I bumped a 100 MP photo to 375 MP using Photoshop’s Super Resolution tool
Mar 12, 2021
Michael Clark
We love it when our readers get in touch with us to share their stories. This article was contributed to DIYP by a member of our community. If you would like to contribute an article, please contact us here.
Share:
Today, March 10, 2021, Adobe dropped its latest software updates via the Creative Cloud and among those updates is a new feature in Adobe Camera Raw (ACR) called “Super Resolution.” You can mark this day down as a major shift in the photo industry. I have seen a bit of reporting out there on this topic from the likes of PetaPixel and F-stoppers, but other than that the ramifications of this new feature in ACR have not been widely promoted from what I can see. The new Super Resolution feature in ACR essentially upsizes the image by a factor of four using machine learning, i.e. Artificial Intelligence (AI). From the PetaPixel article on this new feature they interviewed Eric Chan from Adobe, who was quoted as saying:
“Super Resolution builds on a technology Adobe launched two years ago called Enhance Details, which uses machine learning to interpolate RAW files with a high degree of fidelity, which resulted in images with crisp details and fewer artifacts. The term ‘Super Resolution’ refers to the process of improving the quality of a photo by boosting its apparent resolution,” Chan explains. “Enlarging a photo often produces blurry details, but Super Resolution has an ace up its sleeve: an advanced machine learning model trained on millions of photos. Backed by this vast training set, Super Resolution can intelligently enlarge photos while maintaining clean edges and preserving important details.”
What does this mean practically? Well, I immediately tested this out and was pretty shocked by the results. Though it might be hard to make out in the screenshot below, I took the surfing image shown below, which was captured a decade ago with a Nikon D700 — a 12 MP camera, and ran the Super Resolution tool on it and the end result is a 48.2 MP image that looks to be every bit as sharp (if not sharper) than the original image file. This means that I can now print that old 12 MP image at significantly larger sizes than I ever could before.
What this also means is that anyone with a lower resolution camera, i.e. like the current crop of 24 MP cameras, can now output huge image files for prints or any other usage that requires a higher resolution image file. In the three or four images I have run through this new feature in Photoshop I have found the results to be astoundingly good.
Let’s run through how this works. First off, it works with any image file, whether it is a raw images file, a tiff, or a jpeg. You will have to open the image file in Adobe Camera Raw via Photoshop or Adobe Bridge as shown below. To access the Super Resolution feature, right-click on the image and choose “Enhance” as shown below. A dialog window will come up so you can see how the image will look and you can also toggle back and forth between the original image and the new Enhanced version. The dialog will give you an estimate on how long it will take to create the new Enhanced image, which will show up as a separate image file. Once you are ready simply click the Enhance button in the lower right-hand corner. ACR starts working in the background immediately to build the new image file and it eventually appears right next to the original file you selected wherever that one is stored.
In my testing, as shown below, it took this old 12 MP image from 4256 x 2832 pixels to 8512×5664 pixels. The screenshots below show this enlargement. The top image is the lower resolution (original) version and the bottom image is the one that went through the Super Resolution process. The higher res image looks absolutely amazing. And at 48 MP I could easily blow this up to a 40×60 inch print just as with any image captured using my 45 MP Nikon D850.
Above: The Original image at 4256 x 2832 pixels shown at 100% in Adobe Photoshop.
Above: The new Enhanced image upsized using the Super Resolution feature at 8512 x 5664 pixels shown at 100% in Adobe Photoshop.
Once I upsized the image using the Super Resolution feature, I zoomed into the resulting image and was very impressed. The image seemed just as sharp (if not a little sharper) than the original image file but of course, it is massively larger (in terms of resolution and file size). Kudos to the folks at Adobe for creating a truly revolutionary addition to Photoshop. I have tried some of the Topaz AI software options, like Topaz Gigapixel AI, but I have not seen it work this well.
So what does this mean? For starters, it means that AI technology will have a huge impact on photography. Going forward, the software we use to work up our images (and upres them) might in some instances have a larger effect on the final images than the camera that was used to capture the image. To a certain degree this new tool in Photoshop significantly equalizes the playing field no matter what camera you are working with. All of a sudden my Nikon Z6 and Fujifilm X-Pro 3 (respectively 24 MP and 26 MP cameras) are capable of producing stunning large prints in a way that was previously just not possible.
What about high-resolution cameras you may ask? Where do they end up with all of this? The new Super Resolution tool will allow up to upres any image as long as the resulting “Enhanced” image file is less than 65,000 pixels on the long side and under 500 MP in total. What that means is I can upres the 102 MP images from my FUJIFILM GFX 100 and GFX 100S cameras and produce insane 400 MP image files from a single image. That is getting into the absurd, but that also opens some doors for crazy huge prints. The reality is that this feature is a huge boon to lower resolution (12 to 16 MP) and even medium resolution (24 MP) camera owners. Higher resolution cameras will still yield better image quality but we now have the option of making large prints from relatively low-resolution image files.
