Openexr

production-ready release since OpenEXR 1.0.7. If you have been using
the development 1.1 series, please switch to 1.2.0 as soon as
possible. We believe that OpenEXR 1.2.0 is ready for use in shipping
applications. We have been using it in production at ILM for several
months now with no problems. There are quite a few major new features
in the 1.2 series as compared to the original 1.0 series:

* Support for tiled images, including mipmaps and ripmaps. Note that
software based on the 1.0 series cannot read or write tiled
images. However, simply by recompiling your software against the 1.2
release, any code that reads scanline images can read tiled images,
too.

* A new Pxr24 compressor, contributed by Pixar Animation
Studios. Values produced by the Pxr24 compressor provide the same
range as 32-bit floating-point numbers with slightly less precision,
and compress quite a bit better. The Pxr24 compressor stores UINT
and HALF channels losslessly, and for these data types performs
similarly to the ZIP compressor.

* OpenEXR now supports high dynamic-range YCA (luminance/chroma/alpha)
images with subsampled chroma channels. These files are supported
via the RGBA convenience interface, so that data is presented to the
application as RGB(A) but stored in the file as YC(A). OpenEXR also
supports Y and YA (black-and-white/black-and-white with alpha)
images.

* An abstracted file I/O interface, so that you can use OpenEXR with
interfaces other than C++'s iostreams.

* Several new utilities for manipulating tiled image files.

Detailed Changes:

* Production-ready release.

* Disable long double warnings on OS X. (Drew Hess)

* Add new source files to VC7 IlmImfDll target. (Drew Hess)

* Iex: change the way that APPEND_EXC and REPLACE_EXC modify
their what() string to work around an issue with Visual C++
7.1. (Florian Kainz, Nick Porcino)

* Bumped OpenEXR version to 1.2 and .so versions to 2.0.0 in
preparation for the release. (Drew Hess)

* Imath: fixed ImathTMatrix.h to work with gcc 3.4. (Drew Hess)

* Another quoting fix in openexr.m4. (Drew Hess)

* Quoting fix in acinclude.m4 for automake 1.8. (Brad Hards)

* Imath: put inline at beginning of declaration in ImathMatrix.h
to fix a warning. (Ken McGaugh)

* Imath: made Vec equalWithError () methods const.

* Cleaned up compile-time Win32 support. (Florian Kainz)

* Bug fix: Reading a particular broken PIZ-compressed file
caused crashes by indexing off the end of an array.
(Florian Kainz)

expect to release a stable version, 1.2, around the end of
April. Version 1.1.1 includes support for PXR24 compression, and for
high-dynamic-range luminance/chroma images with subsampled chroma
channels. Version 1.1.1 also fixes a bug in the 1.1.0 tiled file
format.

Detailed Changes:

* Half: operator= and variants now return by reference rather than by
value. This brings half into conformance with built-in types.
(Drew Hess)

* Half: remove copy constructor, let compiler supply its own. This
improves performance up to 25% on some expressions using half.
(Drew Hess)

* configure: don't try to be fancy with CXXFLAGS, just use what the
user supplies or let configure choose a sensible default if CXXFLAGS
is not defined.

* IlmImf: fixed a bug in reading scanline files on big-endian
architectures. (Drew Hess)

* exrmaketiled: Added an option to select compression type. (Florian
Kainz)

* exrenvmap: Added an option to select compression type. (Florian
Kainz)

* exrdisplay: Added some new command-line options. (Florian Kainz)

* IlmImf: Added Pixar's new "slightly lossy" image compression method.
The new method, named PXR24, preserves HALF and UINT data without
loss, but FLOAT pixels are converted to a 24-bit representation.
PXR24 appears to compress FLOAT depth buffers very well without
losing much accuracy. (Loren Carpenter, Florian Kainz)

* Changed top-level LICENSE file to allow for other copyright holders
for individual files.

