2012-01-18: Moved the source code development repository to SourceForge.net.
2009-12-03: The fftconv_p and sum_fftconv_p functions crashes (in the calls to fftw_malloc) using Matlab R2009a and R2009b on Linux 64-bit systems. The export MATLAB_MEM_MGR=system trick does not solve the problem (as it does on older versions of Matlab). There is currently no solution to the problem and if these functions are crucial then use GNU Octave (or an older version of Matlab) instead.
2009-11-13: Version 2.1.3 Release.
- Updated build scripts and dll files for the Windows pthread and FFTW libs for 64-bit Windows.
- Switched to using the gcc compiler (MinGW) to build the GNU Octave oct-files for 32-bit windows. Now the Windows Octave (>= 3.2.2) binary includes the compiler so one no longer need to use (and install) the MSVC Express tools.
- Fixed a documentation bug where the steering angles was given in radians [rad] in the documentation (and help files) which should have been degrees [deg] (thanks to A.-P. B-Dionne).
- Updated the Linux build scripts.
- Updated various m-files.
2008-10-15: Version 2.1.2 Release.
- Experimental 64-bit Matlab support.
- Experimental 64-bit FFTW and Pthread-Win32 libraries for 64-bit Windows available in the download section.
- Updated the user manual with instructions for building the toolbox for 64-bit Matlab on Windows.
- Switched from using fftw-3.1.2 to fftw-3.1.3.
- Changed the FFTW plan setup for fftconv_p and sum_fftconv_p. Now these functions use the (fast) ESTIMATE method when only one output arg is used and the (more accurate and slow) MEASURE method when two output args are used (for more info see the FFTW documentation at http://www.fftw.org); previously only the MEASURE method was used.
2008-10-05: Version 2.1.1 Release.
- Updated Windows build scripts (updated for MSVC 2008 Express Edition and for MinGW 2.4.5/Gnumex 2.0.1)
- Added prebuild FFTW and Pthread-Win32 libraries for Windows to the download section.
- Updated build instructions in the user manual.
2008-02-29: Version 2.1.0 Release.
- MacOS X Support.
- Octave package manager (pkg) install support.
- Improved user manual with a tutorial section.
- Improved build scripts (with hardware auto-detection).
- Three new analytic transducer functions.
- New delay-and-sum and SAFT functions.
- New example scripts for model based imaging and SAFT/delay-and-sum imaging.
- Many bug fixes (in particular for the Octave interface).
2007-02-12: Version 2.0.0 Release. The DREAM toolbox is now open source and should compile on most systems
- Source code available (GPLed)
- Octave support (again)
- 64 bit support (compiled with gcc's -m64 flag)
- Two new parallel functions (sum_fftconv_p and copy_p)
- Many bug fixes.
2006-11-15: The FFTW crash was due to Matlab's memory management routines not being thread safe. A workaround is to set the env. variable MATLAB_MEM_MGR=system (i.e., export MATLAB_MEM_MGR=system).
2006-05-16: There is a conflict between the FFTW library that comes with Matlab and the standard FFTW lib (from http://www.fftw.org ). Therefore, using the fftconv_p function causes a crasch if more than one CPU (thread) is used on SMP systems.
2006-01-15: Version 1.1.1 release. Removed all refs to Octave and added missing PGN files for the DREAM GUI.
2005-11-03: Octave support has been removed since the GPL licence that Octave uses don't allow for binary only distribution.
2005-10-31: Version 1.1.0 Release .
- Octave support (Linux x86)
- Solaris support
- A new fft-based convolution function
- Bug fixes, etc.
2004-02-28: Updated the DREAM generated movies.
2004-02-24: DREAM generated movies added to the examples section.
2004-01-12: First release of the DREAM toolbox : version 1.0.0.