Automatic dust correction in digital photographs: How DxO PureRAW ⁶ uses deep learning to find and remove dust spots

DxO PureRAW ⁶ introduces automatic dust detection and removal: a single click identifies dust spots across the entire image and erases them, automating a tedious manual process. The feature combines a state-of-the-art object detection neural network with DxO's proven inpainting engine.

Key benefits for users

• Fully automatic workflow. Dust detection and removal is a single checkbox. Batch-process an entire shoot, and every image comes out clean.
• Adjustable sensitivity. A slider lets the user balance between catching every possible spot (high sensitivity) and avoiding the risk of false positives (low sensitivity).

(That said, we still recommend cleaning your gear from time to time. )

The problem

Interchangeable-lens cameras tend to accumulate dust on their sensor or lenses. These particles cast small, soft shadows in your images — most visible in smooth, uniform areas such as skies and studio backdrops.

Photographers have long dealt with this in post-processing, using repair, heal, and retouch brushes. On heavily affected images or when processing a high volume of images, this quickly becomes tedious.

DxO PureRAW ⁶ automates this process. A detection algorithm scans the image for dust spots, and an inpainting algorithm erases each one automatically.

Why dust detection is challenging

At first glance, sensor dust seems easy to describe: small, dark, roughly circular blobs. But the apparent simplicity is deceptive. Several properties make robust detection surprisingly difficult.

Extreme subtlety. Most dust spots attenuate only a small fraction of the incoming light — often just 5 to 20 percent. They are faint stains, not opaque blots, and their visibility depends heavily on the underlying image content.

Tiny spatial extent. At full resolution, a typical dust spot spans only a handful of pixels — small enough that general-purpose object detectors, which are optimized for people or cars, struggle to register them.

No rich structure. Unlike the objects that mainstream detectors excel at — a face with eyes, nose, and mouth; a car with wheels and windows — a dust spot offers almost nothing for a neural network to latch onto. It is, in essence, a faint dark smudge.

Enormous variability. The appearance of a dust spot depends on the size and shape of the particle, its distance from the sensor surface, the lens aperture, and the color and brightness of the underlying scene. Some spots are sharp-edged circles; others are soft, diffuse halos. Some appear nearly black against a bright sky; others are barely distinguishable from noise. The diversity is far greater than a casual glance would suggest. Dependency on aperture and scene means that the same physical particle can look quite different from one photograph to the next.

The detection model: RF-DETR

The heart of the feature is RF-DETR, a transformer-based object detection architecture. We evaluated several detection architectures, including multiple generations of CNN-based models. RF-DETR was selected for a combination of reasons:

State-of-the-art accuracy. RF-DETR achieves top scores on standard object detection benchmarks, outperforming many well-known alternatives.

Multiple model sizes. Nano, Small, Medium, Large, and XL variants allow us to choose the best trade-off between accuracy and computational cost. We selected the Medium variant (33 million parameters).

Resolution-agnostic architecture. RF-DETR contains no fully connected layers that would fix the input resolution. This flexibility is important for our tiled inference strategy: the image is divided into overlapping 512×512 pixel patches, and the detection model runs independently on each patch. Results are then merged across the full image.

In standard benchmarks, RF-DETR detects dozens of object categories — people, vehicles, animals, furniture. For our use case, we retrained the model to recognize a single class: dust spot. The challenge lies not in classification but in detection — finding tiny, low-contrast features in a vast image.

Training data

Training a reliable dust detector requires exposing the network to a very large number of examples covering every conceivable combination of dust shape, opacity, blur, and background.

We started by collecting thousands of real photographs with genuine dust spots, all carefully labelled by hand. This real-world dataset already covers a great diversity of dust shapes, sizes, opacity, blurriness, and backgrounds, but we wanted to go further.