* IlmImf: TILED FILE FORMAT CHANGE. TiledOutputFile was incorrectly
interleaving channels and scanlines before passing pixel data to a
compressor. The lossless compressors still work, but lossy
compressors do not. Fix the bug by interleaving channels and
scanlines in tiled files in the same way as ScanLineOutputFile does.
Programs compiled with the new version of IlmImf cannot read tiled
images produced with version 1.1.0. (Florian Kainz)

* IlmImf: ImfXdr.h fix for 64-bit architectures. (Florian Kainz)

* IlmImf: OpenEXR now supports YCA (luminance/chroma/alpha) images
with subsampled chroma channels. When an image is written with the
RGBA convenience interface, selecting WRITE_YCA instead of
WRITE_RGBA causes the library to convert the pixels to YCA format.
If WRITE_Y is selected, only luminance is stored in the file (for
black and white images). When an image file is read with the RGBA
convenience interface, YCA data are automatically converted back to
RGBA. (Florian Kainz)

* IlmImf: speed up reading tiled files as scan lines.
(Florian Kainz)

* Half: Fixed subtle bug in Half where signaling float NaNs
were being converted to inf in half. (Florian Kainz)

* gcc 3.3 compiler warning cleanups. (various)

* Imath: ImathEuler.h fixes for gcc 3.4. (Garrick Meeker)

support for tiled images, multi-resolution files (mip/ripmaps),
environment maps, and abstracted file I/O. We've also released a new
set of images that demonstrate these features, and updated the
CodeWarrior project and Photoshop plugins for this release. See the
downloads section for the source code and the new images.

Detailed Changes:

* Added new targets to Visual C++ .NET 2003 project
for exrmaketiled, exrenvmap, exrmakepreview, and exrstdattr.
(Drew Hess)

* A few assorted Win32 fixes for Imath. (Drew Hess)

* GNU autoconf builds now produce versioned libraries.
This release is 1:0:0. (Drew Hess)

* Fixes for Visual C++ .NET 2003. (Paul Schneider)

* Updated Visual C++ zlib project file to zlib 1.2.1.
(Drew Hess)

* exrdisplay: Fixed fragment shader version. (Drew Hess)

* *Test: Fixed some compiler issues. (Drew Hess)

* Imath: Handle "restrict" keyword properly. (Drew Hess)

* IlmImfExamples: Updated to latest versions of example
source code, includes tiling and multi-res images.
(Florian Kainz)

* exrmakepreview: A new utility to create preview images.
(Florian Kainz)

* exrenvmap: A new utility to create OpenEXR environment
maps. (Florian Kainz)

* exrstdattr: A new utility to modify standard
attributes. (Florian Kainz)

* Updated exrheader to print level rounding mode and
preview image size. (Florian Kainz)

* Updated exrmaketiled to use level rounding mode.
(Florian Kainz)

* IlmImf: Changed the orientation of lat-long envmaps to
match typical panoramic camera setups. (Florian Kainz)

* IlmImf: Fixed a bug where partially-completed files with
DECREASING_Y could not be read. (Florian Kainz)

* IlmImf: Added support for selectable rounding mode (up/down)
when generating multiresolution files. (Florian Kainz)

* exrdisplay: Support for tiled images, mip/ripmaps, preview
images, and display windows. (Florian Kainz, Drew Hess)

* exrmaketiled: A new utility which generates tiled
versions of OpenEXR images. (Florian Kainz)

* IlmImf: Changed Imf::VERSION to Imf::EXR_VERSION to
work around problems with autoconf VERSION macro
conflict. (Drew Hess)

* exrheader: Support for tiles, mipmaps, environment
maps. (Florian Kainz)

* IlmImf: Environment map support. (Florian Kainz)

* IlmImf: Abstracted stream I/O support. (Florian Kainz)

* IlmImf: Support for tiled and mip/ripmapped files;
requires new file format. (Wojciech Jarosz, Florian Kainz)

* Imath: **TMatrix**, generic 2D matricies and algorithms.
(Francesco Callari)

* Imath: major quaternions cleanup. (Cary Phillips)

* Imath: added GLBegin, GLPushAttrib, GLPushMatrix objects
for automatic cleanup on exceptions. (Cary Phillips)

* Imath: removed implicit scalar->vector promotions and vector
comparisons. (Nick Rasmussen)

version adds support for some new standard attributes, such as primary
and white point chromaticities, lens aperture, film speed, image
acquisition time and place, and more. If you want to use these new
attributes in your applications, see the ImfStandardAttributes.h
header file for documentation.

