Photo Corners headlinesarchivemikepasini.com
A S C R A P B O O K O F S O L U T I O N S F O R T H E P H O T O G R A P H E R
Enhancing the enjoyment of taking pictures with news that matters, features that entertain and images that delight. Published frequently.
DxO Releases PhotoLab 9 with AI Masking, PureRaw 5.3, New Modules
2 September 2025
DxO has released PhotoLab 9 with AI masking as well as PureRaw 5.3 and more camera and lens modules.
We published our PhotoLab news a little prematurely today to avoid a delay in publishing the PureRaw and modules news. But we've added to our review of its AI masking and tested a few other new features in this update.
PHOTOLAB 9
We attended a briefing on PhotoLab 9 and have installed and worked briefly with the beta on Apple silicon.Our full report will appear shortly, after we attend to some immediate medical issues.
AI Masking
PhotoLab 9 introduces an AI-informed object masking that, during the demo, was impressive. By hovering over part of the image, a discreet area is selected for masking.
You can also draw a marque around the area to be masked or select a named object to be masked (like "Sky"). But the automatic object identification was what intrigued us.
To activate it, you select the Masking tool and click on the AI Mask button (first in the row). Then as you mouse over your image various areas will be shown with a red overlay (a bit intrusive, it could use more transparency). If you want to mask that, you click on it.
Our first attempt to use the new tool was on a JPEG image captured with the Adobe Indigo camera app. Not, that is, a Raw file. But as the tool is concerned only with pixel boundaries, we thought that would only help.
We tried it on two images.
One was a seascape (similar to this but with an unclouded sky) with a foreground of land area, a middle ground of ocean and a lot of sky. The AI masking tool did distinguish the three areas, although it lumped some islands on the horizon into the sky.
We were impressed to see if identify individual houses in the foreground, though. As we scrolled along them, we noticed the pixel boundaries were not always discreet so the selection was confused, adding buildings together.
But AI masking failed to completely mask trees, leaving the edges of the foliage out of the mask, as seen below:
Selecting a single tree would sometimes do a better job, but as we added trees to the mask, it would constrict. In some cases, however, even a single tree didn't mask completely.
We had more trouble with the second image, a man standing sideways in a room to show off a plaque. In this case, the problem was the skin tones. They had been oversaturated during capture.
We were able to use the AI masking tool to select his face to desaturate it and the results were satisfactory.
We were a little disconcerted, however, to see the edges of the mask did not quite cover the skin of the head. We're not sure is there is some feathering of the mask or not but the selection did not seem precise.
To show the limitations of this AI masking, take a look at the loop below:
We just moused over the various objects in this image to see how PhotoLab would mask them.
The first few frames show how well it did masking discreet tree stumps. But you'll notice that small stumps with shadows falling on them weren't completely masked. A shadow falling on top was included in the mask but the part of the stump in shade was not, for example.
The ground cover was particularly interesting. The grassy area was well masked despite shadows falling on it. But the bark area was poorly masked. Deep shadows were excluded as were highlights. In fact, the foreground bark area failed to generate any mask at all.
In some cases (like foliage), more constricted masking may not matter. But in others, like the bark foreground, you'll have to resort to actual intelligence.
And in still others, the mask previewed on the screen seemed to be approximate. Take a look at this hair mask and the thoroughness of the color change:
The red hair mask, in this case, seems to miss the edges of the hair. But they too seem to turn blue. Or do they? Perhaps not but they are dark enough not to matter so you don't see two colors of hair in the edit.
We were unable to find the option to select Hair as an object.
To some extent PhotoLab's U Point technology can fade an effect inside a mask where a feather does not extend the mask. So you have options.
Other Improvements
Among the other improvements in PhotoLab 9 are:
- DeepPRIME noise reduction technologies and DxO's exclusive Lens Sharpening Optimization can now be targeted using local adjustments.
- Workflow improvements include image stacking, a Favorites system for folders and projects and direct access to the Project palette from Customize mode.
- A new batch filenaming tool. Filenames can be automatically generated using image metadata, Exif information or custom text, streamlining organization and post-processing. You can also create your own presets.
- Support for Apple's iPhone image formats: HEIC/HEIF and ProRaw.
