This is a in-depth review of DxO Photolab 7, based on almost a year of use.

I have long been a strong advocate of Capture One Pro, but with the company’s push towards subscriptions and frankly unjustifiable new pricing, it felt like time to consider other software. Lightroom Classic is the obvious alternative, although it has many of the same drawbacks as Capture One.

So the candidate replacement that I have looked at most comprehensively is PhotoLab, which on paper offers a viable alternative as a high quality RAW editor with just enough digital asset management to be useful.

An Overview of PhotoLab 7

PhotoLab 7’s functionality includes:

• high quality RAW conversion
• flexible image editing options
• basic digital asset management

The software’s claimed unique selling point is that it can squeeze out better image quality from your RAW files than any other software. But is that image quality difference real, and is it enough to make up for any shortcomings in the software?

There are two versions of PhotoLab, ‘Essential’ and ‘Elite’. Most users will want to opt for the more expensive ‘Elite’ variant, as its cheaper sibling skips the most unique and useful functions such as Prime noise reduction.

Another wrinkle is that you also need to buy DxO’s ViewPoint and FilmPack software to unlock the full set of geometric and colour correction tools. It is arguable if these are essential, but FilmPack adds some important tools that you might have reasonably expected to be included in the ‘Elite’ edition.

The user-interface has two modes, PhotoLibrary and Customise, plus a full-screen view intended for quick image review and rating. The user-interface is highly customisable and can be used with a multiple-display setup.

The workflow is very simple: locate the image on disk using the PhotoLibrary file browser, switch to Customise and make your edits, then export the results to a final JPEG, TIFF or DNG file.

The bulk of your time will likely be spent in Customise, where you have left and right columns of tools organised with palettes containing editing tools. Palettes can be docked or free-floating, and you can define your own custom palettes too. The vast number of tools can be daunting, but on the right-hand dock there are buttons that filter the displayed tools by broad function (light, colour, geometry, effects, local-edits), or by keyword search.

Once editing is complete, you can choose Export button to render the result to a file:

PhotoLab’s export handling is extremely flexible, allowing resizing, watermarking, batch processing and different versions of the file to be written. However, there is no option for a final output-specific sharpening stage (ie allowing different final sharpening to be applied to output for screen use vs output for print use).

As with any piece of complex software, it takes a while to get used to the user-interface, but for the most part this is relatively easy and the setup is logical. Many tools have keyboard shortcuts, although these can not be redefined. The interface is pleasant and quick to use, and I much prefer it to Lightroom’s dumbed-down approach.

RAW Processing

PhotoLab’s greatest strength is its RAW demosaicing – the process of turning the raw sensor data in to displayable colour pixels.

The demosaicing process appears to be different in that optical corrections and noise reduction are applied as part of the the demosaicing rather than as a subsequent step. This approach is claimed to yield better detail and lower noise than conventional converters.

The downside is that implementing support for new cameras and lenses takes significantly more work than it would if the software was more modular in design. It can take months or even years for DxO to add support for new hardware and many niche cameras are not supported and will likely never be supported due to the work required.

So how much image quality advantage do you gain from PhotoLab?

Noise Reduction

There are four de-noise tool options: High Quality, PRIME, Deep PRIME and Deep PRIME XD.

High Quality is a conventional noise reduction engine based on image filtering, and is broadly equivalent to the tools found in Capture One or Lightroom. This engine can be used with any image file type that PhotoLab supports, including JPEG and TIFF files.

The ‘prime’ tools, on the other hand, can only be used with supported RAW files. They represent three successive generations of AI based noise reduction, with increasing processing requirements but with progressively better results.

Those processing requirements mean that you can not see PRIME noise reduction in the main image viewer, only in the final exported image. When editing you can see a small preview window which shows a section of the image at a fixed 100% magnification, but the fixed magnification and small size frequently makes it difficult to see the effects of any parameter changes.

Lets take a look at an image taken with an Olympus E-M1 mark II at ISO 6400, processed in PhotoLab 7, Lightroom Classic and Capture One Pro and with crops to show the differences in output.

First the full image, processed in PL7:

A crop showing DeepPrime noise reduction with the default settings:

And the same again using DeepPrimeXD:

DeepPrimeXD might be a smidgeon better, but it is really really hard to see any difference that could not be closed by tweaking the sharpening or micro-contrast adjustments.

