ChangeLog

#------------------------------------------------------------------------------
#
# $Id: ChangeLog 926 2015-08-23 17:36:20Z weegreenblobbie $
#
#  Nsound is a C++ library and Python module for audio synthesis featuring
#  dynamic digital filters. Nsound lets you easily shape waveforms and write
#  to disk or plot them. Nsound aims to be as powerful as Csound but easy to
#  use.
#
#------------------------------------------------------------------------------

Key:
+ Finished item
- Planned item

2021-12-30 Nsound 0.9.5
    + Lots of updates for building and linking with Python3
    + Added github Actions to build on Linux, MacOs and Windows
    + Added FilterMedian
    + Updated Plotter to use latest matplotlib.pyplot
    + Added an Intersteller inspired example: src/examples/interstellar.cc


2015-08-23 Nsound 0.9.4

    + C++11 now required for compiling Nsound
    + Fixed an AudioStream pickling bug, wasn't loading n channels or sample rate
    + Switched to setuptools to build wheel packages on windows (maybe os x)
    + Moved circular_iterator out of Buffer.h
    + Fixed #76: DelayLine: add dynamic delay
    + Fixed #63: Add 'classic' filters with resonance control
        FilterLowPassMoogVcf added
    + AudioStream now using std::vector<Buffer *> instead of raw array
    + Fixed Python expressions for returning references
    + Fixed Python expressions for iterating over Buffer in AudioStreams
    + Reworked circular_iterator, added unit tests
    + Created delayline demo video: https://www.youtube.com/watch?v=9m3KA6JBDH8
    + Created tapedeck demo video: https://www.youtube.com/watch?v=Al6aHlwyEuA


2015-06-13 Nsound 0.9.3

    + Fixed #64: Nsound is now available as a source package from PyPI!
    + Fixed #74: Can't compile c++ lib if swig is not installed
    + Fixed #75: setup.py: use shell environment CC/CXX if defined
    + Some fixes to AudioPlaybackRt


2015-01-01 Nsound 0.9.2

    + Created buildbot config running on my local network [#53]
    + Fixed building on GCC 4.6 (Ubuntu 12) [#65]
    + Fixed compiling without portaudio [#66]
    + Fixed segfault with operator[] [#70]
    + Fixed python module errors with operator<< [#71]
    + AudioStreams and Buffers can now be Pickeled in Python
    + Added a CircularBuffer class which returns the last N samples on read
    + Added src/example/bebot-disp.py which adds a real-time oscilloscope and
      spectrogram display: https://www.youtube.com/watch?v=j4uLvl68PtI
    + Added StreamOperators.h to support serializing Buffers & AudioStream
      (Pickling in Python)


2014-07-01 Nsound 0.9.1

    + Real-time audio playback via AudioPlaybackRt class
        + Requires low-latency driver like JACK Audio Connection Kit for
          smooth playback
    + Added real-time playback example called piano.py
    + Added real-time playback example called bebot.py
        + Which mimics the iPhone App called 'Bebot'
    + Compiles on Mac again (#55)
    + Closed #49: Buffer::getPow() undefined in python module
    + Closed #50: AudioPlayback: use double buffering (added AudioPlaybackRt)
    + Closed #55: nsound_config: update Mac version
    + Closed #60: Filters: add a parameter 'bool is_realtime' like the Generator
        class (added setRealtime(bool) instead)
    + Closed #61: Remove 'reset_first' and add member variable is_realtime for
        the Generator class (added setRealtime(bool) instead)