Our project hosting site, Savannah, is still recovering from a
compromise last month, so in the meantime, we're hosting file
downloads here. Some of the files are not currently available, but
we're working to restore them.

Detailed Changes:

* Fixed a typo in one of the IlmImfTest tests. (Paul Schneider)

* Fixed a bug in exrdisplay that causes the image to display
as all black if there's a NaN or infinity in an OpenEXR
image. (Florian Kainz)

* Updated exrheader per recent changes to IlmImf library.
(Florian Kainz)

* Changed an errant float to a T in ImathFrame.h nextFrame().
(Cary Phillips)

* Support for new "optional standard" attributes
(chromaticities, luminance, comments, etc.).
(Florian Kainz, Greg Ward, Joseph Goldstone)

* Fixed a buffer overrun in ImfOpaqueAttribute. (Paul Schneider)

* Added new function, isImfMagic (). (Florian Kainz)

* Added README.win32 to disted files.

* Fixed OpenEXR.pc.in pkg-config file, OpenEXR now works
with pkg-config.

* Random fixes to readme files for new release.

* Fixed openexr.m4, now looks in /usr by default.

* Added Visual Studio .NET 2003 "solution."

* Fixes for Visual Studio .NET 2003 w/ Microsoft C++ compiler.
(Various)

* Random Imath fixes and enhancements. Note that extractSHRT now
takes an additional optional argument, see ImathMatrixAlgo.h for
details. (Various)

* Added Wojciech Jarosz to AUTHORS file.

* Added test cases for uncompressed case, preview images,
frame buffer type conversion. (Wojciech Jarosz,
Florian Kainz)

* Fix a bug in IlmImf where uncompressed data doesn't get
read/written correctly. (Wojciech Jarosz)

* Added support for preview images and preview image
attributes (thumbnail images) in IlmImf. (Florian Kainz)

* Added support for automatic frame buffer type conversion
in IlmImf. (Florian Kainz)

* Cleaned up some compile-time checks.

* Added HalfTest unit tests.

* [exrdisplay] Download half framebuffer to texture memory
instead of converting to float first. Requires latest
Nvidia drivers.

Industrial Light & Magic has released the source code for an OpenEXR
display driver for Pixar's Renderman. This display driver is covered
under the OpenEXR free software license. See the downloads section for
the source code.

Detailed Changes:

* Fixed IlmImf.dll to use static runtime libs (Andreas).

* Added exrheader project to Visual Studio 6.0 workspace.

* Added some example code showing how to use the IlmImf library.
(Florian)

* Use DLL runtime libs for Win32 libraries rather than static
runtime libs.

* Add an exrdisplay_fragshader project to the Visual Studio 6.0
workspace to enable fragment shaders in Win32.

* Add an IlmImfDll project to the Visual Studio 6.0 workspace.

* In Win32, export the ImfCRgbaFile C interface via a DLL so that
Visual C++ 6.0 users can link against an Intel-compiled IlmImf.
(Andreas Kahler)

* Use auto_ptr in ImfAutoArray on Win32, it doesn't like large
automatic stacks.

* Performance improvements in PIZ decoding, between
20 and 60% speedup on Athlon and Pentium 4 systems.
(Florian)

* Updated the README with various information, made
some cosmetic changes for readability.

* Added fragment shader support to exrdisplay.

* Bumped the version to 1.0.5 in prep for release.

* Updated README and README.OSX to talk about CodeWarrior
project files.

* Incorporated Rodrigo Damazio's patch for an openexr.m4
macro file and an openexr.spec file for building RPMs.

* Small change in ImfAttribute.h to make IlmImf compile with gcc 2.95.

* Updated ImfDoubleAttribute.h for Codewarrior on MacOS.

* Added exrheader utility.

* Update to AUTHORS file.

* Added a README.win32 file.

* Added project files for Visual Studio 6.0.

* Initial Win32 port. Requires Visual Studio 6.0 and Intel C++
compiler version 7.0.

* Added new intersectT method in ImathSphere.h

* Fixed some bugs in ImathQuat.h

* Proper use of fltk-config to get platform-specific FLTK
compile- and link-time flags.

* exrdisplay uses Imath::Math<T>::pow instead of powf now.
powf is not available on all platforms.

* Roll OS X "hack" into the source until Apple fixes their
istream implementation.