Now here is Capture One’s output with the default noise reduction setting of 50:

and Lightroom’s standard noise reduction with a luminance strength of 30 and chrominance 25:

Neither of these come close to DxO PRIME’s output.

Finally, here is the image processed using Lightroom Classic’s new ‘Enhance Details’ functionality, again with the default settings:

This is a much better result, but it still has less detail than DxO. Reducing the correction strength will restore some of that, but at the expense of increased noise. And the enhance tool in Lightroom renders a new intermediate DNG file that is added to the catalog, making it is difficult and costly (in disk space) to use this as a default processing setting.

I generally use Deep PRIME on all images, but with the strength set slightly lower than the default. The only penalty to doing this is an increase in the time needed to export the finished image, which is seldom a problem.

Lens Corrections

Lens corrections cover a number of things, including distortion, vignetting, sharpness and various forms of chromatic aberration. PhotoLab, Lightroom and Capture One all boast custom lens data measurements. Many RAW files now also encode correction data from the camera body and lens, although DxO appears not to use this.

DxO’s lens corrections are unusual in that each combination of lens and sensor has its own calibration data. This should allow a higher quality result than when the lens is measured in isolation. For example, digital Leica M cameras suffer a problem with vintage wide-angle lenses where shallow light-ray angles cause colour smearing and/or vignetting problems that can only be adequately corrected using information about both lens and sensor.

All software that I have used has near perfect distortion and vignetting correction today, so I will not comment on this further – it just works.

Where DxO possibly shows a difference from Lightroom and Capture One is in its sharpness and colour aberration corrections. To illustrate this, the following three image crops show from left to right the output from PhotoLab, Lightroom and Capture One respectively:

This is a heavy crop of an image taken with the Fuji GF80, which is known for its colourful fringing. In each case the settings were adjusted as well as possible to reduce artefacts.

Here I think that DxO is clearly the best, followed by Capture One and then by Lightroom. Note that the crops are roughly equivalent to using a 400% magnification on screen – I could not find any examples of images where the differences in output were clearly visible at lower magnifications, which is perhaps more a commentary on just how good modern lenses are than anything else.

Smart Lighting

DxO Smart Lighting is an image adjustment which adjusts exposure in an image, balancing shadows and highlights.

The effect can be based on a uniform image analysis, or spot weighted with automatic face-detection where possible.

As far as I can tell, it does not do anything that could not be achieved with the usual exposure tools found both in DxO and other applications, but its single-slider makes it easy to get natural looking, balanced exposures from high contrast scenes very quickly and easily.

I tend to use it as an alternative to the ‘auto-adjust’ functions in Lightroom and Capture One, although these both usually need more additional editing to match the result from Smart Lighting.

ClearView Plus

DxO ClearView Plus is PhotoLab’s dehaze tool. Like the similar tools in Lightroom and Capture One, it appears to work by intelligently adjusting the tone curve and micro-contrast.

Here is a slide-show with a mildly hazy image processed with 0, 10%, 50% and 100% correction in PL7:

The same again but with Lightroom Classic Dehaze:

And lastly with Capture One’s Dehaze tool:

At the 10% setting, all three applications do a good job, but as the intensity is increased PhotoLab holds up better, without the colour shifts and blocked shadows seen with Lightroom and Capture One.

In reality, you would almost never apply a strong de-haze setting to the entire image. Fortunately all three applications allow the dehaze to be applied as a local correction if necessary.

Colour

Colour has been one of the main areas of improvement in recent PhotoLab versions, and there are now a broad selection of colour editing tools, the most important of which can be used both globally and locally.

Once white balance has been set correctly, normally the first place to visit is the Color/B&W Rendering tool, where you can choose the basic colour profile to apply, how strong the rendering effect should be, and an option to help pull-in extremely saturated colours.

Once a basic rendering has been chosen, you can make further changes using the HSL tools. If you have FilmPack installed there is also a Channel Mixer which is essential for the emulation of colour contrast filters when processing to black-and-white (it is absurd that you need to pay extra for this).

PhotoLab has also recently added a LUT (Look-up Table) tool. LUTs are widely used for colour grading video images, but there is no reason why they can not also be applied to stills. You can find many LUT files online and there are also 3rd party applications that can edit custom LUTs.