2013-12-24 Nsound 0.9.0

    + Added precompiled static libs of libao & libportaudio for Windows
    + Added Visual Studio compiler again for Windows builds, 64-bit builds
    + Can now build 64-bit Python module for Windows (no cygwin\mingw required)
    + Added Visual Studio 2010+ project files (using VS 2012 Express) in msvs/
    + Refactored SCons configuration a bit (nsound_config.py).
    + Added some SCons custom builds in site_scons/site_tools
    + Fixed an init problem with Moving Average Filter
    + Added Buffer.mul() for multiplying over a sub-region of the buffer.
    + Added getFrequencyOffset(f1, octaves, cents)
    + Added kwargs to most Plotter functions
    + Added Macros.h and C++ exceptions
    + Added Python module C++ exception wrapping
    + Added Python Buffer.toList() method, converts Buffer to Python List
    + Added Utils.h, computes octaves and cents away from reference frequency
    + Improved the Vocoder band-pass filter bank, vocoder example sounds much
      better
    + Refactored some code to assert values and pointers, throwing exception on
      error
    + Refactored lots of Plotter methods, now supports kwargs, added Axes class
    + Changed Plotter::subplot(), now supports sharex, sharey (see example5.cc)
    + Renamed AdsrEnvelope to EnvelopeAdsr and reworked completely to be more
      useful, calculates sustain time automatically.
    + Removed duplicate plotting code, Nsound Python module calls the smae
      C++ code (Pylab.cc)

2012-04-15 Nsound 0.8.2

    + Fixed bug 3341096 FFTransform windowing

    + Added Spectrogram class for calculating and plotting spectrograms

    + Added AudioPlayback::getStateString()

    + Added AudioPlayback::scanDevices() which will try to play a sample on
      all the available audio drivers detected

    + Added Nsound::use(AudioBackendtype)

    + Added Nsound::getBackends(), returns a list of strings of available
      audio backends (libao or libportaudio).

    + Updated AudioBackendLibao::setOptions(), now can select specific driver

    + Updated AudioBackendLibportaudio::setOptions(), now can select specific
      driver

    + Added AudioStream::pad(), ensures channels in the AudioStream are the
      same length

    + Added AudioStream::transpose(), performs a matrix transpose by treating
      the AudioStream like a 2D matrix

    + Added AudioStream::pow() & operator^ for powers

    + Added Buffer::pow() & operator^ for powers

    + Added Buffer::log(), log10()"

    + Added Buffer::findPeaks() to find indices of local peaks

    + Removed some FFTChunk functions

    + Added Generator::gausssianNoise() that draws random number from the
      normal distribution

    + Added Plotter::imagesc() to plot AudioStreams, treating them  like a 2D
      matrix, used by Spectrogram::plot()

    + Enhanced Wavefile::read() to decode the format tag, but still only reads
      uncompressed PCM wavefiles

    + Added example cepstral_pitch_estimate that calculates the cepstrum for
      pitch estimation


2011-06-25 Nsound 0.8.1

    + Added basic sound card playback via libao or libportaudio!

    + Examples now play to the sound card!

    + Made the scons build system smarter in searching for the correct python
      lib to link with

    + Enhanced the Wavefile class to handle unknown chunk ids and skip them

    + Added User's Guide documentation for Audio Playback and Instruments

2011-03-26 Nsound 0.8.0

    + Fixed Python Buffer slicing (bug 3204997)
    + Fixed build on MacOSX (bug 3186360)
    + Fixed scons error with setup.py (bug 3168738)
    + Fixed Generator::operator= (bug 3163766)
    + Closed Hat::play() demo has small artifact (bug 3148915)

    + Fixed Python AudioStream expressions like:
        a = AudioStream(100, 2)
        a << 1 << 2
        a[0] <<= 2  # Didn't work before
        a[1] <<= 3  # Didn't work before

    + Added Buffer(const FloatVector & list) constructor, allows Python
      expressions like "b = Buffer([1,2,3,4,5])"

    + Added notice to SCons build to contact Nick with bugs and help requests

    + Added lots of documentation to Generator including LaTeX formulas!

    + Added small section "Getting Started" in the User's Guide

    + Changed the signature to Generator::drawGaussian(), now specifies
        duration, mu, sigma

    + Added Buffer::exp(), std(), zNorm()

    + Added Buffer::operator^, operator^=

    + Added Buffer::getNBytes()

    + Added Generator::drawSine(), uses math, not a wavetable for precise
        waveforms with dynamic frequency.