With its expertise in image and signal processing, our research team developed a dust synthesizer: a compact algorithm that generates a dust spot — indistinguishable from a real one — and composites it onto a random photographic or synthetic background. The synthesizer models the key physical properties of real dust: the irregular blob shape, the per-channel light attenuation in linear space, the blur that softens the edges, and the optional directional shading that some particles exhibit. Every parameter is randomized within carefully calibrated ranges derived from statistical analysis of real dust spots.

This synthetic approach ensures even distribution of dust characteristics and backgrounds throughout the training set, avoiding the biases that inevitably arise in any manually collected dataset. It guarantees, for example, that the network sees enough very faint spots, enough very small spots, and enough unusual backgrounds — combinations that would be underrepresented in a purely real-world collection.

In total, our dust detection network has seen approximately one million dust spots — a mix of real and synthetic — during its training.

DeepPRIME XD3: Fourth-generation AI denoising and demosaicing

DxO PureRAW⁶ introduces DeepPRIME XD3 for Bayer sensors, the latest generation of DxO's deep-learning engine for raw image processing. A single neural network now performs three tasks simultaneously — denoising, demosaicing, and chromatic aberration correction — delivering images with even finer detail than its predecessor.

The technology rests on three pillars: a new multi-task formulation that adds chromatic aberration correction to the network's responsibilities, an optimized convolutional architecture discovered through extensive research, and a significantly improved training pipeline that closes the gap between synthetic training data and real-world RAW images.

Key benefits

• Better image quality. Cleaner color reconstruction, finer detail, and fewer artifacts, especially on high-frequency textures and edges, and particularly on recent sensors without an optical anti-aliasing filter.
• Same processing speed. Despite a substantially more capable network, DeepPRIME XD3 runs as fast as DeepPRIME XD2s on consumer hardware.
• Broad compatibility. DeepPRIME XD3 unites all our recent advances in RAW image processing, and it now supports all sensor types.

A six-year journey

Raw conversion — the process of turning a camera sensor's mosaic of noisy single-color samples into a full-color photograph — has been at the heart of DxO's expertise for over two decades. In 2020, DxO introduced DeepPRIME, the first commercially available neural network to perform denoising and demosaicing jointly in a single pass.

Since then, we have worked relentlessly to push quality further. Deep learning and this holistic approach were also what allowed us to finally support X-Trans sensors, a variant found in part of Fujifilm's camera lineup. These sensors had never been supported by our classical denoisers. In 2022, we introduced the "XD" (eXtreme Detail) family — a second tier of DeepPRIME engines that reach for the highest possible image quality, at the cost of significantly heavier computation that demands a powerful GPU — or a good measure of patience.

2020 — DxO PhotoLab⁴
DeepPRIME. Joint denoising and demosaicing in a single deep neural network (Bayer only).

2022 — DxO PureRAW ²
DeepPRIME extends to X-Trans sensors.

2022 — DxO PhotoLab⁶
DeepPRIME XD ("eXtreme Detail"). More capable architecture and perceptual loss function, encouraging finer detail (Bayer only).

2023 — DxO PureRAW ³
DeepPRIME XD extends to X-Trans sensors.

2024 — DxO PureRAW ⁴
DeepPRIME XD2. Adversarial discriminator loss for a more natural rendering (Bayer only).

2024 — DxO PhotoLab⁸
DeepPRIME XD2s. Improved noise calibration for selected camera bodies.

2025 — DxO PureRAW ⁵
DeepPRIME 3. Three joint tasks: denoising, demosaicing, and chromatic aberration correction (Bayer and X-Trans).

2025 — DxO PhotoLab⁹
DeepPRIME XD3. More capable architecture and two-phase training (X-Trans only).

2026 — DxO PureRAW ⁶
DeepPRIME XD3 extends to Bayer sensors.