Despite all the above, I still regard Capture One as retaining an advantage with its advanced colour editor, which allows an arbitrary number of colour edits and which also has modes optimised for use with skin tones. Capture One also has the huge advantage of a luma-curve tool, which avoids the saturation side-effects of RGB curve adjustments.

Local Editing

PhotoLab provides extensive local editing tools with a useful selection masking tools and adjustments.

As well as traditional graduated filter and brush tools, there are NIK-style Control Point and Control Line tools and a Smart Brush tool. There are currently no AI based masking tools.

A single mask can be made using multiple control lines and control points, as well as their inverses which prevent an area from being masked. And the whole mask can also apply selectively based on colour and luminance values selected with an eye dropper and two sensitivity sliders. Excellent.

The Smart Brush tool allows masks to be defined using edge detection. The brush has an inner and an outer circle: anything brushed by the inner circle is always included in the mask, while anything brushed only by the outer circle is included only up to a detected boundary. It works quickly and effectively, and is almost as fast as using some of the newer AI tools in other software. However, the mask edges are ‘hard’, and I often wish that there was an option to feather the edges for a smoother transition between masked and unmasked areas.

One issue with PhotoLab’s masking tools is that they do not respect geometry changes made after the mask has been applied.

For example, if you make a set of mask based local edits and then rotate the image slightly, the image rotates but the masks do not, leaving them misaligned. Anything which affects the geometry will do this, including the horizon (small image rotations), the perspective correction, or ViewPoint’s Reshape tool. As a minimum PhotoLab should warn the user if a change is being made that might invalidate local edits, but it does not. The only work-around is therefore to be absolutely certain that geometry changes are complete and correct before creating any local mask.

Once you have a mask, you have a good range of tools that can be applied. These include exposure, contrast, white balance, Clear View Plus, HSL, sharpness and blur. Here is an example of a local colour edit made using the Smart Brush and HSL tools:

Be aware that some local tools are more limited than their global equivalents, and the type of tools that can be used are limited. The most obvious omission is the lack of a curves control, something that both Capture One and Lightroom have.

Overall, I think that Capture One still provides the best local editing functionality. Its combination of AI and conventional masking options, plus the unique ability to edit almost any image parameter locally – including Luma Curve – give it a significant edge.

The PhotoLab Database and Sidecar Files

PhotoLab is usually praised for giving you the freedom to manage your photographs on disk as you see fit. There is no catalogue and no ‘import’ process needed before you can see your images.

But this is not quite the case.

In fact, PhotoLab stores all its information about image edits, metadata and projects (albums) in a single, hidden SQL catalogue, located at ~/Library/DxO PhotoLab v7/DOPDatabaseV7.dopdata. Photographs are implicitly imported simply by the act of viewing them in PhotoLab.

Your image edits are stored in that database, in exactly the same way as they would be in Lightroom or Capture One. However, PhotoLab also writes those edits to a sidecar file, which lives in the same directory as the photograph file and has the same name only with ‘.dop’ appended.

For example, if you edit Photos/Italy-2024/DSCF1234.RAF, PhotoLab will create a parallel .dop file which contains the image adjustments and any key-wording that you made:

  Photos/Italy-2024/DSCF1234.RAF
  Photos/Italy-2024/DSCF1234.RAF.dop

It is easy to think that the .dop files are where the edits to your photographs are actually saved, but they are not. In reality, they are backups to the data written to the hidden database and are only used if PhotoLab can not find any other information. You can see this by deleting a .dop file and reopening PhotoLab – so long as the database is not corrupt, your edits will still be present.

Where the .dop files are useful is if you move or rename an image. Do this without a .dop file, and you will appear to loose your edits, because the link between the database information and the file (ie its name and location on disk) has been broken. However, providing that you similarly move or rename the .dop file, then when the image is next opened and PhotoLab cannot find an entry in its database, it will create a new database entry using the whatever the .dop file indicates.

So, PhotoLab gives the illusion that .dop files contain the edits.

One case where this can go wrong is if you attempt to reverse the edits on a single image by restoring its .dop file from a backup. PhotoLab will ignore the changed .dop file and overwrite it with the current data from its database if further edits are made. If you want this to work, after restoring the .dop file you need to immediately force PhotoLab to re-read its contents by using the File Menu’s ‘Sidecars : Import’ action.