    + Added Generator::drawSine2(), uses math, not a wavetable for precise
        waveforms with dynamic frequency and phase.

    + Added Generator::generate2(), waveforms with dynamic frequency and phase.

    + Added CudaTool.py, for compiling Cuda .cu files (EXPERIMENTAL)

    + Added StretcherCuda.cu, Cuda accelerated functions. (EXPERIMENTAL)

    + Renamed Plotter::axisX to xlim, axisY to ylim

    + Updated Doxyfile and set formula font size.

    + Changed example1 to sound like Pac Man.

2010-12-31 Nsound 0.7.5

    Minor interface changes:
    + Changed AudioStream::getMono(), now returns an AudioStream
	+ Changed Buffer.toPythonList() to Buffer.toList()

    + Added the first draft of the Nsound User's Guide!

    + Fixed a bug in Wavefile, ID3v1 tag called fclose() twice
    + Fixed the Stretcher class for time shiffting Buffers.
    + Fixed Wavefile bug, now able to read unsigned 8 bit wavefiles.

    + Added convenience functions zeros, ones, rand, for Buffers & AudioStreams
    + Added AudioStream::select(), returns AudioStreamSelection
    + Added Buffer::getResample(), dynamic case
    + Added Buffer::select(), returns BufferSelection
    + Added chorus setting to generators
    + Added ability to sync generators together like an analog synth
    + Added dynamics to FilterDelay, variable delay
    + Added Sawtooth Generator class
    + Added WindowTypes.h
    + Added Generator::drawWindow()
    + Added FilterLeastSquaresFIR::setWindow()
    + Added FFTransform::setWindow(Window)
    + Added FFTChunk::getFrequencyAxis()
    + Added Plotter::grid()
    + Added Plotter::hold()
    + Added Plotter::set_xscale(), set_yscale()
    + Added Square(sr, n_harmonics) constructor, sum of sine waves
    + Added FilterFlanger
    + Added FilterPhaser
    + Added BufferWindowSearch
    + Enhanced Buffer::getResample() to use windows of the input buffer,
      conserving memory usage.

    + Allow Filter plot data in dB to go to -60 dB

    + In the Python module, updated __getitem__() and __setitem__() functions
      to handle slice objects, returns new copy of a Buffer or AudioStream


2010-07-09 Nsound 0.7.4

    + Fixed bug 2944836 GCC 4.4.1+ errors
    + Fixed bug 2948680 NsoundConfig_Cygwin.py
    + Fixed bug 2944837 Python: FilterLowPassIIR.plot() fails
    + Fixed bug 2974228: Wavefile read
    + Fixed some byte swapping in Wavefile
    + Fixed Buffer::getResample(), now matches Matlab's resample.m

    + Added BufferResample_UnitTest()

    + Added Buffer::operator<<(AudioStream), concatenates a mono version of the
      AudioStream to the Buffer
    + Added FilterLeastSquaresFIR, creates a FIR filter with least squares fit
      to custom frequency response
    + Added Generator::drawWindowKaiser()
    + Added Stretcher class, implemented WSOLA for dynamic pitch & time
      shifting!
    + Added examples/stretcher.cc & stretcher.py that performs dynamic pitch &
      time shifting!

    + Added cool plugins for The GIMP, import/export wavefiles, high & low pass
      filters!

    + Added ID3v1 tags to the Wavefile class, this breaks the RIFF standard
      but is only used by the GIMP plugin to store image width & height.