- DeepPRIME XD3 X-Trans for X-Trans sensors
Price, Availability
DxO PhotoLab 9 is available now from the DxO Web site, for macOS and Windows machines:
- New license: $239.99/€239.99£219.99
- Upgrade from DxO PhotoLab 7 or 8: $119.99/€119.99£109.99
In addition, through Sept. 5, Photo Corners readers can get 20 percent off any DxO software (including Photo Lab 9 by using the discount code LABOURDAY20 at checkout.
PURERAW
DxO PureRaw 5.3, a free update, brings DeepPRIME XD3 X-Trans out of beta for X-Trans sensors, removing all previous limitations and offering full support for fifth-generation X-Trans cameras such as the Fujifilm X-T5, X-H2 and X100VI.
For more information see the news release below.
MODULES
New DxO modules are available today in DxO PhotoLab, DxO PureRaw, DxO ViewPoint and DxO FilmPack.
New camera modules include Canon PowerShot V1, DJI Mavic 4 Pro, Fujifilm GFX 100RF, Nikon Coolpix P1100, Panasonic S1 II/S1 IIE and OM System OM 5 II.
New lens modules include Canon PowerShot V1 Lens, Canon RF 20mm f1.4 L VCM, DJI Mavic 4 Pro lens, Fujifilm GFX 100RF Lens, Panasonic Lumix S 24-60mm f2.8, Samyang AF 14-24mm f2.8 FE (Sony FE), Sigma 14mm f1.4 DG DN A (Sony FE), Sigma 28-45mm f1.8 DG DN A (L-mount), Sigma 50mm f1.2 DG DN A (Sony FE), Sony FE 400-800mm f6.3-8-GB OSS, Sony FE 400-800mm f6.3-8-GB OSS with SEL TC1.4x and Sony FE 400-800mm f6.3-8-GB OSS with SEL TC2x.
For more information see the news release below.
DxO PhotoLab 9 Introduces a New Era of Editing, Adding Exceptional AI Masks, High-Precision Local Adjustments for DeepPRIME Denoising and Sharpening and More
New features also include powerful batch renaming, a host of improvements for a smoother workflow and support for Apple HEIC/HEIF and ProRAW files.
PARIS, France -- DxO, a pioneering force in photo editing software for more than two decades, today announces the launch of DxO PhotoLab 9, the world's most advanced end-to-end Raw photo editing software. Version 9 introduces masking powered by AI that guarantees supremely accurate selections, allowing photographers to create local adjustments with more precision than ever. In addition, new features include: local adjustments for denoising, demosaicing and Lens Sharpness Optimization; a range of workflow upgrades for a smoother editing experience; powerful batch renaming options that use metadata; and support for Apple HEIC/HEIF and ProRAW images. "Version 9 is a major step forward in Raw photo editing," explains Jean-Marc Alexia, VP of Product Strategy. "The AI-powered masking offers unmatched accuracy an when you combine this with all of the other features in PhotoLab 9, it's probably our biggest upgrade to our flagship editing software of the last twenty years."
Introducing DxO AI Masks, the new standard for automatic object selection
The new AI Masks harness the power of artificial intelligence to create smart, pixel-precise selections in an instant. Users can quickly create masks through three different methods: hovering and clicking on different parts of the photo; drawing a box around the area of an image that contains the object to be selected; or choosing from various Subject Types, a list of predefined objects, including sky, people, faces and hair.
The list of predefined objects offers more than just quick selection. For example, if a photographer selects "Hair" and makes some adjustments, these changes can be intelligently copied and pasted between images or saved as a preset that is contextually aware of the selection. Complex retouching becomes faster, smarter and effortlessly consistent.
Combining DxO AI Masks and U Point Technology to achieve unmatched precision
By integrating effortlessly with other Local Adjustments tools, AI Masks deliver an unparalleled level of flexibility and control. Selections can be refined using DxO's renowned U PointTM technology: using a combination of AI Masks, Control Points, Control Lines, Graduated Filters and the Brush tool, photographers can create highly complex masks that would otherwise be too sophisticated for even the most advanced AI. This offers a significant advantage over competing tools, where refining complex selections can be laborious.
Exceptional control with local adjustments for targeted noise reduction and Lens Sharpness Optimization
For photographers who demand the most precise control, DeepPRIME noise reduction technologies and DxO's exclusive Lens Sharpening Optimization can now be targeted using local adjustments.