PhotoLab’s hidden database is also used to store projects (equivalent to a LrC collection or COP album), keyword lists, and search information. These have no equivalent to the per-image .dop backup file, so if you use these features you should make use of the File Menu’s ‘Project Database : Create a Backup…’ action.

By the way, one cool aspect of the .dop sidecars is that they are plain text files and can be opened in any text editor. This opens up the (unsupported) possibility of automating changes to settings by using a script. For example, this would allow you to fix issues with key-wording that can be very difficult to address with the user-interface.

File and Folder Management

In the PhotoLibrary display, the left-hand column shows a basic file browser. Clicking on a folder will show the images in that folder on the right.

Note that all you see in the folder are the images, but if you look with the Finder you will also see .dop and .xmp files too:

If you move or rename images from within PhotoLab, the sidecar files are implicitly updated as well.

PhotoLab will not create .xmp files by default, but it will read them. To write them you will need to enable an option under the application preferences.

Hovering over an image thumbnail will show an overlay with image file details, metadata, keywords and some processing information:

In the example above there is too much information to show in the popup, so there is a scroll bar. But if I move the mouse to try to scroll the display the pop-up vanishes. Who knows what secrets are contained in those last missing lines?

The Folders column is frustrating if you have a complex disk structure, and it can take a lot of scrolling to find what you need. Fortunately, if you are using macOS, there is a ‘Favourites’ option in the disk browser that will show whatever files you have marked as favourites in the Finder (ie the ones that show up in the left-hand panel in a Finder window). It would be helpful if a future version of PhotoLab had an option to define its own favourite folders, independently of the Finder.

Another omission with the file browser is that there is no option to show images recursively within a folder tree. At the moment I am traveling in Asia, and I have a top-level project folder together with dated sub-folders from each day of shooting. With PhotoLab’s browser, I can only ever see one day’s worth of images – I can not see or edit all the images via the enclosing folder.

Image Tagging and Rating

While PL7 does not support a full DAM (Digital Asset Managment) database, image metadata can be customised and this used to filter the list of images displayed. The options are:

• Tags (‘Picked’, ‘Rejected’ or ‘Untagged’)
• Star Ratings
• Colour Labels

You can use these any way that you like. The tag functionality is obviously useful to quickly rank images for either further processing or deletion. Keyboard shortcuts allow these to be changed in both PhotoLibrary or Customise modes, plus there is also a full-screen mode designed to speed up image review:

The full-screen mode is an excellent way to quickly assess an image, but it lacks the ability to compare two or more images side-by side. In fact, nowhere in PL7 is it possible to show two different images side-by-side, which is frustrating if, for example, you want to compare two images for focus accuracy. This is one area where third-party software such as FastRawViewer is useful.

Projects

Aside from simply using standard disk folders and filenames to organise your images, PhotoLab 7 also provides virtual image collections, termed projects. A project is essentially the same as a Lightroom collection or a Capture One album, where you can group related images independently of their disk location. A single image can be used in many projects, and adding it to a project does not move the file on disk.

You can create projects and ‘project groups’ (virtual folders which contain projects) under the PhotoLibrary tab. To add images to a project you can select the images to add and then drag them to the project icon. Alternatively, you can select one or more images and then right click to bring up a menu with options to add the files to an existing project or to create a new project from the selection.

Of these, bar far the best option is the pop-up menu, as drag-and-drop becomes extremely painful when you have many disk folder or projects and need to scroll the to find the drag destination.

There is no way to add images to a project via the keyboard.

And, unforgivably, there is no undo. If you make a mistake adding or removing images the only way to fix things is to go back and fix things manually, image by image. With drag-and-drop combined with an auto-scrolling toolbar it is all too easy to make a mistake, so use the contextual menu to minimise frustration.

Projects themselves act exactly like normal folders in the disk browser, and can be filtered or sorted by various criteria. But there is no means to manually reorder images, which is baffling given that one of the primary applications for projects is to create image sequences and narratives.