Focusing on X-Trans first during the development of DeepPRIME XD3 was a natural choice: the X-Trans version of DeepPRIME XD was older and easier to surpass than DeepPRIME XD2s, which Bayer users already enjoyed. But it led to a somewhat complex situation for the latter. On most images, DeepPRIME XD2s delivered the highest quality, yet on certain low-ISO images affected by chromatic aberrations, DeepPRIME 3 could actually yield better results. The release of DeepPRIME XD3 for Bayer sensors finally brings us back to a simplicity we had not enjoyed since 2023: whatever camera you use, you can choose between two RAW conversion networks — one that strikes a balance between speed and image quality, and one that reaches for the utmost in image quality.

The RAW image restoration challenge

Every digital image captured by a CMOS sensor contains three fundamental defects, all introduced before any software touches the pixels:

Color mosaic. The sensor does not capture full color at each pixel. Instead, a grid of tiny color filters lets each photosite record only one of three colors (red, green, or blue). Reconstructing the two missing color values at every pixel is the task of demosaicing. Two filter patterns are common in digital photography: Bayer, used by approximately 95% of all digital cameras, and X-Trans, found in the remaining 5%.

Sensor noise. Each photosite collects a random number of photons. The resulting shot noise is an inescapable property of light itself, and electronic read noise compounds it further. At high ISO sensitivities, noise can obscure fine detail entirely.

Chromatic aberrations. Most lenses don't focus all wavelengths of light to exactly the same point. The result is small lateral shifts between the red, green, and blue channels, visible as colored fringes along high-contrast edges.

Traditional RAW processing treats these three problems independently: a demosaicing algorithm interpolates the missing colors, a separate denoiser suppresses noise, and a third module corrects chromatic aberrations. Each module works in isolation, unaware of the others' decisions, and each can introduce its own artifacts that complicate the next step. DxO's approach, starting with DeepPRIME in 2020, has always been to solve multiple problems jointly in a single neural network. With DeepPRIME XD3, that principle now extends to all three defects.

Three defects, one network

The case for solving denoising, demosaicing, and chromatic aberration correction jointly is a matter of fundamental interdependence.

Consider what happens when these tasks are separated. Denoising a RAW image requires some understanding of how the mosaic pattern relates to the underlying scene — essentially, an implicit demosaicing on the fly. Conversely, demosaicing a noisy image requires the ability to see structure through the noise — essentially, an implicit denoising — because distinguishing a real edge from a noise fluctuation is critical for correct color interpolation. And demosaicing an image affected by chromatic aberrations is very nearly the same problem as correcting those aberrations: if the red, green, and blue channels are laterally shifted relative to one another, then reconstructing the correct color at each pixel requires imagining what the image would look like if the channels were aligned.

Splitting these three tasks across three separate networks — even networks trained to cope with the artifacts produced by the previous stage — would require more weights and more computation globally, because each network would need to internally replicate part of the intelligence of the others. The result would be longer processing times for equivalent quality, or lower quality for equivalent speed.

A single network, by contrast, can share internal representations across all three tasks. The features it learns to detect edges for demosaicing also help it distinguish signal from noise and identify lateral chromatic shifts.

Synthetic training data

A neural network is only as good as the data it learns from. For DeepPRIME XD3, the quality and realism of the training data are every bit as important as the architecture of the network itself.

The training data problem

When research on DeepPRIME began at DxO in 2018, a fundamental question was: how do we obtain the training examples that a supervised neural network needs — pairs of degraded input images and their corresponding perfect originals?

All options were on the table. Taking pairs of real photographs — a clean, low-ISO shot alongside a noisy, high-ISO shot of the same scene — seemed natural, but proved impractical: the two exposures never align perfectly, moving subjects cause inconsistencies, and the approach would have to be repeated for every camera body and every ISO sensitivity DxO supports. The noise-to-noise approach, which substitutes burst sequences for clean references, suffers from similar scaling limitations. And classical labeling — the backbone of most supervised learning — is simply impossible here: no human can look at a noisy mosaic of single-channel pixel values and propose the correct full-color, noise-free output for billions of pixels.