    + Added Python operator Buffer << list, Buffer >> list, len()

    + Header include cleanup to ensure Python.h is included first
    + Some scons build system work


2009-12-07 Nsound 0.7.3

    + Added Python2.6 support on Windows (aready worked on Linux/OSX)
    + Added typedef std::vector< boolean > BooleanVector
    + Added typedef std::vector< BooleanVector > BooleanVectorVector
    + Added AudioStream::mono()
    + Added AudioStream::getMono()
    + Added class AudioStreamSelection
    + Added class BufferSelection
    + Added example7
    + Added FilterStageIIR::operator=()
    + Added Generator::drawWindowBartlett()
    + Added Generator::drawWindowBlackman()
    + Added Generator::drawWindowBlackmanHarris()
    + Added Generator::drawWindowHamming()
    + Added Generator::drawWindowHanning()
    + Added Generator::drawWindowNuttall()
    + Added Generator::drawWindowParzen()
    + Added Generator::drawWindowRectangular()
    + Added Instrument Hat
    + Added Instrument OrganPipe (A pipe organ)
    + Added TODO.txt, a list of things Nick would like to see get done
    + Added Wavefile 64 & 32 floating point support (Most apps can't open
      64-bit, but usefull for unit testing or saving high precision waveforms)
    - Removed AudioStream::setLessThan(), new operators replace functionality
    - Removed AudioStream::setGreaterThan(), new operators replace functionality
    - Removed Buffer::setLessThan(), new operators replace functionality
    - Removed Buffer::setGreaterThan(), new operators replace functionality
    + Fixed FilterLowPassFIR::filter(AudioStream, float64) infinite recursion
      error
    + Fixed FilterLowPassFIR::filter(x,y), was calling itself
    + Fixed Generator::ctor(sample_rate, Buffer) memory copy
    + Fixed NsoundConfig.py::checkForMatplotlibCAPI() bug, need to return 0;
    + Fixed swig/plotting.i to report a warning when paylab is not available

2009-09-11 Nsound 0.7.2

    + Code fixes to compile with GCC 4.3.3
    + Added AudioStream::AudioStream(sample_rate, channels, preallocate)
    + Added AudioStream::convolve() convolves the signal with any other signal
    + Added AudioStream::downSample()
    + Added AudioStream::limit(Buffer min, Buffer max)
    + Added AudioStream::limit(min, max)
    + Added AudioStream::resample()
    + Added AudioStream::setGreaterThan(threshold, value)
    + Added AudioStream::setLessThan(threshold, value)
    + Added AudioStream::sqrt()
    + Added AudioStream::upSample()
    + Added Buffer::circular_iterator operators += and -=
    + Added Buffer::convolve() convolves the signal with any other signal
    + Added Buffer::downSample()
    + Added Buffer::limit(Buffer min, Buffer max)
    + Added Buffer::limit(min, max)
    + Added Buffer::preallocate(n)
    + Added Buffer::resample()
    + Added Buffer::round()
    + Added Buffer::setGreaterThan(threshold, value)
    + Added Buffer::setLessThan(threshold, value)
    + Added Buffer::sqrt()
    + Added Buffer::upSample()
    + Added Cosine generator
    + Added DrumBD01 based on Istvan Varga's drum kits
         http://www.csounds.com/istvan/
    + Added Docs/How_To_Reverse_Engineer_Csound_Instrument.odt
    + Added FFTransform::fft(Buffer) and FFTransform::ifft(Buffer)
    + Added Filter::filter(AudioStream, float64)
    + Added Filter::filter(Buffer, float64)
    + Added FilterBandPassFIR::filter(AudioStream, float64, float64)
    + Added FilterBandPassIIR::filter(AudioStream, float64, float64)
    + Added FilterBandRejectFIR::filter(Buffer, float64, float64)
    + Added FilterBandRejectIIR::filter(Buffer, float64, float64)
    + Added FilterHighPassFIR::filter(Buffer, float64)
    + Added FilterHighPassIIR::filter(Buffer, float64)
    + Added FilterLowPassFIR::filter(AudioStream, float64)
    + Added FilterLowPassIIR::filter(AudioStream, float64)
    + Added FilterStageIIR::filter(AudioStream, float64)
    + Added Generator::buzz(), generates a set of harmonics
    + Added Generator::buzz() unit tests
    + Added Generator::reset()
    + Added Generator::setPhase()
    + Added Generator::tell()
    + Added Plotter::run_string(string), sends raw string command to Python
    + Fixed FilterTone::filter(), a bug when using dynamic frequencies
    + Fixed Generator::generate_(), changed some if's to while's
    + Fixed Plotter, added an internal state to avoid initializing the Python
         interpreter multiple times, prevents crashes

