU and hard drive-hungry super-resolution file on the right. This causes the score to be "1-0" in favor of doing things the old fashioned way.
Base image scaling (left) vs. super resolution (right) (Image credit: Future)
2.) The Rhino Trial
So we tried to collate the 2 techniques in this photograph of a rhinoceros. This time, we're going to jump right into the side-by-side comparison, as there's no way you're going to tell the difference otherwise.
View the full-size version here (Image credit: Future) This time, our subject shows a slight improvement in texture and sharpness when we use super resolution (right) compared to the basic resize tool (left ). You can see the full size version here.
Base image scaling (left) vs. super resolution (right) (Image credit: Future) However, our image is also miscolored, leaving purple aberrations that now need to be corrected. Even if they weren't there, we'd still wonder if Adobe's new AI-powered tool was worth it. Resizing an image using Photoshop's low-cost resizer, in these first 2 examples, produced similar or better results and was absurdly fast (around XNUMX seconds vs. over XNUMX seconds) by comparison. with the super resolution tool.
3.) Redemption of Porto
However, not everything is dark for super resolution; our experiences suggest that there are going to be times when the feature is worth its weight, if not gold, then at least something reasonably robust. For example, here is a photo from the bank side in Porto.
See the full-size version here (Image credit: Future) And now a super violent crop of the exact same image is shown. After cropping, we end up with an image that's just XNUMX x XNUMX pixels, which is pretty tight if you want to print.
View the full-size version here (Image credit: Future) We resized the image using Photoshop's good old resize and Super Res tool, and the results are truly amazing. For a more in-depth look, check out the full-size version.
Base Image Scaling (left) vs. Super Resolution (right) (Image credit: Future) You still need to be careful, but the base resizing version (left) is considerably smoother than the Super version Resolution to the right. There's more detail, sharpness, and while the trained eye notices a slightly high amount of chromatic aberration (purple and green fringing), it's not worth mentioning.
4.) The Tiger Test
Our last super resolution test was this photograph of a tiger, which was cropped from that original shot.
See the full-size version of this crop (Image credit: Future) And, after scaling using both super resolution and good old Photoshop image resizing, here's the side-by-side comparison of the results. Or you can take a closer look by viewing the full size version here.
Upscaling from base image (left) vs. super resolution (right) (Image credit: Future) At this point, we start to question our eyes, our sanity, our purpose. The basic resize, which took a few seconds, is noticeably cleaner than the Super Resolution version on the right.
The verdict
All of this leaves us scrambling to call. There's really no buy resolution as such - if you have a Creative Cloud subscription that includes Photoshop, you already have Super Resolution, so the question is whether it's worth the extra time. In some cases, significantly, yes, you'll be able to enlarge images, preserving or even creating additional sharpening as you go. In other cases, notably not: your images could end up being a bit smoother, and if anything, you're going to be torn between Photoshop doing a basic job, or spending more time on it. Get Adobe Camera Raw to do your 'Enhance work.
The alternative: Gigapixel
All this before estimating the popular Gigapixel AI software from Topaz. It's long and expensive at €XNUMX and ninety-nine, but given our vintage of Port to work with, it generated an image remarkably stronger than Photoshop's sacrifices. This is uniquely noticeable in the full-size image.
Adobe Super Resolution (left) vs Gigapixel AI (right) (Image credit: Future) So we tried using Gigapixel on our blackbird image, just to test sanity, you see, and it all fell apart again, with our image of Gigapixel introducing a previously absent charge of noise, marking them, fine details of our bird. If ever there was an image-editing topic that warranted the shrug emoji's display, it appears to be resizing the image. Take a closer look with the full size version of the comparison now.
Adobe Super Resolution (left) vs Gigapixel AI (right) (Image credit: Future) Yet despite these mixed results, the writing is indeed on the wall for image editing: artificial intelligence and editing based on the cloud will be the way of the future. Did Adobe crack it? Super resolution is an interesting but extremely inconsistent way of juddering images, especially when compared to some great third-party software. But if you don't want to spend money on Gigapixel, if you're frustrated with the way Photoshop is upgrading its traditional tools, let's try the cloud.