Comparing Topaz Gigapixel AI vs Photoshop Super Resolution – Which is best?

Mar 22, 2021

John Aldred

John Aldred is a photographer with over 20 years of experience in the portrait and commercial worlds. He is based in Scotland and has been an early adopter – and occasional beta tester – of almost every digital imaging technology in that time. As well as his creative visual work, John uses 3D printing, electronics and programming to create his own photography and filmmaking tools and consults for a number of brands across the industry.

Comparing Topaz Gigapixel AI vs Photoshop Super Resolution – Which is best?

Mar 22, 2021

John Aldred

John Aldred is a photographer with over 20 years of experience in the portrait and commercial worlds. He is based in Scotland and has been an early adopter – and occasional beta tester – of almost every digital imaging technology in that time. As well as his creative visual work, John uses 3D printing, electronics and programming to create his own photography and filmmaking tools and consults for a number of brands across the industry.

Join the Discussion

Share on:

YouTube video

Topaz Labs’ Gigapixel AI has probably been the industry standard when it comes to upscaling images to massive proportions. It offers some pretty incredible abilities, claiming to let you to resize up to 600% from the original size “while perfectly preserving image quality”. Recently, Adobe announced their new Super Resolution feature, which uses AI to intelligently upsample images.

But how do the two compare with each other? Which is better? Which is faster? Are they both just as good as each other? That’s what Finnish photographer Peter Forsgård wanted to find out. So, in this video, he compares the two side-by-side on a number of images and talks about the pros and cons of each.

It’s an interesting comparison, which also takes the cost of the two into account (perpetual license vs subscription) and while Peter does believe that Gigapixel AI is the overall winner, especially when it comes to things like architecture, cityscapes and landscapes, it is much slower than the Adobe offering. Of course, it offers a lot of options you can tweak to get the best results and if you want the best results, what’s a few minutes of waiting time?

Photoshop is clearly ahead when it comes to speed, though, producing results in a matter of seconds rather than minutes. And if you’re resizing portraits, there really isn’t all that much difference between the images resized by the two different applications from the example shown in Peter’s video. This is where Photoshop’s speed definitely offers huge performance gains over Gigapixel AI – especially for wedding and event photographers who might potentially have quite a lot of images to upscale.

When it comes to cost, the immediate advantage is going to go to Adobe. More people will already have a CC subscription with Photoshop already installed on their system and they basically get the Super Resolution for free as part of the monthly subscription they’d be paying anyway. Gigapixel AI is going to be a separate purchase, but it’s a one-time deal – assuming you’re happy to not upgrade to newer versions in the future.

If you want to try out Gigapixel AI and see how it compares for yourself, they do offer a free trial that you can download. Although if you do want to buy it, you’ll pay just under $100. Adobe also has a free trial of Photoshop CC available, too.

Have you tried them both? Which do you prefer?

Filed Under:

Tagged With:

Find this interesting? Share it with your friends!

John Aldred

John Aldred

John Aldred is a photographer with over 20 years of experience in the portrait and commercial worlds. He is based in Scotland and has been an early adopter – and occasional beta tester – of almost every digital imaging technology in that time. As well as his creative visual work, John uses 3D printing, electronics and programming to create his own photography and filmmaking tools and consults for a number of brands across the industry.

Join the Discussion

DIYP Comment Policy
Be nice, be on-topic, no personal information or flames.

Leave a Reply

Your email address will not be published. Required fields are marked *

9 responses to “Comparing Topaz Gigapixel AI vs Photoshop Super Resolution – Which is best?”

  1. Steven Gotz Avatar
    Steven Gotz

    I did my own comparison as soon as it came out in Photoshop. I used a crop of the eye of an Owl for the comparison.

    The only advantages to Photoshop that I can see, is that it is saved to a DNG, so if you are only really concerned with using Lightroom and not Photoshop, that helps. Especially when it comes to Lightroom Classic. That, and yes, it is faster.

    The Gigapixel Ai, which I limited to the 2X multiple, was slower, and just a touch better with the sharpening.

    But I usually want to set an exact pixel dimension, and Gigapixel Ai will do that, while Photoshop won’t.

    And sometimes I want to go over 2X and Gigapixel Ai will do that too.

    If you don’t already have Gigapixel Ai then stick with Photoshop. If you have it, continue to use it.

  2. Supreme Dalek Avatar
    Supreme Dalek

    I did my own test — https://youtu.be/7fxdRCCXxCY — and got the opposite results: I found that Adobe Super Res worked slightly better than Gigapixel AI for 4x upsampling (which is all that ASR can do… Gigapixel AI can do more sizes and has more fine-tuning adjustments, which may be important for some people.)