    - Removed DMALLOC stuff from NsoundConfig.py

2009-04-01 Nsound 0.7.1

    + Added Python module, you can now use Nsound in Python!
    + Added Mesh2D class, uses a mesh of delay lines to simulate 2D membrane
    + Reworked classes Envelope and AdsrEnvelope a bit, these are untested
    + Fixed a bug in Filter, reset() wasn't being called
    + Updated INSTALL with instructions for building the Python module in Linux
    + Made some small updates to the Pluck class
    + Changed default compiler flags to include optimizations

2009-02-14 Nsound 0.7.0

    + Added Mac OS X support, nsound compiles on the Mac!
    + Added option to disable 64-bit data types,     $ scons disable-64=1
    + Added option to compile a static library,      $ scons static=1
    + Added option to disable linking with python,   $ scons disable-python=1
    + Added option to enable dmalloc linking,        $ scons dmalloc=1
    + Added option to compile with lots of warnings, $ scons extra-warnings=1
    + Added a FilterMovingAverage class
    + Changed the Vocoder Class to use less memory
    + Added FilterBandPassVocoder, better performance than FilterBandPassIIR
    + Changed the raw data holder to std::vector in the Buffer class
    + Added Buffer::iterator, Buffer::circular_iterator
    + Reworked the AudioStream class, made sure to extend methods form the
      Buffer to the AudioStream class.
    + Changed the AudioStream constructor to specify sample_rate first, like
      other classes in Nsound
    + Added Buffer_UnitTest to the Unit Test dir
    + Added Plotter::text to add text at location x,y
    + Added RandomNumberGenerator interface
    + Added RngTausworthe class, based on GNU's Scientific Library (GSL)
      taus.c random number generator

2008-12-23 Nsound 0.6.1

    + Added FilterDelay, reworked from DelayLine, DelayLine will get removed
    + Added FilterCombLowPassFeedback, used for simple reverb/echo
    + Added FilterAllPass
    + Added ReverbrationRoom, simulates the echo of a room (freeverb)
    + Added Vocoder, talk like Walle!
    + Added FilterSlinky, delays signals by frequency, like traveling waves in
          a slinky
    + Added AudioStream(std::string &) constructor, to read wavefiles
    + Added Buffer(std::string &) constructor, to read wavefiles
    + Updated examples

2008-11-01 Nsound 0.6.0

    + Dropped the Microsoft compiler, use cygwin + mingw to compile on winblows
    + Added Instruments GuitarBass, FluteSlide, Clarinet, DrumKickBass
    + Wavefile class can now write 32, 48, 64 bit values
    + Generator class now can be an arbitray sample rate
    + Generator::generate() can now be used for real-time code
    + Added ParametricEqualizerFilter class based on the Csound pareq opcode
    + Added Generator::drawExp()
    + Added AudioStream::abs().
    + Renamed FIRLowPassFilter to FilterLowPassFIR
    + Renamed FIRHighPassFilter to FilterHighPassFIR
    + Renamed FIRBandPassFilter to FilterBandPassFIR
    + Renamed FIRBandRejectFilter to FilterBandRejectFIR
    + Added FilterLowPassFIR_UnitTest
    + Added FilterHighPassFIR_UnitTest
    + Added FilterBandPassFIR_UnitTest
    + Added FilterBandRejectFIR_UnitTest
    + Renamed IIRLowPassFilter to FilterLowPassIIR
    + Renamed IIRHighPassFilter to FilterHighPassIIR
    + Renamed IIRBandPassFilter to FilterBandPassIIR
    + Renamed IIRBandRejectFilter to FilterBandRejectIIR
    + Added FilterLowPassIIR_UnitTest
    + Added FilterHighPassIIR_UnitTest
    + Added FilterBandPassIIR_UnitTest
    + Added FilterBandRejectIIR_UnitTest