This is just the start of the AI revolution. It also shows quite clearly that many of the advancements in image quality are going to come from the software side of the equation as we start to see cameras with incredible specs that might be hard to dramatically improve upon in the coming years. I am super excited about this new option in Photoshop as it will allow me to offer much larger prints than I have been able to create previously–and they will look stunning.
UPDATE 4:53 PM – MARCH 10, 2021
After talking with some photographer friends about this new feature I played around with images from a variety of different cameras to see how it varies. I ran a few images through from my Nikon Z6 and also a few from my FUJIFILM GFX 100. With the GFX 100 image, the Super Resolution feature popped out a 376 MP image file that was damn near identical to the original image file, just four times larger. My jaw hit the floor when I zoomed into 100% and compared it to the original! You can see both the original and the Enhanced images below. There is no way to actually convey the 100% image size here as I have no control over the viewers’ screen resolution but regardless, they both look wicked sharp.
Above: The Original FUJIFILM GFX 100 image at 11205 x 8404 pixels shown at 100% in Adobe Photoshop.
Above: The new Enhanced image upsized using the Super Resolution feature at 22409 x 16807 pixels (376 MP) shown at 100% in Adobe Photoshop.
From what I can tell, the Super Resolution tool seems to do an even better job with higher resolution cameras and in particular with cameras that do not have an anti-aliasing filter in front of the sensor. My Nikon Z6 images when enhanced with this tool still look impressive but not as jaw-dropping as the example above. The Z6 has a very strong anti-aliasing filter, basically a filter that slightly blurs the image to reduce digital artifacts. In addition, it seems like the amount of sharpening or noise reduction applied to the image is also magnified so playing around with how the image is worked up may have a significant effect on the final image quality. I will have to do some more testing.
If you have gotten this far, and are still reading this full-on pixel-peeping madness, then you might have realized that this could be the best upgrade to any and every camera ever. This is certainly one of the most incredible features Adobe has ever released in Photoshop. Try it out for yourself and let me know how it works out for you in the comments below.
About the Author
Michael Clark is an internationally published outdoor photographer specializing in adventure sports, travel, and landscape photography. He has risked his life and limbs on a variety of assignments to bring back stunning images of rock climbers, mountaineers, kayakers, big-wave surfers, and mountain bikers in remote locations around the world. If you would like to see more of his work, take a look at his website, Behance gallery, and Vimeo, follow him on Twitter and Instagram, and like his Facebook page. This article was also published here and shared with permission.
We love it when our readers get in touch with us to share their stories. This article was contributed to DIYP by a member of our community. If you would like to contribute an article, please contact us here.
Related Posts
Does Adobe’s Super Resolution eliminate the need for high resolution cameras?
The Met Museum digitizes the entire collection and releases over 375,000 images for free
Apple just got bumped down to 4th place in the smartphone market
Artist photoshops celebs to fit today’s beauty standards and the results are just sad
Join the Discussion
DIYP Comment Policy
Be nice, be on-topic, no personal information or flames.
17 responses to “I bumped a 100 MP photo to 375 MP using Photoshop’s Super Resolution tool”
The big caveat with this, of course, is that it cannot create detail that was not there. Previously, the algorithms used to up-scale based what it filled in on the surrounding pixels, using various less or more sophisticated assumptions for the interpolation. The change is that Adobe has used millions of photos of different scenes, at very high-res, to learn what kinds of details are there that your low-res image might be leaving out. Based on what it thinks is in your picture (a chair? a tree? a face?), it goes to its vast array of “what does this kind of thing look like at higher res?” and tries to fill in the missing information.
So, just to repeat: It cannot recover information you never recorded. This is just a more intelligent form of guessing at what was left out. For most people, that’s great and it’ll work fine. For some purists, or those who photograph things that the ML/AI was not trained on or not trained well on, the results are sketchy at best.
For example, CSI’s “enhance, enhance, enhance” and suddenly a fingerprint appears over what used to be 1 pixel … that’s not gonna happen with this. Or Star Trek’s “enhance,” again is not something that’s possible to magically make the alien ship that the ML/AI was never trained on magically come into high-res. But, if you photograph a redhead at low res and “enhance,” it knows that they often have lots of freckles, so it could add those in if you would have detected them if you’d shot higher-res.
Not to mention not needing long glass. Shoot 200mm, Super Res, Crop. voila 400mm equivalent at similar res.
enhance works only with raw files – it is stated even in the release notes!
Adobe doesn’t believe you.
“Super Resolution also works on other file formats such as JPEGs, PNGs, and TIFFs. Here’s an example where I captured a time-lapse sequence in raw format, then composited them in Photoshop to produce a TIFF file. I then applied Super Resolution to this composite.”
https://blog.adobe.com/en/publish/2021/03/10/from-the-acr-team-super-resolution.html
No it doesn’t. That’s only a limitation of photoshop atm. Get Adobe Bridge, right click the image and near the top click “Edit in camera raw”. Once there right click and click enhance. I’ve tested with Raw,JPEG,PNG,DNG and TIFF so far.
