CalibRaw


CalibRaw — Free Download. RAW photo editor

CalibRaw is a non-destructive RAW image processor that performs all rendering through GPU compute shaders. The application provides a photo library with culling tools, complete color and tone controls including curves and color grading, local adjustment masks, lens corrections, HDR merging, advanced demosaicing for Bayer and Fujifilm X-Trans sensors, and a set of creative effects. Optional artificial intelligence instruments operate offline on the device for subject, sky, and object masking, noise reduction, and object removal. Edits are stored in sidecar files, so original RAW data remains unaltered.

★★★★★
5.0(1 ratings)
File size: 15.1 MB
The latest version of CalibRaw is: 1.2.1
Operating system: Windows, Mac OS, Linux, Android
Languages: English
Price: $0.00 USD (Open Source (GPL-3.0))

What’s new in CalibRaw

  1. Android users can now send photos to CalibRaw from other apps via the Open with or Share options
  2. Exporting and sharing a photo can now be done in a single action on Android
  3. A button in the Export panel lets you replay your edits on Android
Full update (+9 more) →
  • Point Color. This instrument samples a specific color from the image using an eyedropper and creates a custom Hue, Saturation, and Luminance range around that sample. The operator can refine each component of the range independently and preview the affected pixels through a live selection mask overlay. Adjustments apply only to the sampled color range, which permits precise recoloring or luminance correction of individual objects or tones without altering other parts of the frame.
  • 8-Band Color Mixer. This control divides the color spectrum into eight distinct bands, each corresponding to a primary or intermediate color region. For every band, the operator adjusts Hue, Saturation, and Luminance values independently. The mixer provides a structured method for color correction, allowing selective modification of skin tones, foliage, sky, or any other color group without influencing neighboring bands.
  • 3-Way Color Grading. This module offers four color wheels: one for shadows, one for midtones, one for highlights, and one global wheel. All wheels map their adjustments through the Oklab color space, which maintains perceptual uniformity during tonal transitions. The operator can introduce color casts or neutralize existing ones in specific luminance ranges, achieving complex color balances that resemble cinematic grading workflows.
  • Tone Curves. The curve panel contains a luminance spline and three individual splines for the red, green, and blue channels. Each curve operates with monotonic limits, which prevents the creation of overshoots or inversions that could produce artifacts. The operator shapes contrast, brightness, and color response across the entire tonal range by placing and dragging control points on the graph.
  • Sigmoid Tone Mapping. This tone mapping operator applies a smooth, film-like compression to highlight values, avoiding the harsh clipping that conventional curves can produce. It works in conjunction with Opposed and LCh highlight recovery methods, which reconstruct detail from channels that are not clipped. The result is a gradual roll-off into white with preserved texture in bright areas.
  • Bayer RCD Demosaicing. For image sensors that use the standard Bayer color filter array, this algorithm performs directional interpolation based on the red-chroma-difference method. It examines edge directions in the raw data and interpolates missing color values along the dominant direction, which reduces color fringing and zipper artifacts on fine details. The processing occurs in a compute shader for fast execution.
  • Fujifilm X-Trans Demosaicing. Sensors with the X-Trans color filter array require a specialized algorithm because of their irregular pixel pattern. This implementation uses a directional derivative and homogeneity demarcation pipeline, which identifies the direction of image structures and interpolates accordingly. The method preserves fine details and reduces moiré patterns that commonly appear with conventional demosaicing on X-Trans sensors.
  • Dual Demosaicing. This mode combines two demosaicing passes: one that prioritizes edge preservation and another that applies smoothing. The algorithm blends the results to suppress chroma noise and artifacts that appear in high-ISO images. The operator can select this method when working with noisy raw files where single-pass demosaicing produces color speckles or false detail.
  • Lens Correction. This instrument uses the Lensfun database to automatically identify the camera and lens combination from metadata. It then applies correction profiles for geometric distortion, lateral chromatic aberration, and vignetting. The operator can enable or disable each correction type individually and adjust the strength of the applied profile when the automatic value does not match the desired result.
  • HDR Merge. This function combines multiple bracketed RAW exposures of the same scene into a single image with extended dynamic range. The algorithm aligns the frames, accounts for exposure differences, and merges the data into a high-bit-depth result. The operator can then process the merged file through the standard tone and color tools without the risk of clipping highlights or crushing shadows.
  • Masks. The masking system includes brush, linear gradient, radial gradient, and shape masks. Each mask carries its own set of adjustment controls, including curves and color tools, which operate independently from global adjustments. The operator can stack multiple masks on one image and refine their shapes, feathering, and opacity to target specific regions with precision.
  • Local AI Tools. This collection of offline models includes subject masking, sky masking, and object masking, all of which run on the local device without network access. Additional AI instruments provide denoising and object removal. The operator applies these tools to isolate complex subjects or remove unwanted elements, then combines the results with manual masks or global adjustments.
  • GPU Creative Effects. This set of stackable effect cards applies to either the full image or individual masks. The available effects include Light Rays, Lens Blur with adjustable bokeh characteristics, Motion Blur, Radial Blur, Fog, Smoke, Glow, and Neon. The operator arranges and combines these effects to construct atmospheric or stylistic treatments without leaving the application.
  • Tiled Export and CLI. For large RAW files that exceed available GPU memory, the export system splits the rendering into tiles and processes each tile sequentially. A command-line tool provides the same rendering pipeline for batch operations, which allows the operator to process multiple files or integrate the application into automated workflows without launching the graphical interface.
  • JPEG XL Export. This export path produces lossless .jxl files at either 8-bit or 16-bit precision. The operator can include EXIF metadata in the exported file when needed. JPEG XL provides efficient compression while retaining full image quality, which suits archival purposes or workflows that require a compact but lossless output format.

CalibRaw is developed by Duecki1 with contributions from the open-source imaging community. The project began in 2026 as a native Rust application built with egui and wgpu, using WGSL compute shaders for image processing. The codebase is organized as a modular Rust workspace with separate crates for core logic, GPU pipelines, artificial intelligence models, user interface, command-line tools, and Android integration. The development draws inspiration from darktable for algorithms such as Bayer RCD, X-Trans Markesteijn, Opposed highlight recovery, and Sigmoid tone mapping, and from GIMP and Ansel for inpainting and color normalization methods. LibRaw, Rawler, and Lensfun provide RAW decoding, DNG metadata extraction, and lens profile databases.