2008-08-09 Nsound 0.5.1

    Bugs Closed:
    + Fixed scons path link encoding, closed bug 1921838
    + Added IIRFilter::clearHistory(), closed bug 1899358

    New Features:
    + Created src/bin to hold nsound tools, e.g. ns_readwaveheader
    + Added Plotter::spectrogram(), plot frequency content of a signal
    + Added Plotter::imshow(), interpret a Buffer as a matrix and draw it
        (like Matlabs imagesc)
    + IIR plots work, easily plot filter's frequency response!
        (try src/test/test_fft)

    Other:
    + Removed Pluck::generate2(), it was low quality and cpu expensive
    + Changed filters to use pointers instead of array indexing, faster


2008-03-05 Nsound 0.5.0

    + Changed Generator::generate(duration, frequency, acceleration) to
              Generator::generate(duration, frequencies)
    + Removed Plotter::quickPlot()
    + Added AudioStream::plot()
    + Added Buffer::plot()
    + Added FFTChunk::plot()
    + Added Plotter::show(), must be called to make plots visible.
    + FIR & IIR are now derived classes from Filter.
    + Added FIR::filter(float64 x), FIR::Filter(float64 x, float64 freq1),
            FIR::filter(float64 x, float64 freq1, float64 freq2) -- FIR filters
            can do sample by sample filtering, like IIRs
    + Added Wavefile::get/setDefaultSampleRate()
    + Added Wavefile::get/setDefaultSampleSize()

2008-02-18 Nsound 0.4.4

    + Fixed a bug with Buffer::subbuffer()
    + Moved IIRFilter to IIRFilterStage
    + Created base class IIRFilter that all IIR Filters inherit
    + Added class FFTransform to perform FFTs & IFFTs
    + Added class FFTChunk to represent the result of an FFT
    + Added Plotter::axhline(), draws horizontal line
    + Added Plotter::axvline(), draws horizontal line
    + Added Plotter::xlabel(), sets x axis label
    + Added Plotter::ylabel(), sets y axis label
    + Added Plotter::axisX(), sets plot xmin, xmax
    + Added Plotter::axisY(), sets plot ymin, ymax
    + Added Plotter::quickPlot(FFTChunk)
    + Added Plotter::quickPlot(FIRFilter), plots frequency response, phase
    + Added Buffer::dB() to convert a signal to dB
    + Update Doxyfile.in
    - Matplotlib still not working for cygwin, special case for linking with
        Python that I haven't had time to figure out :(

2007-12-09 Nsound 0.4.3

    + Fixed a bug with the Plotter class when the Matplotlib's C API is not
        found
    + Build libNsound statically for profiling, when -pg is in CXXFLAGS
    + Added a few more tests to test/test_buffer.cc
    + BufferChunk now uses pointers, not an array
    + Allow the Buffer class to control the BufferChunk size
    + Fixed bugs associated with building on Big Endian systems
    - Matplotlib still not working for cygwin, special case for linking with
        Python that I haven't had time to figure out :(
    + Updated INSTALL

2007-11-20 Nsound 0.4.2

    + Integrated Python matplotlib, awesome plot package, plotting with gnuplot
        will become deprecated
    + Added test_plotter.cc in the test directory
    + Added operators to allow easy writing of wave files:
        buffer >> "somefile.wav";
        audio_stream >> "somefile.wav";
    + Added mynameis-filtered example
    + Don't do Scons configuration when cleaning the build
    + Cleaned up the examples a tiny bit

