130 lines
3.1 KiB
Plaintext
130 lines
3.1 KiB
Plaintext
|
|
2.0
|
|
|
|
Add new analyzer methods bandpass_band_ff() and band_q().
|
|
|
|
Add new analyzer methods band_analysis_support() and
|
|
band_synthesis_support().
|
|
|
|
Add support for adjusting the amount of overlap between bands.
|
|
|
|
Add support for selecting a local or global phase convention, and make
|
|
the local phase convention the default in the 2.0 API.
|
|
|
|
Extend the supported ff_min range to 0.0001...0.5.
|
|
|
|
Add support for linear and mel frequency scales.
|
|
|
|
Analyzers no longer share data via static variables as this could
|
|
cause race conditions in multithreaded applications.
|
|
|
|
Disable the copy constructor and assignment operator of the analyzer
|
|
class.
|
|
|
|
The number of bands per octave is no longer restricted to integer
|
|
values.
|
|
|
|
Further improve the performance of analyzing short signal blocks.
|
|
|
|
Document bands_begin() and bands_end(), and expand the documentation
|
|
of band numbering in general.
|
|
|
|
1.7
|
|
|
|
Miscellaneous bug fixes.
|
|
|
|
Support lower numbers of bands per octave, down to 4.
|
|
|
|
Further improve the performance of analyzing short signal blocks.
|
|
|
|
The "Frequency-Domain Filtering" and "Streaming" examples now use
|
|
a white noise and impulse signal, respectively.
|
|
|
|
1.6
|
|
|
|
Add "API Introduction" documentation section that was missing
|
|
from version 1.5, causing broken links.
|
|
|
|
Improve analysis and resynthesis performance when using PFFFT or vDSP
|
|
by automatically enabling the use of real rather than complex FFTs
|
|
where applicable.
|
|
|
|
1.5
|
|
|
|
Add navigation links to the HTML documentation.
|
|
|
|
Add a code example demonstrating synthesis of musical notes.
|
|
|
|
Add a function process() for iterating over coefficients sets with
|
|
greater flexibility than apply(). Also add a function fill() for
|
|
algorithmically creating new coefficients.
|
|
|
|
Make the C++ declarations in the API reference documents more closely
|
|
resemble actual C++ code.
|
|
|
|
Add a method gaborator::analyzer::band_ref() returning the band number
|
|
corresponding to the reference frequency.
|
|
|
|
1.4
|
|
|
|
Support building the library as C++17, while retaining compatibility
|
|
with C++11.
|
|
|
|
Further improve the performance of analyzing short signal blocks, and
|
|
of signal blocks not aligned to large powers of two.
|
|
|
|
Add a code example mesasuring the resynthesis signal-to-noise
|
|
ratio (SNR).
|
|
|
|
1.3
|
|
|
|
Eliminate some compiler warnings.
|
|
|
|
Declare gaborator::analyzer::band_ff() const, making the code match
|
|
the documentation.
|
|
|
|
Fix incorrect return type of gaborator::analyzer::band_ff() in the
|
|
documentation.
|
|
|
|
Improve performance of analyzing short signal blocks.
|
|
|
|
Remove special-case optimization of analyzing signal slices of all
|
|
zeros, as it caused incorrect results.
|
|
|
|
Support up to 384 bands per octave.
|
|
|
|
1.2
|
|
|
|
Add overview documentation.
|
|
|
|
Add real-time FAQ.
|
|
|
|
Actually include version.h in the release.
|
|
|
|
Fix off-by-one error in defintion of analyzer constructor ff_min
|
|
argument.
|
|
|
|
Fix incorrect return value of band_ff() for DC band.
|
|
|
|
Add streaming example code.
|
|
|
|
Add analyzer::analysis_support() and analyzer::synthesis_support().
|
|
|
|
Document analyzer::band_ff().
|
|
|
|
Improve signal to noise ratio at low numbers of bands per octave.
|
|
|
|
Note the need for -mfpu=neon on ARM in render.html.
|
|
|
|
1.1
|
|
|
|
Added CHANGES file.
|
|
|
|
Added reference documentation.
|
|
|
|
New include file gaborator/version.h.
|
|
|
|
1.0
|
|
|
|
Initial release
|