That left synthetic data generation: starting from pristine, high-quality photographs and simulating the defects that a real camera sensor would introduce. Each training example is thus a pair: a synthetically degraded image, and the original pristine version serving as ground truth. On paper, this is the most scalable solution by far. DxO supports over 600 camera bodies across roughly 20 ISO settings each, creating over 12,000 possible configurations. And this figure accounts only for noise: chromatic aberrations depend on the lens, the aperture, the zoom setting, and the focusing distance. If we wanted to capture real image pairs for every camera–ISO–lens combination, the number of configurations would explode into the millions. Synthetic generation can cover all of them from the same pool of ground-truth images.

The distribution gap

The challenge with synthetic data is a phenomenon known as the distribution gap: the statistical difference between the simulated training images and the real RAW files the network will encounter in production.

A naïve simulation — shifting color channels slightly to mimic chromatic aberrations, removing two color values out of three to simulate the Bayer mosaic, then adding white Gaussian noise — is enough to generate the above illustrations for this white paper. It is not enough to train a neural network. A network trained on such idealized data would perform well on synthetic images drawn from the same simulation, including images it has never seen during its training, but it would fail on real RAW files from real cameras.

Real RAW images differ from a naïve simulation in countless ways:

Noise is not purely white Gaussian. Photon shot noise is indeed white and signal-dependent, guaranteed by the physics of light. But real sensor data is a mixture of photonic and electronic noise. Electronic noise — read noise, dark current, banding — can exhibit spatial correlations, non-Gaussian tails, and fixed patterns that vary from one sensor design to the next.

Chromatic aberrations vary across the field. Lateral color shifts are not uniform — they change in magnitude and direction from the center of the image to the corners, following the optical properties of each specific lens.

"RAW" files are not truly RAW. Before the data is written to the memory card, the camera applies a series of in-camera processing steps that alter the signal: black level correction, fixed-pattern noise subtraction, static defective pixel correction, focus pixel interpolation. Some manufacturers go further and apply lossy compression or even noise reduction to what they label as RAW data.

Sensor behavior changes with usage. Noise characteristics can shift depending on sensor temperature, shutter mode (mechanical vs. electronic), and other operating conditions. All of this varies across manufacturers and across camera generations. Manufacturers do not document their internal processing. We must infer what they do based on careful observation.

Closing the gap

Since 2018, DxO has leveraged everything at its disposal to minimize the distribution gap: two decades of expertise in image signal processing and, crucially, a proprietary calibration database that has no equivalent in the industry. For every supported camera body, at every ISO setting, DxO's laboratory has captured and analyzed calibration images — both photographic content and dark frames — to characterize not just the standard deviation of the noise, but its full statistical profile: its distribution, any spatial correlations introduced by in-camera processing, and how these properties change across the sensor and across operating conditions. This database, originally built to feed DxO's classical denoising algorithms, turned out to be an invaluable foundation for training neural networks.

Sometimes, however, some cameras reveal gaps that the existing simulation does not cover. A recent example illustrates the challenge: Fujifilm's X-Trans sensors of the 4th and 5th generations, where something changed relative to the first three generations. Despite extensive efforts, our DeepPRIME XD2 training pipeline never managed to produce satisfactory results for these sensors, which is why DeepPRIME XD2 and XD2s were released as Bayer-only.

For DeepPRIME XD3, properly supporting these sensors was a top priority. Over months of investigation, the team dissected how the newer X-Trans sensors differed from their predecessors and systematically adjusted the training data synthesis until the distribution gap became small enough for the network to generalize well to real images from these cameras.

Finding the best architecture

Adding a third task and demanding better demosaicing quality required a more capable network. The team began with a broad exploration. Transformer architectures, which dominate many fields of deep learning today, were tested alongside multiple convolutional neural network (CNN) designs. For this particular task — recovering fine, local image detail from noisy and incomplete data — CNNs proved more effective. Their inherent local bias, which focuses on small spatial neighborhoods, naturally encourages the smoothing of noise without hallucinating structure that is not there. Transformers, which model long-range dependencies, tended to let noise through rather than suppress it. For a denoiser, the CNN's bias toward local regularity is a feature, not a limitation.

