Darktable is a free, open-source photography application for organizing photographs and developing raw camera files on Windows, macOS and Linux. Its lighttable holds the photo collection, while its darkroom turns camera data into finished images through adjustable processing modules. Edits stay separate from the original files, so a photograph can have several interpretations without requiring several copies of the raw image.
Darktable’s photo library
Importing photographs registers them in Darktable’s library database. Images imported from the same folder form a film roll, giving the collection a starting structure based on their location on disk. Other collections use selection criteria such as filenames, camera metadata or tags instead of requiring another physical folder.
Star ratings and color labels help separate selections within a shoot. Tags can group photographs across folders, while individual titles and descriptions belong in the metadata editor. Collection filters narrow the current selection further, allowing a photographer to work with a particular camera’s images, a tagged subject or selected ratings. Metadata can remain with the library and XMP sidecars or accompany exported photographs.
Raw development
Darktable exposes the processing that converts raw sensor data into a viewable photograph. The darkroom can therefore look flatter than the camera’s embedded JPEG preview: the raw conversion is a starting point for the photographer’s decisions about brightness, contrast and color.
Exposure adjusts overall brightness, with a picker that can match a selected area to a target lightness. Automatic exposure can also analyze raw-image histograms, which is useful when applying consistent corrections across a sequence. Color calibration handles the light source and chromatic adaptation; selecting a neutral patch provides one way to correct a color cast.
Its scene-referred workflow retains an extended range of brightness values during processing before mapping them to the display. The filmic rgb module controls this mapping through scene black and white points and a curve that distributes contrast across shadows, midtones and highlights. Color balance rgb then provides global adjustments or separate treatment of those tonal regions.
Local adjustments
Masks restrict an adjustment to part of a photograph. Drawn masks use brushes, circles, ellipses, paths and gradients, with adjustable opacity and softened edges. Their shapes can be reused by other modules, and they follow the same image area through transformations such as lens correction.
Parametric masks select pixels through values such as brightness, hue or saturation. Combining these with a drawn region lets an adjustment depend on both position and image content. For example, an exposure correction could affect a selected bright region without changing the rest of the frame. The mask determines how strongly the module affects each pixel, so the boundary need not be a hard selection.
Optional AI object masking starts with clicks on a subject and produces editable path shapes. It requires AI support in the installed build, enabled AI features and an active masking model. Processing runs locally, and the resulting shapes remain available for manual adjustment. The Flatpak package does not include the AI features.
Lens correction
Darktable can correct lens distortion, chromatic aberration and vignetting using supported correction metadata embedded in a raw file or profiles from the Lensfun database. Camera and lens identification comes from Exif data, with manual choices available when identification is missing or incorrect. The controls also allow the photographer to choose which corrections to apply and adjust scaling to avoid empty corners.
Correction coverage depends on the available data for the lens and camera combination. Manual vignetting controls provide another way to change corner brightness when a suitable profile is absent. A separate rotate and perspective module can straighten a horizon or adjust converging lines, while cropping sets the final boundaries and aspect ratio.
Noise reduction
The profiled denoise module uses camera and ISO information to select a sensor noise profile. Its wavelet mode separates noise by detail scale and provides control over brightness and color noise. The alternative non-local means method averages similar neighborhoods. Strength and detail controls let the photographer balance smoothing against the retention of fine texture.
Optional machine-learning tools provide another approach when the installed build supports them. RGB denoising works on an already demosaiced image, while raw denoising combines denoising and demosaicing, writes a LinearRaw DNG and imports that result for subsequent editing. These tools use an active model chosen in AI preferences and run when the user requests the operation.
Retouching
The retouch module removes unwanted details using clone, heal, blur and fill tools inside drawn shapes. Cloning copies a source patch to a target; healing also adapts the sampled color and brightness to the surrounding area. Source and target positions remain adjustable after placement.
Wavelet decomposition separates the photograph into layers of fine and coarse detail before retouching. This allows a correction to target a broader blemish without applying the same operation to fine texture. The selected detail layers are recombined to form the output, with the retouch controls showing which scales contain edits.
Editing history
Darktable records processing settings, metadata and tags in its database and normally synchronizes them to XMP sidecar files beside the originals. Multiple editing versions can share the same source image while keeping independent settings and sidecars. Keeping those sidecars with the source files preserves the work needed to reconstruct each version.
The history stack records adjustments and lets the photographer return to an earlier state. It differs from the processing order: the stack shows when controls were changed, while the module pipeline determines when each operation runs on the image. Presets retain settings for individual modules, and styles apply groups of module settings to selected photographs. Presets from other raw editors are not directly interchangeable with Darktable styles.
Photo exports
Export renders the chosen processing into a separate image file. Darktable can export the current photograph or a selection from the lighttable, with JPEG, PNG, TIFF and high-dynamic-range formats among the output options. Format controls cover properties such as compression, quality and bit depth where applicable.
Export settings also control maximum dimensions, scaling, output color profile and included metadata. Filename templates can incorporate variables such as a sequence number. Multiple export presets can produce several sizes in one run, while an export-only style can add processing without rewriting the photograph’s saved history. On Windows, printing requires exporting and using another application’s print function because Darktable’s print module is unavailable there.
Tethered capture
With a supported camera connected by USB, Darktable can capture photographs directly into a film roll. The shutter can be triggered from the program or the camera, and supported cameras can provide live view. Capture controls include timed sequences and exposure brackets, with session names separating new captures into different film rolls.
Tethering depends on camera support through gphoto2. Windows also requires a compatible camera USB driver; its usual camera driver does not provide the connection Darktable needs. Camera compatibility therefore matters separately from support for reading that camera’s raw files.