2007-09-30 Nsound 0.4.1

    + Added Pluck Generator for simple string simulation (Karplus-Strong)
    + Added Buffer::getMax(), getMaxMagnitude(), getMin()
    + Added Buffer::normalize()
    + Added AudioStream::getMax(), getMaxMagnitude(), getMin()
    + Added AudioStream::normalize()
    + Added class ToneFilter (based on Csound's tone opcode)
    + Added class DelayLine (based on Csound's delayr, delayw opcode)
    + Added class DCFilter, filters out DC offsets from signal
    + Added Wavefile::write(std::string, Buffer, bits, sample rate)
    + Added class IIRFilter, IIRLowPassFilter, IIRHighPassFilter,
        IIRBandPassFilter, IIRBandRejectFilter
    + Added example6.cc

2007-07-26 Nsound 0.4.0

    + Implemented FIR Low pass, High pass, Band Reject, Band Pass filters!!!
    + New filters can dynamically filter audio, look at:
        src/test/test_filters.cc
        src/examples/dynamic_filter.cc
    + Added class Granulator for granular synthesis!
    + Added AudioStream::getDuration(), returns length of stream in seconds
    + Added NsoundAll.h for convience
    + Added Buffer::abs()
    + Added Instrument.h
    + Added Generator::silence()
    + Added Buffer::smooth(), implements moving average filter.
    + cleaned up Buffer and AudioStream header files, only returns an object
        when necessary
    + Added Generator::drawGaussian()
    + Added Generator::drawFatGaussian()
    + Added Generator::drawDecay()

2007-04-10 Nsound 0.3.2

    + Added Cygwin build target
    + Removed old GNU Build System, SCons all the way!
    + Added DistTar.py, creates nsound-A.B.C.[tar.gz|tar.bz2]
    + Added DistZip.py, creates nsound-A.B.C.zip
    + Added AudioStream::stepAverage(uint32 n_samples_per_step); degrades the
        AudioStream
    + Added Buffer::speedUp(float32 x); resamples the stream by x
    + Added Buffer::speedUp(Buffer x); resamples the stream by x on a per
        sample basis, note that a less than 1.0 value will really slow down the
        stream, so the name is a tiny bit missleading.
    + Added AudioStream::speedUp(float32 x); calls Buffer::speedUp() accross
        all channels
    + Added AudioStream::speedUp(Buffer x); calls Buffer::speedUp() accross
        all channels
    + Added AudioStream::subStream(float32 start_time, float32 n_seconds);
    + Added Buffer::subbuffer(uint32 start_index, uint32 n_samples)

2007-02-13 Nsound 0.3.1

    + Started using subversion
    + Extensive testing implemented in the test dir for classes:
        Buffer, AudioStream and Mixer
    + Fixed Buffer::operator-(buffer_t d, Buffer b)
    + Fixed Buffer::operator/(buffer_t d, Buffer b)
    + Fixed AudioStream::pan(buffer_t d), missing '='
    + Fixed Mixer::getStream(float32 start_time, float32
    + Removed BufferChunk::alocated_; not needed
    + Changed BUFFER_CHUNK_SIZE_ to non-const to aid in testing
    + Added AudioStream::operator=(const Buffer * rhs)
    + Added AudioStream::operator+,-,*,/(const buffer_t d, const AudioStream &
        rhs)
    + Added AudioStream::subStream(uint32 offset, uint32 n_samples);

