libDSP News

Dec 17, 2005		5.1.0 development version.
			Changed to use libcommoncpp and got rid of semi-
			external modules like Inlines, etc.
			Also added SSE3 optimizations.

May 13, 2005		5.0.2 stable version.
			Fixed a bug in SSE2 complex multiply-add. Support for
			gcc-4.x. Number of other minor fixes.

February 15, 2005	5.0.1 stable version.
			Added assembler optimized version of complex multiply-add.

November 7, 2004	5.0.0 stable version.
			Implemented DCT into clDSPOp and also a new lockless
			ringbuffer class clRingBuffer. Fixes has been made
			for gcc-3.4.x support.

February 29, 2004	4.9.2 development version.
			Extended support for SSE2, also more
			assembler optimizations for x86-64.

October 26, 2003	4.9.1 development version.
			Removed obsoleted interfaces (clDecimator, clFilter,
			clReBuffer2). Renamed clFilter2 -> clFilter and
			clReBuffer3 -> clReBufferT. Added a class for
			flipping odd frequency bands for use with
			undersampling/quadrature modulation/decimation.

September 21, 2003	RPM packages released.

September 7, 2003	4.9.0 development version.
			Support for Intel's Integrated Performance Primitives
			(IPP). FFTW support updated to match version 3 API.
			Various code cleanups.

June 16, 2003		4.6.0 release version.
			Bugfixes, added resampling by non-integer factor
			with and without averaging. Tested on Solaris/SPARC.

October 5, 2002		4.5.1 development version.
			Bugfixes, optimizations and optimized multiply-add.

August 11, 2002		4.5.0 development version.
			New baseclass for multirate filters and new
			interpolator classes. Filters for multirate filters can
			be selected from FFT/FIR/IIR. Support for IIR filters
			have been added. Also number of smaller changes have
			been made.

August 10, 2002		It's now possible to make donations for the libDSP
			project using PayPal. Link for doing so is on
			the licensing page.

May 25, 2002		4.4.2, support for GCC 3.1 and Intel C++ v6,
			clReBuffer3 template bugfix, small filter improvements
			based on tests and fuction to get filter delay.

April 14, 2002		4.4.1, number of smaller bugfixes and some finetuning
			of the clFilter2 class.

February 16, 2002	4.4.0, wrapper for C, clReBuffer bug fix and smaller
			bugfixes and documentation updates.

December 29, 2001	4.3.0, transform updates, FFT filter improvements
			(automatic size and adjustable overlap), Dolph-
			Chebyshev window and support for Intel C++ compiler.

October 20, 2001	4.2.0, optimizations, E3DNow! (Athlon/Duron) support
			and better complex root.

September 27, 2001	4.1.2, bugfixes and optimizations in complex exp(),
			log(), log10(), pow() and root() functions.

July 26, 2001		4.1.1, just changed the default compiler in Makefile
			from ng++ to g++. Documentation updated.

July 24, 2001		4.1.0, DSP vector and rebuffer C++ templates. Makefile
			updates for *BSDs.

June 24, 2001		4.0.0, new decimation filter classes, new recursive
			decimation class and new rebuffering classes. Also
			number of small fixes and optimizations.

April 16, 2001		3.3.8, Only documentation updates.
			Because I'm constantly exceeding my current web space,
			this project is now hosted on SourceForge.

April 1, 2001		3.3.7, New documentation, QNX port and fixes in
			delayed crosscorrelation (normalization factor).

March 10, 2001		3.3.6, Performance and precision improvements.

February 10, 2001	3.3.5, Hankel (modified) transform and bugfixes.

January 1, 2001		3.3.4, new FIR decimation filter set and bugfixes.

December 26, 2000	3.3.3, cleanups and clFilter::GetCoeffs().

December 17, 2000	3.3.1, FFT filter optimizations, cleanups and Pack().

December 3, 2000	Started this page.
			3.3.0.1 snapshot, better code split from HASAS and
			various cleanups.