Lastly, projects are fragile. PhotoLab encourages the use of the filing system to organise images, but if an image in a project is renamed or moved outside of PhotoLab, the project will all lose track of the image and just display a blank grey rectangle with a question mark. This defeats one of PhotoLab’s often proclaimed advantages – namely that you organise the image files as you see fit.

So frankly, PhotoLab projects are, as presently implemented, more frustrating than useful.

Is there an alternative?

One possibility on the Mac is to use the Finder to create a folder in which you then put aliases (a form of symlink) to the master images. There are several advantages over PhotoLab:

• Finder Aliases remain functional even if the original file is moved or renamed
• the folder can be accessed from PhotoLab just like any other folder
• you can create an alias either to the RAW file and also its sidecars or just to the RAW file; if you do the latter then the alias files can have their own independent edits from the original
• the Finder Aliases can be arbitrarily renamed to change the sorting order, without changing the original files
• the Finder can undo file rename or move operations (unlike PhotoLab!)

And of course, if aliases make you nervous and you are not stressed for disk space, you can just copy the files entirely.

None of this is remotely as quick or as easy as using an album in Capture One, but it is still light years ahead of PhotoLab’s crude attempt at replicating that functionality.

Search

PhotoLab’s library module has a search function that can theoretically find images from metadata or filename information.

Search works behind the scenes via PhotoLab’s hidden database. Images are implicitly added to the search index automatically when they are opened, and you can also ask PhotoLab to index all images in a folder tree.

Unfortunately, if you move or rename files outside of PhotoLab, the search indexes are not updated. In the above screenshot, you can see each image twice – once with a Thumbnail and once with a question mark. This is because the folder containing the images was renamed and then reindexed.

There seems to be no way to tell PhotoLab that these images no longer exist, nor even any way to filter out the missing images from the display. The only remedy appears to be to delete the entire database and re-index everything from scratch?!

Metadata searches also seem similarly broken. An IPTC search for one lens type returned just two images, when in reality there are more than 1000. With a different lens I get more plausible results, but I simply do not trust that the search is reliable.

Film Pack Additions

If you have a DxO Film Pack licence, PhotoLab gains some features.

Most obviously, you will see numerous new presets, including some very accurate film emulations. There are a huge number of these, and you are bound to find something of value, whether it is Portra colour film emulation, a vintage daguerrotype, or an Aerochrome emulation.

The presets potentially combine many settings to get a final result, but you can also find the colour profiles for the film emulations in the Color/B&W Rendering tool, along with a whimsical time-based film-style browser:

FilmPack also adds a Luminosity Mask option to the Local Adjustments panel:

Luminosity masks are very useful. They set a default mask based on how bright an image area is, and you can add or remove from this subsequently by using the brush or eraser tools.

FilmPack also adds eight more tools:

Filter: adds a simple colour filter effect
Grain: adds film-grain
Channel Mixer: colour based channel mixing – essential for emulating B&W colour contrast filters
Creative Vignetting: adds a vignette effect
Blur: soft-focus and vintage lens blur effects; nothing to do with the pop-group
Frame: adds a (very tacky) photo-frame effect
Texture: adds a texture effect, for example mimicking paper textures
Light Leak: various (also very tacky) attempts to emulate light-leaks in vintage film camera

There are also some less obvious enhancements. For example, the contrast panel normally just has two sliders, Contrast and Microcontrast (clarity), but with FilmPack installed you gain four more settings, adding Fine contrast (clarity) and separated contrast adjustments for the shadow, mid-tones and highlights. Allegedly the split toning panel has more options when FilmPack is installed, although I could not understand exactly what these were.

I can understand why the film emulations might be conditional on FilmPack. There are a lot of emulations and presets that are added, and developing an accurate rendition takes a huge amount of time and effort.

However, much of the functionality that FilmPack adds has nothing to do with film emulation, and it is unjustifiable that basic features such as luminance-masking and the colour channel mixer (essential for black-and-white editing) are not included in PhotoLab by default.

ViewPoint Additions

As with FilmPack, if you have DxO ViewPoint installed, PhotoLab will again gain some features.

The first is ViewPoint’s Volume Deformation tool. This is a form of perspective correction that can be used to correct the geometric distortions at the extreme edges of a wide-angle images. The classic example is a person at the edge of shot taken with a full-frame equivalent 16mm lens. It turns out that I do not shoot like this unless the distortion is an essential part of the image, so I could not find any image where this tool was useful.

