EXR Format Guide

Convert from EXR

Convert to EXR

About EXR Format

EXR (OpenEXR) is an open-standard high dynamic range raster image format created by Industrial Light & Magic (ILM) and released as open source in 2003. The format was originally developed for the visual effects industry to address the limitations of existing HDR formats, providing a robust container for storing images with 16-bit half-float or 32-bit full-float precision per channel. OpenEXR supports an arbitrary number of channels (RGB, RGBA, depth, motion vectors, cryptomatte, and custom AOVs), multi-layer and multi-view images, deep data (multiple samples per pixel for volumetric effects), and both lossless (PIZ, ZIP, ZIPS, RLE, B44) and lossy (DWAA, DWAB) compression methods. The .exr file extension is universally recognized across VFX, film, animation, and game development software. OpenEXR's ability to store the full dynamic range of a scene -- from the darkest shadows to the brightest specular highlights -- with floating-point precision makes it the industry standard for compositing, rendering, and color grading in professional film and television production.

History of EXR

OpenEXR was developed internally at Industrial Light & Magic beginning in the late 1990s, driven by the need for a high-fidelity image format that could handle the demanding requirements of feature film visual effects production. The format was first used in production on films like Star Wars: Episode II -- Attack of the Clones and Harry Potter and the Sorcerer's Stone. ILM's Florian Kainz led the development effort, creating a format that could store images with 16-bit half-float precision (IEEE 754 half-precision) -- a crucial innovation that provided sufficient dynamic range for compositing while keeping file sizes manageable compared to 32-bit float alternatives. In January 2003, ILM released OpenEXR as open source under a modified BSD license, making the reference implementation freely available to the entire visual effects and computer graphics community. The format was rapidly adopted by major VFX studios including Weta Digital, Pixar, and DreamWorks Animation. Version 2.0, released in 2013, introduced deep data support (storing multiple depth samples per pixel for improved compositing of volumetric effects like smoke and fog) and multi-part files (storing multiple images with different compression and channel configurations in a single file). The Academy Software Foundation (ASWF) took over stewardship of the project in 2019, and version 3.0 was released in 2021 with a completely rewritten core library (Imath), improved security, and better performance. Today, OpenEXR is the de facto standard for intermediate image storage in virtually every major VFX and animation studio worldwide.

Key Features and Uses

OpenEXR stores pixel data using either 16-bit half-float (IEEE 754 binary16) or 32-bit float (IEEE 754 binary32) precision per channel. The half-float format provides a dynamic range of approximately 10 orders of magnitude with 10 bits of mantissa precision, while 32-bit float offers the full IEEE single-precision range. The format supports an arbitrary number of named channels, allowing storage of not just RGBA color but also depth (Z), motion vectors (forward and backward), object IDs, cryptomatte data, and any number of custom arbitrary output variables (AOVs) used in modern rendering pipelines. Multiple compression methods are available: PIZ (wavelet-based lossless, best for noisy CG renders), ZIP and ZIPS (zlib-based lossless, good general-purpose), RLE (run-length encoding for simple images), B44 and B44A (fixed-rate lossy for real-time playback), and DWAA/DWAB (DCT-based lossy with configurable quality, excellent compression ratios for delivery). The tiled storage mode enables efficient random access to image regions, essential for texture mapping and interactive viewing. Multi-resolution support (mipmaps and ripmaps) allows a single file to contain multiple resolution levels. Deep image support stores variable-length lists of samples per pixel, enabling improved compositing of volumetric effects and transparent surfaces without the traditional limitations of flat compositing.

Common Applications

OpenEXR is the backbone of modern visual effects and animation production pipelines. In film VFX, every major studio uses EXR as the primary format for render outputs, storing beauty passes alongside dozens of AOVs (diffuse, specular, reflection, refraction, ambient occlusion, shadow, depth, motion vectors, cryptomatte) that compositors combine in Nuke, Fusion, or After Effects. 3D rendering engines including Arnold, V-Ray, RenderMan, Cycles (Blender), and Unreal Engine all output EXR natively. Game developers use EXR for HDRI environment maps, lightmaps, and offline baking of complex lighting data that gets converted to runtime formats. Digital intermediate (DI) and color grading workflows in DaVinci Resolve and Baselight frequently use EXR for scene-referred linear light data. Photogrammetry and 3D scanning pipelines store depth maps and normal maps in EXR for maximum precision. Scientific visualization, medical imaging, and astronomical imaging benefit from EXR's floating-point precision for accurate data representation. The format is also used for texture painting in tools like Mari and Substance Painter, where high bit-depth prevents banding artifacts in subtle gradients and lighting variations.

Advantages and Disadvantages