Need help with tev?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

484 Stars 50 Forks BSD 3-Clause "New" or "Revised" License 724 Commits 15 Opened issues


High dynamic range (HDR) image comparison tool for graphics people. With an emphasis on OpenEXR images.

Services available


Need anything else?

Contributors list

# 87,488
610 commits
# 141,557
4 commits
# 118,251
3 commits
# 21,519
2 commits
# 461,318
1 commit
# 369,370
1 commit

tev — The EXR Viewer

A high dynamic range (HDR) image comparison tool for graphics people. tev allows viewing images through various tonemapping operators and inspecting the values of individual pixels. Often, it is important to find exact differences between pairs of images. For this purpose, tev allows rapidly switching between opened images and visualizing various error metrics (L1, L2, and relative versions thereof). To avoid clutter, opened images and their layers can be filtered by keywords.

New: tev can display true HDR on Apple extended dynamic range (EDR) and 10-bit displays.

While the predominantly supported file format is OpenEXR certain other types of images can also be loaded. The following file formats are currently supported: - EXR (via OpenEXR) - PFM (compatible with Netbpm) - DDS (via DirectXTex; Windows only. Shoutout to Craig Kolb for adding support!) - Supports BC1-BC7 compressed formats. - Low-dynamic-range (LDR) images are "promoted" to HDR through the reverse sRGB transformation. - HDR, BMP, GIF, JPEG, PIC, PNG, PNM, PSD, TGA (via stb_image) - stb_image only supports subsets of each of the aforementioned file formats. - Low-dynamic-range (LDR) images are "promoted" to HDR through the reverse sRGB transformation.


Screenshot A false-color comparison of two multi-layer OpenEXR images of a beach ball. Image courtesy of openexr-images.


Graphical User Interface

Images can be opened via a file dialog or by dragging them into tev. They can be reloaded, closed, or filtered at any time, so don't worry about opening more images than exactly needed.

Select an image by left-clicking it, and optionally select a reference image to compare the current selection to by right-clicking. For convenience, the current selection can be moved with the Up/Down or the 1-9 keys. For a comprehensive list of keyboard shortcuts click the little "?" icon at the top (or press "h").

If the interface seems overwhelming, you can hover any controls to view an explanatory tooltip.

Command Line

tev takes images as positional command-line arguments:

$ tev foo.exr bar.exr

By default, all layers and channels are loaded, but individual layers or channels can also be specified. In the following example, the depth layer of foo.exr and the r, g, and b channels of foo.exr and bar.exr are loaded.

$ tev :depth foo.exr :r,g,b foo.exr bar.exr

Other command-line arguments exist (e.g. for starting tev with a pre-set exposure value). For a list of all arguments simply invoke

$ tev -h

Over the Network

tev can also be controlled remotely over the network using a simple TCP-based protocol.


argument specifies the IP and port tev is listening to. By default, tev only accepts connections from localhost (
), which is useful, for example, as frontend for a supported renderer like pbrt version 4.

The following operations exist:

| Operation | Function | :--- | :---------- |

| Opens an image from a specified path on the machine tev is running on. |
| Creates a blank image with a specified name, size, and set of channels. |
| Updates the pixels in a rectangular region. |
| Closes a specified image. |
| Reloads an image from a specified path on the machine tev is running on.

tev's network protocol is already implemented in the following languages: - Python by Tomáš Iser - Rust by Karel Peeters

If using these implementations is not an option, it's easy to write your own one. Each packet has the simple form

[uint32_t total_length_in_bytes][byte operation_type][byte[] operation_specific_payload]
where integers are encoded in little endian.

There are helper functions in Ipc.cpp (

) that show exactly how each packet has to be assembled. These functions do not rely on external dependencies, so it is recommended to copy and paste them into your project for interfacing with tev.

Obtaining tev

macOS / Windows

Pre-built binaries for Windows (32-bit and 64-bit) and macOS (64-bit) are available on the releases page.

On macOS, tev can also be installed via homebrew:

brew install --cask tev


Building tev

All that is required for building tev is a C++17-compatible compiler. Begin by cloning this repository and all its submodules using the following command:

$ git clone --recursive

If you accidentally omitted the

flag when cloning this repository you can initialize the submodules like so:
$ git submodule update --init --recursive

tev uses CMake as its build system. The following sections detail how it should be used on various operating systems.

macOS / Linux

On macOS and most Linux distributions CMake can be obtained via a package manager (Homebrew on macOS, apt on Ubuntu/Debian, etc.). Most Linux distributions additionally require xorg, gl, and zlib development packages and zenity. On Ubuntu/Debian simply call

$ apt-get install cmake xorg-dev libglu1-mesa-dev zlib1g-dev zenity

Once all dependencies are installed, create a new directory to contain build artifacts, enter it, and then invoke CMake with the root tev folder as argument as shown in the following example:

$ mkdir build
$ cd build
$ cmake ..

Afterwards, tev can be built and installed via

$ make -j
$ make install


On Windows, install CMake, open the included GUI application, and point it to the root directory of tev. CMake will then generate Visual Studio project files for compiling tev. Make sure you select at least Visual Studio 2017 or higher!


tev is available under the BSD 3-clause license, which you can find in the
file. TL;DR you can do almost whatever you want as long as you include the original copyright and license notice in any copy of the software and the source code.

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.