    I didn’t test as many samples as Peter did, but one important thing I did do differently was evaluate prints, not just onscreen comparisons. I did this because while pixel-peeping is fun, the most common NEED for upsampling is when sending medium-resolution files to a large-format printer.

  3. Marc Albrecht Avatar
    Marc Albrecht

    in my comparisons, Photoshop did a much better job on good source material by not introducing the terrible artifacts Gigapixel is famous for. However, a stacked enlargement (not using even step sizes) in PS provided even more pleasing results, so I don’t use PS new feature with good materials anyway.
    Gigapixel excels at (some) poor source materials, but at the cost of “creating random data” if you will.

  4. JP dJ Avatar
    JP dJ

    We need more precision, in the discussion. All experiences in the past must be related to versions. In PS and LR/CR some scaling existed. PS had two algorithms. The “Enhance” function now is added to LR/CR, not PS, in a narrow sense. Note that Topaz Gigapixel AI is called that because of its AI. IMO the PS algorithms do some mathematically precise, naive wrt AI, upscaling. Note that TGPAI does its own raw processing. Which AI is better? It depends on how it works. If based on “learning” the training examples are critical to success (think repertoire, styles, use cases). And this implies requirements to comparison testing. A year ago, I felt that TGPAI did much better than PS.
    As to magnification numbers, we need to understand “linear” versus area. 2x linear enlargement is 2*2=4 times area. MP are an area number, TGPAI scales “up to 32K on the longest side” so a 24MP shot if 6000*4000 can get a max linear enlargement if just over 5 times (25 times the area).
    I see others distinguish upscaling from upsampling, too. My definition is that upscaling does not look at context with “domain knowledge” but mathematically only, when upsampling attempts to do better – compare raw processing that causes Moiré, versus removing, versus preventing.
    Let’s say I’ll test LR/CR enhance as soon as I can process my raw images without suffering from what seems random “noise” but is nasty Moiré from randomly textured surfaces.
    I didn’t buy TGPAI at the test time as the installation was buggy, and the app had some issues – these have been addressed, AFAIK.

  5. Toys Samurai Avatar
    Toys Samurai

    My finding is that Photoshop super res does a slightly better job at upscaling text and lineart when the original is too small to be recognized. Photoshop often manages to upscale them to a more natural upscaled version, while Gigapixel would introduce too much artificial pixels that don’t help recognizing the text and the lineart. Gigapixel does present sharper result, however, while Photoshop would produce much softer result that even after sharpening, still wouldn’t match the clarity of Gigapixel’s result and introducing more artifacts.

  6. SooksVI Avatar
    SooksVI

    Although I have an Adobe subscription, I always resort to Gigapixel. More often than not, I also want to enhance detail and not just upscale. Photoshop’s upscale is suitable for issues with great detail and no flaws, but I you need enhance a poor image, Gigapixel takes the cake. I also regularly use AI Sharpen and Video Enhance (they’re even testing deinterlacing).

  7. jamesgordonpatterson Avatar
    jamesgordonpatterson

    Sadly (for me), I’ve found the only topaz plugin that was worth what I paid was Simplify. I guess this stuff must work for somebody, but the demonstrations are not nearly what I can get. And I’ve tried many many times. The answers seem to be in the category of you need a better photo to start out with. Yeah, I need everything perfect, then use a plugin. Gotcha. The “enhance” feature in Adobe does what it says it will do. It may or may not sharpen a photo just a wee bit, but mostly it quadruples an image in size without any wear and tear on the image.

    1. Mencken Avatar
      Mencken

      I’m a Topaz loyalist and attest to the quality of their products although some of their effects will simply not appeal to everyone. This is a given with any plugin/filter and does not diminish the quality or capability of the product.

      You do bring up a very valid point I hear few people raise. In virtually all cases, those who market plugins and filters begin demonstrations with Grade A, superlative photos that could easily stand on their own. If it’s a really bad photo, no filter will rescue it. In large part, the adage “garbage in, garbage out” applies to photography as much as it does to the processing of data.

      In the end, it is the photographer who decides if further enhancement of a very good photo will improve it further in order to reach their vision. The often abused HDR effect would be one example of altering a good photo to achieve an effect the photographer could not arrive at in camera at the time of the shoot.

  8. Mencken Avatar
    Mencken

    Note to Editor: Check your high school grammar book. Look up “comparative case.” Title should read “…–WHICH IS BETTER,” not “…WHICH IS BEST.” (Sorry. Things like this raise my hackles.) Carry on.