2006-10-18 Nsound 0.3.0

    + Updated all examples with new useage and cleaned them up
    + Added Buffer::reverse()
    + Added AudioStream::reverse()
    + Code cleaned up by turning on all the g++ warnings, to use it try:
        $ scons warn=g++
    + Cleaned up and implemented operator=() and copy constructor for all
        classes
    + Added a test to SConstruct to test for the M_PI macro
    + Removed M_PI macro declaration in .cpp files
    + Class Envelope now uses an arbartary number of channels.
    + Removed any class Allocated() crap, don't need it.
    + Removed AudioStream number of channels limitation, now any number of
        channels may be stored
    + Added AudioStream::operator[+-*/](Buffer)
    + AudioStream's channel indicies now start at 0 like it should
    + Renamed AudioStream::write() to AudioStream::setChannel()
    + Renamed AudioStream::getNumberOfChannels() to AudioStream::getNChannels()
    + Added AudioStream::setNumberOfChannels(int)
    + Added Buffer Wavefile::operator<<(Buffer, std::string)
        e.g. Buffer b1 << "some_file.wav";
    + Added AudioStream Wavefile::operator<<(AudioStream, std::string)
        e.g. AudioStream as1 << "some_file.wav";
    + Rewrote Wavefile to read/write an arbartrary number of channels
        to/from a file.


2006-08-01 Nsound 0.2.2

    + Fixed and tested gnuplot usage on windows
    + Reworked scons build system a bit to check for endianess and sizeof
    + Added ACGenerateFile.py, an scons replacement for AC_SUBST()
    + Added Generator::drawParabola() for exponential curves
    + Added Generator::tanh() that will draw the tanh funciton
    + Added Generator::whiteNoise()
    + Added class LowPassFilter
    + Added example5 that simulates a drum
    + Added test_lfp that uses LowPassFilter
    + Added class ComplexNumber, not tested
    + Replaced tabs with spaces

    These don't work yet, I'm doing research:
    - Started class CascadeFilter
    - Started class LowPassFilterIIR

2006-05-01 Nsound 0.2.1

    + Added Scons (www.scons.org) build system, works on linux and windows
    + Nsound now builds natively on windows; CAUTION: windows development is
        hazardous to your health!
    + Removed a "using namespace" from Wavefile.h, shame on me!
    + Added MSVC++ project files under nsound-0.2.1/msw in hopes that it may
        be useful

2006-03-22 Nsound 0.2.0

    + Reorganized directory structure.
    + Updated configure.ac to capture CFLAGS, CPPFLAGS, and CXXFLAGS correctly
    + Put everything in one namespace called Nsound.
    + Removed the NSOUND_DELETE macro from Nsound.h.
    + Moved the SWAP_BYTES macro from Nsound.h to private members of the
        Wavefile class.
    + Removed unused AudioStream.typeMatches().
    + Added AudioStream.pan() method.
    + Buffer class math operations now don't require buffers of same length,
        only operates over the shorter buffer.
    + Added -=, +=, *=, and /= operators to Buffers and AudioStreams.
    + Changed the Generator class:
        The Generator class now holds a generic waveform in a wavetable.  The
        Sine, Cosine, Triangle and Square all fill the wavetable with their
        waveform so all classes get the generate() method for free.
    + Added Triangle, Square derived from Generator.
    + Removed the Line class and implemented a drawLine method for class
        Generator.  drawLine() method takes (duration,amplitude_start,
        amplitude_finish).
    + Added a Envelope base class.
    + Reworked AdsrEnvelope to be a derived class from Envelope.
    + Updated Buffer::add(), works for all cases.
    + Updated test/test_buffer.cc to brute force test the Buffer::add() method.
    + Added examples/example4.cc to show a sine wave turn into a triange to a
        square wave.
    + Added a check in Wavefile::write() to avoid clipping.

2005-10-26 Nsound 0.1.2

    + Found and fixed bug in Buffer::operator<<(Buffer)
    + Replaced legacy NsoundSystem::LinkList class with std::vector
    + Added example3.cc

2005-08 Nick Hilton <weegreenblobbie2_gmail_com>

    * Nsound-0.1.0 released under GPL

-----

Copyright (C) 2005  Nick Hilton

This file is part of Nsound.

Nsound is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2, or (at your option)
any later version.

Nsound is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with Nsound; see the file COPYING.  If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.