Having tested Topaz Gigapixel AI, I found it to be able to make convincing upscaled images up to 32,000 pixels on the image’s longest side. Much better than Photoshop’s best algorithm.
I didn’t buy it yet, because if a couple issues.
An application cannot add what was not there, true.
But when GP AI rendered a couple hikers in the far distance of my landscape shot, it seemed like other applications just made much more information disappear.
Yesterday I worked with Enhance and it works as described in the article. Amazing. My new iPhone ProRAW DNG files just jumped from 12MP to 48MP. Makes the iPhone 12 Pro Max an even more capable camera. Exciting. Add Halide and maybe Apple will embed this technology and we have a stunning transformation for a walk-around camera.
I did a comparison with Gigapixel and G had more color noise and maybe a bit less sharp. I did not try to tune G by adjusting parameters. Enhance is fast and much handier to use, just a click in ACR. Maybe limiting Enhance to a simple doubling upsize has led to better results. Not suffering from Swiss Army Knife effect.
It does work with other file types, not just RAW. It retains prior edits.
Hard to believe where this is all going, esp with likely embedding of this in the powerful iPhone. Adobe and Apple are in cahoots with ProRAW.
I’ve been spending the past week doing something very similar to this with my art. Now all of my pieces can scale to pretty much any size without losing quality and it’s such a great feeling!
I’m assuming creating vectors of your Art? If not and its physical art, so long as you have a camera with a decent MP count (Mine is a T7i), you should be able to scale photos infinitely. Tried this with a playing card an a macro lens, details resolves up to the ink spots and because its resolving the finest detail boom. Ran it with enhance 1s to 8000×12000 then a second time for 16000×24000
I’m using AI enhancement on my mixed media pieces. It’s far superior than any vector process I’ve used in the past. Having my pieces go from let’s say 900*1200 to 36000*48000 is something I could have only dreamed of just a few years ago. I would have laughed at anyone who said it was possible. As far as people saying “It can’t add to what’s not there” it absolutely can and in pretty much all of my pieces it’s added textures and even paint strokes in my style all on its own.
Louder please. Its the same people who cried about pixels after film. And Manual being superiors to AF. Adobe giving this technology for free to anyone and everyone is going to change the entire world of art in every aspect.
To the people crying about “This just makes the image worse by making mistakes” or “It can not add what is not there” first of all have you tried it? Side by side?
The data that is captured by RAW images is specifically upscaled and enhanced natively rather than after processing and conversion to a readable format that is viewable on a pc. That’s how it enhances detail. That is the point of this. Another example of a similar method is ISO. The error rate is 0% for Raw Images and 0.0% – 0.5% if you are upscaling JPEG/PNG ect. If it all. So please just shut up. Honestly. Go back to film then and crying about pixels
PSA on how to do this on any image format including JPEG,PNG ect.
This doesn’t just work on RAW Images only. (That’s a limitation of Photoshop Camera Raw atm).
Get Adobe Bridge (Free).
Load your photos, right click and select “Open in Camera Raw”
Now just right click again and click “Enhance”
I’ve tried it on RAW. JPEG, PNG, DNG, TIFF so far. All work.
Also you don’t have to only do it once. If you have an APSC camera (Like me) with 6000×4000.
If you run it once you get 12000×8000. Run it again and you get 24000×16000 (This will be at least 800MB).
But honestly at this point. Thanks to this the gap between Full Frame and APSC in terms of resolution, is even more negligible now.
Bridge isn’t needed for JPEGs and other formats. In the Photoshop open dialog, select your file then change Format to Camera Raw before clicking open.
When will this be fast enough to just consider it a form of image compression?
I am really sorry, but I don’t like your upscaled surfer photo at all. The colors seemed spot on, but I was looking at them at 1080p size. When I looked closer, I could see the nasty work Adobe had done. The problem is simple. Take a photo, blow it up 4x and make it look exactly the same. Why copy a million photos? That’s the wrong approach in my view. Its going to make the photo look different. And guess what? It does look very different from the original. And that’s the problem: guiding these upscalers to a successful conclusion. Sorry Adobe. Now if you’d just fix the Bridge…
“it cannot create detail that was not there”
However others have taken multiple images (handheld, see below) and stacked them with super-resolution mode to recover the lost detail of a single image. Thereby simulating the super-resolution bracketing included in Olympus, Panasonic and Pentax cameras (among others). Of course this only works for stationary subjects.
Why handheld? The super resolution in-camera modes shift the sensor slightly between shots. Imagine vertical lines whereby a single image only captures every other line due to the resolution limits of the sensor. Shot one captures lines 1, 3, 5, etc. Shot two captures lines 2, 4, 6, etc. These can now be stacked together (in camera for the above brands) to capture lines 1, 2, 3, 4, 5, 6, etc.
If the camera (without a super resolution mode) is on a tripod then all the shots will be identical. For the above paragraph, only lines 1, 3, 5 will be captured. But if handheld then body movements will capture additional detail in the other shots. The best example I saw (do some web searching for other examples of this feature), showed 12 stacked images and the results were superior to what is demonstrated here. OF COURSE, that won’t work for the surfer.