An early prototype of DeepPRIME XD3 achieved the desired quality, but ran three times slower than DeepPRIME XD2s — too slow for a production tool used on thousands of images. The challenge, then, was to find an architecture that could be just as intelligent while fitting within the same computational budget. The team explored different convolutional block designs, separable convolutions in place of the full 3D convolutions used in earlier generations, different activation functions, and varying amounts of computation allocated to each scale of the U-Net.

Each candidate architecture was trained for approximately three weeks on an Nvidia H100 GPU. Around 50 configurations were evaluated in total, amounting to roughly three years of cumulative H100 GPU time dedicated solely to architecture exploration.

This entire process was carried out twice: first for X-Trans, then for Bayer. This is the principal reason why the Bayer version arrives only now in DxO PureRAW ⁶, while the X-Trans version was already released six months earlier in DxO PhotoLab⁹.

The outcome is a network with significantly more parameters than DeepPRIME XD2s, arranged in a way that keeps inference time essentially the same on consumer hardware. More weights, more intelligence, but no significant penalty in processing speed.

Renoising, rethought

Almost twenty years ago, DxO's researchers made an observation that still holds today: it is very difficult to make a denoiser remove only part of the noise. Denoisers — from the earliest wavelet and non-local means filters to modern neural networks — generally perform best when asked to remove all noise. Attempting partial removal tends to produce artifacts. The better the denoiser, the more detail it preserves in the process, but even the best denoisers inevitably erase some fine structure along with the noise.

To avoid the "plastic" look that results from fully denoised images, our researchers devised a simple but effective technique: let the denoiser do its job completely, then add a small fraction of the removed noise back to the image. Reintroducing part of the original noise, rather than synthetic white noise, has a crucial advantage — it also reintroduces part of the fine detail that was lost in the process. The first product to feature this technique was DxO OpticsPro 5, released in 2008. Even though DeepPRIME XD3 is vastly more capable than the denoising and demosaicing algorithms of that era, the principle remains as valid as ever.

For DxO PureRAW ⁶, we reworked how this noise reintroduction interacts with our lens corrections, specifically with vignetting and distortion correction. Both corrections are now applied before adding the residual noise back to the image, which allows us to treat the main signal and the noise component differently.

Vignetting. The noise level in RAW images depends on the signal level in a nonlinear way. With a lens that exhibits strong vignetting, the signal-to-noise ratio decreases significantly in the corners. When we amplify the corners to produce a uniformly bright image, we also amplify the noise, leaving it visibly stronger than in the center. The solution is to use the noise model — the known relationship between signal level and noise level — to derive a correction factor that produces homogeneous noise across the frame, and to apply this factor to the noise before adding it back.

Distortion. Distortion correction requires geometric interpolation of the pixel grid. When applied to white noise, interpolation introduces two unwanted effects: it creates spurious structure in the noise, and it causes periodically varying noise levels. At positions where the interpolated coordinate coincides with a real pixel, the noise is preserved as-is, while at positions that fall between pixels, the noise is smoothed and its level drops. In DxO PureRAW ⁶, we address this by applying a specialized interpolation algorithm to the noise component separately, ensuring that its level remains uniform after distortion correction.

Both effects are most visible at high ISO settings, where the residual noise — even though it is only a fraction of the original — is clearly perceptible.

This improved renoising pipeline applies to both DeepPRIME 3 and DeepPRIME XD3. It is a good example of how much we care about the details: our ambition is not "only" to build the world's best denoiser, but the world's best RAW conversion engine.

The results

The practical effect of all these advances depends on image content and shooting parameters. Compared to DeepPRIME XD, which DeepPRIME XD3 replaces for X-Trans sensors, the new engine generally yields cleaner, more natural results. Compared to DeepPRIME 3, it almost always produces images that are both cleaner and more detailed, at all ISO sensitivities. The difference with DeepPRIME XD2s is more subtle: DeepPRIME XD3 shows its advantage most clearly on images with fine textures, sharp lenses, sensors without an optical anti-aliasing filter, and lenses exhibiting chromatic aberration. Improvements in demosaicing and chromatic aberration correction are best visible at low ISO, while improved detail preservation is most apparent at intermediate to high ISO settings.