Next up, we have a Miniature Effect tool, which produces a fake tilt-shift blur effect:

And finally, a Reshape tool that can apply arbitrary deformations to the image:

The above example is a frivolous, but the reshaping works best when the edits are so subtle that you do not realise that they are there.

I found the tool useful with PhotoLab’s perspective corrections, where sometimes the correction does not feel quite correct, or where a small amount of image stretching can be used to avoid cropping outside of the image. AFAIK, neither LrC nor COP offer a similar tool that works non-destructively.

Of the three tools, I only use Reshape, and that only occasionally.

Working around Unsupported Cameras in PhotoLab 7

I have one camera that I use regularly and which is not supported by PhotoLab: a Leica Q2 Monochrom. This has a sensor without a colour filter array, and so can only shoot images in black and white. It is a niche camera and so unsurprisingly it is not supported by DxO and (unlike its Q2 and Q3 colour siblings) probably never will be.

But it is possible to process unsupported RAW files in PhotoLab 7 if they are first demosaiced by other software. If this can be automated with a script then it might be viable to pre-process the files from unsupported cameras and use PhotoLab to finish the image edits. Two free and open-source based ways to do this are with either Dave Coffin’s dcraw and darktable‘s command line tool, darktable-cli.

Darktable has an advantage in that it is possible to perform some corrections to the images, typically by applying preset settings which you can setup within the main darktable application. In this example, all darktable corrections are disabled, so only the minimum is done to convert the Q2M’s native greyscale .DNG files in to a .TIFF that DxO will accept.

The darktable-cli command line used was:

darktablecli --verbose q2m-sample.dng q2m-sample-dtc.tiff --core --conf plugins/imageio/format/tiff/bpp=16 plugins/imageio/format/tiff/compress=2 plugins/imageio/format/png/compress=2 plugins/Imageio/format/webp/comp_type=1 plugins/imageio/format/tiff/compresslevel=6

Here is the image file, first as rendered by Capture One, then the output from darktable-cli, and finally the edited version after processing in PhotoLab 7:

The results are surprisingly good. Here are two crops from the image centre from both Capture One and the darktable-cli plus PhotoLab 7 combination:

Note that while you can work with JPEG and TIFF files in PhotoLab, you will loose access to PRIME noise reduction and DxO’s lens-specific optical corrections. However, you can still use the non-AI noise reduction and correct distortion manually (as had to be done here).

So, at least for the Q2M, it is to produce good quality images that were edited in PhotoLab. And while the process looks difficult, in reality it is easy to automate the initial darktable-cli processing using a script which is run automatically when images are copied in to a folder. Once in PhotoLab you can then use a custom preset to configure default sharpening, distortion correction and a tone curve.

The primary downside to doing this the need to generate those intermediate .TIFF files, which are about double the size of the original .DNG and take about 10 to 15 seconds of processing for each conversion.

Undo

It seems bizarre to write this in 2024, but undo in PhotoLab 7 is a frequent source of pain.

Undo only works for some actions. If you adjust brightness then saturation, pressing CMD-Z repeatedly does exactly what you might expect. But if you adjust brightness then rotate the image by 90 degrees, pressing CMD-Z undoes the brightness change, not the rotation.

In fact, undo breaks in all kinds of ways. Make some edits, then change the currently selected image, and the undo history is completely lost. Worse still, make some edits with many images selected, change the selection, and there is then no way to undo the multiple edit.

Undo does not work on 90 degree image rotation, but it does on smaller ‘horizon’ adjustments. It does not work on keyword changes, and it does not work on any edits made to projects.

Seriously?!

PhotoLab 7 Bugs

All software has bugs, but PhotoLab seems unusually problematic. Some problems seen while writing this review:

• Visible image banding when applying strong exposure/shadow lifts to low-ISO GFX100s files (above).
• Smart Brush local masks show the total area brushed, not the area that will be affected by the edit.
• Sometimes drag-and-drop just does not work. ie trying to drag an image from a folder to a project, but finding that it can not be dropped. Restart PL7 and it will work.
• Sometimes the image browser just stops working. ie click on an image, and but the image display is blank. double clicking an image no longer opens the image editor. quit and restart PL7 and it starts working again.
• Occasional application crashes were seen while dragging images in to a project.
• At one point, the image search function was clearly giving incorrect results. Re-indexing the image folders did not correct the problem. The ultimate solution was to delete the entire PhotoLab database and re-index, but of course this un-recoverably lost all project data.

