processing

Control how images are processed.

🔗image processing

always use LittleCMS 2 to apply output color profile: If this option is activated, darktable will use the LittleCMS 2 system library to apply the output color profile instead of its own internal routines. This is significantly slower than the default but might give more accurate results in some cases.; If the given ICC is LUT-based or contains both a LUT and a matrix, darktable will use LittleCMS 2 to render the colors regardless of this parameter’s value (default off).
pixel interpolator (warp): The pixel interpolator used for rotation, lens correction, liquify, crop and final scaling.; Whenever we scale or distort an image we have to choose a pixel interpolation algorithm (see wikipedia for details). For warping modules, darktable offers bilinear, bicubic or lanczos2. In general, bicubic is a safe option for most cases and is the default value.
pixel interpolator (scaling): The pixel interpolator used for scaling. The same options are provided as for the warp modules, but with the addition of lanczos3.; lanczos3 can cause pixel overshoots leading to artefacts but sometimes gives a more crisp visual appearance. This option is therefore only provided for transforming (scaling) algorithms and is the default value.
LUT 3D root folder: Define the root folder (and sub-folders) containing LUT files used by the LUT 3D module
auto-apply pixel workflow defaults: Choose which modules and module order are applied to new RAW image edits by default:

scene-referred (filmic) (default) assumes that most processing will be performed in a linear RGB color space. Selecting this option automatically enables the filmic rgb, exposure and color calibration modules for new edits and sets the module order to v3.0 RAW.

The exposure module includes an automatic exposure adjustment of +0.7 EV (to provide a midtone brightening comparable to the +0.5 to +1.2 EV typically added by in-camera tone curves), and automatically enables the “compensate camera exposure” option for the filmic workflow. Both of these settings are intended to provide a reasonable starting-point for RAWs produced by a broad range of SLR and mirrorless cameras, but they can be overridden with an automatically-applied preset if the defaults produce consistently dark images for your camera.

In the scene-referred workflows, the color calibration module acts in conjunction with the white balance module as the modern way to handle white balance and chromatic adaptation with improved color science. Note that when using the color calibration module, the white balance module needs to be active and set to “Camera Reference” mode (this will be done automatically and warnings will appear if the two modules have inconsistent settings). When using both modules as prescribed, it is still possible to auto-detect white-balance from a specific area of the image by selecting the CCT picker tool in the CAT tab of color calibration.
scene-referred (sigmoid) follows the same assumptions and overall flow as scene-referred (filmic), with the exception that it auto-enables the sigmoid module for tone mapping in place of filmic rgb.
display-referred (legacy) is the legacy mode (used by default in darktable 2.6 and earlier) and assumes that most processing will be performed in the Lab color space. Selecting this option automatically enables the base curve module for tone mapping and sets the module order to legacy. This workflow uses only the white balance module for chromatic adaptation.
none sets the module order to v3.0 RAW and uses the white balance module for chromatic adaptation. No other exposure or tone mapping modules are enabled by default.

auto-apply per camera basecurve presets: Use a per-camera base curve by default (if available) instead of the generic manufacturer one. This should only be used in conjunction with the display-referred workflow defined above (default off).
detect monochrome previews: Enable this option to analyse images during import and tag them with the darkroom|mode|monochrome tag if they are found to be monochrome. The analysis is based on the preview image embedded within the imported file. This makes for a more convenient workflow when working with monochrome images, but it slows down the import, so this setting is disabled by default.
show warning messages: Enable this option to display warning messages in processing modules where non-standard and possibly harmful settings have been used in the pipeline. Such messages can sometimes be false-positives (because of intentional non-standard settings) and can be disregarded if you know what you are doing. Disable to hide these warnings. (default on).

🔗CPU / memory

darktable resources

Choose how much of your system and graphics card (GPU) memory will be used by darktable. Four options are provided by default:

small takes roughly 20% of your system memory and 40% of your GPU memory. This might be acceptable on very large systems, especially if you’re not exporting images. Mostly, though, this can only be recommended if you are using a lot of other demanding applications at the same time as darktable.

default takes roughly 60% of your system memory and 70% of your GPU memory. This mode is recommended if you’re not exporting a lot of images, have at least 16Gb of system memory and 4Gb of GPU memory, and also are running a lot of other application at the same time as darktable.

large takes roughly 75% of your system memory and 90% of your GPU memory. This is the best option if you are only using darktable on your system and/or are exporting a lot of images.

unrestricted is not generally recommended. In this mode darktable may attempt to use more memory than your system has available. This might be possible if your system uses swapping when all of its system memory is taken, but it could lead to system instability. Use this mode with care, only when exporting very large images that darktable cannot otherwise handle.

See the memory & performance tuning section for more information.

prefer performance over quality

Enable this option to render thumbnails and previews at a lower quality. This increases the rendering speed by a factor of 4, and is useful when working on slower computers (default off). This also improves the performance of slideshow image rendering.

🔗OpenCL

activate OpenCL support

Your GPU can be used by darktable to significantly speed up processing. The OpenCL interface requires suitable hardware and matching OpenCL drivers on your system. If one of those is not found this option is grayed out. OpenCL support can be switched on and off at any time and takes immediate effect (default on).

OpenCL scheduling profile

Defines how preview and full pixelpipe tasks are scheduled on OpenCL enabled systems:

default: the GPU processes the center view pixelpipe; the CPU processes the preview pipe.

very fast GPU: both pixelpipes are processed sequentially on the GPU.

multiple GPUs: both pixelpipes are processed in parallel on different GPUs – see the multiple devices section for more information.

use all device memory

Enable this option to allow darktable to use all OpenCL memory on all devices except for a safety margin (headroom). The default headroom is 600MB and can also be specified on a per-device basis.

OpenCL drivers

In most cases darktable is able to find the correct OpenCL driver but this depends on how your operating system handles installation. Generally speaking, darktable must:

not use unreliable drivers

only have one active driver per hardware device. Problems are typically found where the vendor-provided driver is installed as well as another driver (like rusticl) or (on Windows) where the OpenCLon12 driver has been installed via the OpenCL Compatibility Pack.

Toggle-switches are offered for most available drivers, though on more exotic hardware like ARM boards you will have to enable the “other platforms” fallback option. Select the drivers you want to use from the list. If you suspect that a driver is malfunctioning you can explicitly disable it here.

See the memory & performance tuning section for more information.