How DxO’s pioneering approach makes DNG files four times smaller without impacting quality

DxO PureRAW ⁶ introduces a new high-fidelity compression option for the DNG format, reducing file sizes by approximately 4x compared to the current lossless compression, while preserving full perceptual image quality.

DxO’s new High-Fidelity Compression technology combines two complementary techniques: Dynamic Range Compression and the JPEG XL image codec.

Key benefits

• 4x smaller files — A 50 MP camera's Linear DNG drops from ~200 MB to ~50 MB, making Linear DNG practical for everyday use and high-volume workflows. Smaller files mean faster imports, faster cloud syncs, and less disk usage.
• High fidelity — The compression is perceptually transparent, even under aggressive editing.
• Compatibility — The output remains a standard DNG file. Any DNG-compatible application (Adobe Lightroom, Capture One, etc.) can open and edit these files normally.

Why compress more?

Linear DNG is DxO's recommended output format for DxO PureRAW because it preserves maximum editing latitude while being universally compatible with third-party RAW processors. However, even with the lossless compression built into the DNG specification, a typical Linear DNG weighs in at approximately 4 MB per megapixel. For a 50 MP camera, that is 200 MB per image.

Clearly, there is a strong motivation to compress these files more aggressively.
But how far can we go without compromising quality?

From lossless to perceptually lossless

Lossless compression is the most reassuring approach for both developers and users alike, since it guarantees that the decompressed file is mathematically identical to the original, bit for bit. However, this class of algorithms is inherently limited in efficiency, especially when the signal being compressed contains information that, from a perceptual standpoint, is useless.

For DxO PureRAW 6, our image scientists have developed a compression scheme that targets this useless information, removing it before compression and thereby achieving much better compression ratios. The result is what is known as perceptually lossless compression: the mathematical loss it introduces is not perceivable by a human observer under typical viewing and editing conditions.

We identified two types of perceptually irrelevant information in Linear DNG files:

1. Excess pixel precision. Digital camera RAW files are typically encoded at 12 or 14 bits per pixel; the output of our DeepPRIME pipeline uses 16 bits. However, images always retain some residual noise, intentionally left in place to prevent the unnatural “plastic” appearance caused by complete denoising. As we explain below, the more noise a signal contains, the less its full numerical precision is relevant. Removing the unused precision is the role of Dynamic Range Compression (DRC).

2. Exact texture and grain shape. In practice, slight differences in the exact shape of noise grain or fine texture are imperceptible. Simplifying these micro-details is a classic principle in image and video compression, and is the domain of the JPEG XL codec.

Both techniques require standard DNG mechanisms so that any compatible software can open the resulting files transparently. DRC is encoded via the DNG Linearization Table tag, and JPEG XL is a compression mode introduced in DNG specification version 1.7. Both are supported by common RAW processing applications.

Dynamic Range Compression

Dynamic Range Compression (DRC) is a well-known technique in audio signal processing. A compressor reduces the dynamic range of a signal by applying a non-linear transfer function: in audio terms, loud parts are attenuated, and quiet parts are boosted so that the signal fits more efficiently within a given bit budget. The same principle turns out to be remarkably well-suited to RAW digital images.

Why DRC works for RAW images

Digital images are affected by photonic (shot) noise, a fundamental property of light itself. The standard deviation of this noise grows with the square root of the signal intensity.
This has a profound consequence for compression of linear images:

• In dark regions, noise is very low, and the signal is finely structured. Every bit of precision can carry genuinely useful information — 14 or even 16 bits may be needed.
• In bright regions, noise is comparatively large. The useful signal precision is far lower than what 14 or 16 bits represent. Those extra bits encode noise more precisely than anyone would ever need or could ever see.