While writing this review DxO released a software update that completely broke any attempt at local editing on a Mac. To their credit, within 48 hours they had released a fix, but something must be deeply broken in their quality assurance process to release software with such a huge and unmissable problem.

Give all of this, it is particularly worrying that much of PhotoLab’s state is stored in a single hidden database. As the search issue suggests, issues with that database can be subtle and not easily caught and the essential claim that one of the main strengths of PhotoLab is that there is no catalogue to break is simply incorrect.

PhotoLab as a plugin and PureRaw

PhotoLab includes a plugin that allows you to process a RAW file from Lightroom, then returning the output back to Lightroom for further processing if needed.

In principle this gives you the best of both worlds: DxO’s raw conversion and editing with Lightroom’s cataloguing. The only downside is the need to create an intermediate file (DNG or TIFF), just as Lightroom’s ‘Enhance’ function does.

If the only features that you need are PRIME Noise Reduction and DxO’s optical corrections (ie the RAW conversion), there is an option to return the file back to Lightroom as a linear DNG with only those corrections applied. The resulting image can then be finished using Lightroom’s tools.

Here is an example in Lightroom showing a comparison between Lightroom’s rendering of an image alongside the corresponding PhotoLab linear DNG rendered with PRIME XD Noise Reduction and Optical Corrections only:

DxO also sells a cut-down version of PhotoLab that is intended to be used with precisely this workflow. PureRaw provides PhotoLab’s noise reduction and optical corrections with none of the other baggage. It costs less too.

You can also use either PhotoLab or PureRaw in a similar fashion with Capture One, although because Capture One does not support plugins you will need to manage the export/import process manually.

Whether using PhotoLab or PureRaw, the primary disadvantage of this workflow is disk space.

For example, the GFX100s RAW file used above example is 85MB in size, while the intermediate linear DNG product from DxO is a whopping 266MB. Do this for a lot of images and you very quickly find yourself needing a larger disk.

Conclusion

I really want to like PhotoLab. It can give excellent results, I enjoy using it, and there are some useful features such as the AI noise reduction and Clear View Plus that other software cannot match. But forget about making serious use of the DAM functions, which are rudimentary at best and unreliable at worst.

In summary –

Pros:

• excellent results quickly and easily
• significantly better noise reduction than other software
• more detail extracted from RAW files than by other converters, especially with smaller sensors and high ISOs
• some unique tools such as Clear View Plus, Smart Lighting or ViewPoint’s Reshape
• FilmPack has some great renderings
• powerful and customisable user-interface, yet easy to approach for the first time
• no subscription requirement

Cons:

• the need to buy FilmPack to get decent B&W editing and the luminosity mask tool
• the DAM functionality is mostly worthless
• undo is broken by design
• geometry changes (rotation, perspective, etc) will break any existing local edit mask positioning
• there is no way to compare two different images side-by-side
• using it with officially unsupported cameras is possible but painful
• if your camera or lens is new, it usually takes many months before DxO adds support (if ever)
• seems unusually prone to bugs and poor software QA

One of the motivations for writing this review was to help figure out whether or not it would make sense for me to migrate from Capture One to PhotoLab. But unfortunately, PhotoLab’s lack of support for some cameras, its broken DAM implementation, and the sheer number of bugs that I have tripped over mean that it is impossible for me to use it as my primary image processing tool, no matter how much I like the results.

Of the three camera systems I use, one works perfectly (Olympus), one works partially (Fujifilm GFX, with banding evident in some images), and one does not work without drastic workarounds (Leica Q2M).

This makes it difficult to make any blanket recommendation to other potential users, as anyone’s impression of the software is going to depend very much on what camera they use and what kind of use that may make of PhotoLab’s search, key-wording, or projects.

In contrast PureRaw offers a much more concise and well thought out feature set, while still delivering DxO’s image quality. For existing Lightroom or Capture One users that occasionally need better noise reduction and detail it is a much easier recommendation than PhotoLab.