While PhotoLab's automated processing already delivers outstanding results, this gives users the ability to fine-tune their image with a supreme level of detail, especially given that the new AI Masks -- combined with existing local adjustment tools -- offer incredible precision.
A smarter, faster workflow
DxO PhotoLab 9 brings a host of workflow improvements designed to make photo organization faster and more intuitive. New features include image stacking, a Favorites system for folders and projects and direct access to the Project palette from Customize mode. And as part of a cleaner, more versatile editing experience, you can also easily relocate folders that have been moved.
Powerful batch renaming
DxO PhotoLab 9 introduces a powerful new batch renaming tool, giving users full control over how original and processed files are named. Filenames can be automatically generated using image metadata, Exif information or custom text, streamlining organization and post-processing. Users can also create and save their own presets, making it easy to apply consistent naming conventions across entire projects or workflows.
Support for iPhone images
Version 9 introduces support for Apple's iPhone image formats -- HEIC/HEIF and ProRaw -- making it easier than ever to edit photos taken using iOS devices. This expanded compatibility ensures a seamless workflow for photographers who want to harness the power of DxO's renowned image processing tools with photos captured on their iPhone.
Unrivaled noise reduction and detail extraction for X-Trans sensors: Introducing DeepPRIME XD3 X-Trans
Two years ago, DxO introduced DeepPRIME XD, the 'eXtra Detail' engine designed to use more intense processing for taking on the most demanding files. The technology continues to evolve, ensuring unprecedented results for images captured at extreme ISO levels.
Today, DxO is proud to announce that, following the technology preview released as part of DxO PureRaw 5, DeepPRIME XD3 has emerged from beta and now processes Raw files from all X-Trans cameras.
The image quality is unprecedented, able to recover noise and retrieve details from images captured in very low light like never before.
Price and availability
DxO PhotoLab 9 is available today from the DxO Web site (shop.dxo.com), for macOS and Windows machines:
New license: $239.99/€239.99£219.99
Upgrade from DxO PhotoLab 7 or 8: $119.99/€119.99£109.99
A 30-day trial is available from dxo.com/en/dxo-photolab/download.
DxO PureRaw 5.3 unlocks incredible image quality as DeepPRIME XD3 X-Trans becomes available for all Fujifilm X-Trans sensors
The industry's most advanced Raw processing technology for noise reduction and detail extraction emerges from beta and is now fully compatible with fifth-generation X-Trans cameras.
PARIS, France -- DxO announces the release of DxO PureRaw 5.3, a free update that brings its most advanced denoising and demosaicing technology, DeepPRIME XD3 X-Trans, out of beta for X-Trans sensors. The update removes all previous limitations and now offers full support for fifth-generation X-Trans cameras such as the Fujifilm X-T5, X-H2 and X100VI.
"We've been overwhelmed by the community's response following our call for feedback," says Jean-Marc Alexia, VP of Product Strategy at DxO. "Photographers from across the globe sent us an extraordinary volume of responses -- thousands of comments, images and results and it was extremely positive. We used this input to refine and test the algorithm on a wide range of real-world images."
DeepPRIME XD3 X-Trans technology offers unmatched image quality for the most demanding files
First introduced in DxO PureRaw 5.0, DeepPRIME XD3 X-Trans was launched as a public beta for earlier-generation Fujifilm cameras. Now, following extensive internal testing and thousands of customers submitting their feedback, DxO is delivering the final release with expanded support and even greater performance.
DeepPRIME XD3 X-Trans combines powerful machine-learning denoising with DxO's signature demosaicing and optical corrections. In internal tests, it delivers superb results on high-ISO images, maintaining fine textures while eliminating noise and color artifacts.
Unparalleled image quality for Fujifilm photographers
With DxO PureRaw 5.3, photographers using any X-Trans-based Fujifilm camera can now enjoy the full power of DeepPRIME XD3 X-Trans, transforming their Raw files into images with greater clarity, dynamic range and detail.
The pioneering technology in DxO PureRaw 5.3
- DeepPRIME is DxO's pioneering noise reduction engine for Bayer and X-Trans sensors. Now in its third generation, DeepPRIME 3 delivers spectacularly clean Raw files thanks to the power of machine learning.