It is precisely these perceptually useless high-precision samples in the highlights that make lossless compression less efficient: the compressor must faithfully encode bits that carry no meaningful information.

• DRC addresses this by applying a companding function — concretely, a curve close to the square root — to the linear pixel values before compression. This is conceptually related to a variance-stabilizing transform: after the square root, the noise standard deviation becomes approximately constant across the entire tonal range. Precision is thereby allocated where it matters — many levels in the shadows, fewer in the highlights — without discarding any information that was perceptually meaningful to begin with.

At decompression time, the inverse function (stored in the DNG Linearization Table) restores the original linear encoding, exactly as the DNG specification intends. The process is fully transparent to any downstream application.

The number of quantization levels was chosen conservatively and validated against worst-case editing scenarios such as large exposure pushes combined with extreme shadow recovery to ensure that quantization artifacts remain invisible in all practical uses.

JPEG XL compression

After DRC, the conditioned image is compressed using JPEG XL, the next-generation image codec standardized by the JPEG committee.

What makes JPEG XL better than legacy JPEG?

Legacy JPEG dates from 1992 and relies on a fixed 8x8 block transform with relatively simple entropy coding. While groundbreaking in its time, this approach leaves significant compression performance on the table by today's standards. JPEG XL incorporates over two decades of advances in image compression research:

Variable-size transforms — As small as 2x2 and up to 256x256, these allow the encoder to use large, efficient blocks in smooth regions and small, precise ones near edges, adapting to local image content rather than forcing a one-size-fits-all grid.

Perceptually optimized color space — JPEG XL's internal color representation is modeled on the human visual system, enabling smarter allocation of bits to the aspects of the image that matter most to perception.

Advanced entropy coding — Modern and significantly more efficient coding techniques extract more redundancy from the data than legacy approaches could.

Sophisticated prediction and context modeling — The encoder builds a statistical model of the image as it goes, capturing fine-grained local structure and reducing the amount of truly unpredictable information that must be stored.

Native high bit-depth support — unlike legacy JPEG, JPEG XL is designed from the ground up for high bit-depth content, making it an ideal compression layer for RAW imaging pipelines.

We apply JPEG XL with a near-lossless quality setting, meaning the mathematical loss introduced by the codec is negligible — far below the noise floor of any real-world image. The combination with prior DRC is what makes the compression so effective: by removing perceptually irrelevant precision before handing the data to JPEG XL, we give the codec a signal that is inherently easier to compress, without asking it to make any quality-damaging decisions.

DxO PhotoLab 9.6: New levels of image quality, control, and efficiency

DxO PhotoLab 9.6 is now available, delivering major advances in image quality and workflow efficiency. This latest update introduces DeepPRIME XD3 for Bayer sensors, expands creative control with AI Mask diffusion, and debuts High-Fidelity Compression, dramatically reducing DNG file sizes without compromising RAW quality.

With these innovations, DxO PhotoLab 9.6 removes key limitations faced by demanding photographers, offering more freedom at every stage of the editing process, from noise reduction and local adjustments to export and archiving.

A fully featured 30-day free trial is available now.

DeepPRIME XD3 delivers outstanding image quality for all sensors

With version 9.6, DeepPRIME XD3 is now available for both Bayer and X-Trans sensors, bringing DxO’s most advanced RAW processing technology to an even wider range of cameras.

Built on the foundations of DeepPRIME, DeepPRIME XD3 is designed for the most demanding shooting conditions. It produces cleaner, sharper images while preserving natural textures and accurate color reproduction, even in extremely high-ISO files. From night photography to highly detailed landscapes, it sets a new benchmark for noise reduction and detail extraction.

Combined with DxO’s exclusive DxO Modules, which provide the most precise camera and lens corrections available, DeepPRIME XD3 delivers a complete image quality pipeline for photographers who refuse to compromise.

AI Masks gain diffusion for smoother, more natural adjustments

DxO PhotoLab’s acclaimed AI Masks become even more versatile in version 9.6 with the addition of diffusion controls. This new option allows photographers to soften mask edges and create smoother transitions, resulting in more natural-looking local adjustments.

Whether refining portraits, enhancing landscapes, or working with complex tonal transitions, diffusion offers greater creative flexibility while preserving the intuitive workflow that has made AI Masks a cornerstone of DxO PhotoLab.

High-Fidelity Compression: smaller DNGs without compromise

DxO PhotoLab 9.6 also introduces High-Fidelity Compression for DNG export, producing files that can be up to four times smaller than standard uncompressed DNGs.

This breakthrough allows photographers to significantly reduce storage requirements without sacrificing image quality, dynamic range, or editing flexibility. Large-scale processing, long-term archiving, and day-to-day workflows all become faster and more efficient — while maintaining a fully professional RAW workflow.

What is DxO PhotoLab?

DxO PhotoLab is our flagship RAW photo editing software, built on more than 20 years of cutting-edge research and innovation. Awarded year after year for its image quality, DxO PhotoLab gives photographers powerful tools driven by science, not hype — from DeepPRIME noise reduction and our exclusive lens correction modules to local adjustment technologies like U Point.

Version 9 builds on this foundation by introducing the powerful and versatile AI Masks, targeted noise reduction and Lens Sharpness Optimization, and smart workflow refinements. Whether you shoot landscapes, portraits, street, or still life, DxO PhotoLab⁹ helps you craft images with spectacular fidelity.

To learn more about DxO PhotoLab⁹, visit the Overview page, and to see a comprehensive list of features, click here.

DxO PureRAW ⁶ launches with groundbreaking DeepPRIME XD3 for Bayer sensors, and dramatic workflow gains

Next-level noise reduction, breakthrough compression, and blazing-fast processing.

DxO today introduces DxO PureRAW 6, a major new release that pushes RAW image quality and workflow speed to unprecedented levels. Already established as the essential first step of any professional RAW workflow, version 6 now delivers even sharper, cleaner, more detailed images — and introduces a suite of intelligent upgrades designed to give photographers total freedom and efficiency.

With this release, DeepPRIME XD3 — previously exclusive to X-Trans sensors — now brings its elite noise-reduction and detail-extraction performance to both Bayer and X-Trans cameras. Combined with breakthrough DNG compression and powerful workflow improvements, DxO PureRAW 6 sets a new benchmark for what photographers can expect at the start of their editing process.

Next-level quality: DeepPRIME XD3 expands to Bayer sensors

DeepPRIME XD3 represents the most advanced evolution of DxO’s acclaimed DeepPRIME technology. Engineered for the most challenging files, XD3 uses a larger neural network to clean noise, restore detail, and preserve natural textures with exceptional precision — even in extremely low-light, high-ISO conditions.

By extending this technology to Bayer-sensor cameras, DxO PureRAW 6 now delivers its highest-ever level of denoising and demosaicing quality — performed simultaneously for unprecedented results — to more photographers than ever before. From night photography to finely detailed compositions, XD3 delivers extraordinary clarity and depth that redefines what’s possible straight out of the RAW file.

Dramatically smaller files, no loss of quality? Say hello to compressed DNGs

DxO PureRAW 6 introduces a breakthrough in efficiency with new High-Fidelity Compression, delivering RAW-quality output in DNG files up to four times smaller than standard uncompressed versions.

Photographers can now enjoy the full dynamic range and flexibility of a RAW workflow while saving huge amounts of storage space — a major advantage for large projects and long-term archiving. Heavy file management is made easier with a lighter, faster, more streamlined experience.

Powerful workflow upgrades

Batch parallelization for faster throughput

DxO PureRAW 6 significantly accelerates high-volume processing with batch parallelization, which intelligently prepares the next image before the previous one has finished. This results in dramatically faster throughput and smoother performance when working with large sets of files — a major advantage for demanding workflows.