- DeepPRIME XD technology uses an extended neural network and applies more intense processing -- ideal for images shot at ultra-high ISO levels. As well as eliminating noise, it renders accurate colors and extracts exceptional detail, despite low light. DeepPRIME XD2s is for Bayer sensors, while the new DeepPRIME XD3 X-Trans is specifically for Raw files from Fujifilm X-Trans sensors.
Price and availability
DxO PureRaw 5.3 is available today from the DxO Web site (shop.dxo.com) for macOS and Windows machines:
- New license $129.99
- Upgrade from DxO PureRaw 3 or 4 $89.99
A 14-day trial is available from dxo.com/en/dxo-pureraw/download.
Version 5.3 is available as a free update to owners of DxO PureRaw 5.
DxO's Latest Modules Push Image Quality to New Heights With Cutting-Edge Cameras and Lenses From Fujifilm, Panasonic, Sony and More
This September, DxO adds over 1,200 new DxO Modules, expanding its industry-leading library to almost 105,000 camera and lens combinations.
PARIS, France -- DxO Labs has announced immediate DxO Camera and Lens Module support for some of the most exciting new releases in photography, including the recently released Panasonic Lumix S1 II, the unique Fujifilm GFX 100RF and Sony's versatile FE 400-800mm f6.3-8-GB OSS super-telephoto zoom.
Whether you're shooting expansive landscapes, wildlife at a distance or ultra-high-resolution studio work, DxO's latest Modules ensure that every image reaches its full potential. Built in DxO's independent laboratory, each Module is a unique profile for a specific camera-lens pairing, correcting distortions, aberrations and softness while delivering the exceptional detail that photographers demand.
Highlights
Highlights include:
- Fuji GFX 100RF: This high-resolution medium-format camera unlocks breathtaking detail -- now refined even further with DxO's bespoke optical corrections.
- Panasonic S1 II and S1 IIE: A workhorse full-frame hybrid, the S1 II gains extra clarity and sharpness when paired with DxO's precision Modules.
- Sony FE 400-800mm f6.3-8-GB OSS: This newly released super-telephoto zoom expands reach for sports and wildlife shooters and with DxO Modules, every millimeter is optimized for edge-to-edge crispness -- even with teleconverters.
Where optical innovation meets scientific precision
These new Modules join DxO's ever-expanding collection, which now covers almost 105,000 camera and lens pairings -- a figure that reinforces DxO's reputation for unrivalled optical expertise. The September update also includes support for exciting new glass from Sigma and Samyang, plus gear from Canon, DJI, Nikon and OM System.
Full list
Full list of newly supported equipment
Cameras:
- Canon PowerShot V1
- DJI Mavic 4 Pro
- Fujifilm GFX 100RF
- Nikon Coolpix P1100
- Panasonic S1 II
- Panasonic S1 IIE
- OM System OM 5 II
Lenses:
- Canon PowerShot V1 Lens
- Canon RF 20mm f1.4 L VCM
- DJI Mavic 4 Pro lens
- Fujifilm GFX 100RF Lens
- Panasonic Lumix S 24-60mm f2.8
- Samyang AF 14-24mm f2.8 FE (Sony FE)
- Sigma 14mm f1.4 DG DN A (Sony FE)
- Sigma 28-45mm f1.8 DG DN A (L-mount)
- Sigma 50mm f1.2 DG DN A (Sony FE)
- Sony FE 400-800mm f6.3-8-GB OSS
- Sony FE 400-800mm f6.3-8-GB OSS with SEL TC1.4x
- Sony FE 400-800mm f6.3-8-GB OSS with SEL TC2x
The complete list of DxO-supported cameras and lenses can be found on the DxO Supported Cameras & Lenses page.
What makes DxO Modules unique?
Unlike generic lens profiles, DxO Modules are meticulously engineered from thousands of measurements taken across every focal length, aperture and focusing distance in DxO's purpose-built lab. This unrivalled optical precision means photographers get sharper, cleaner and more natural images automatically, whether using flagship gear or more affordable setups.
DxO Modules are available today in DxO PhotoLab, DxO PureRaw, DxO ViewPoint and DxO FilmPack.
Almost 105,000 Modules -- and counting
Developed over two decades, DxO's vast library of Modules remains unmatched in the industry, spanning every major camera and lens manufacturer. To discover how these Modules are created and how they transform image quality, visit the dedicated milestone page.