| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2013-06-19 | Initialise ooura struct with right amount of memory for ooura FFT. Fixes #11 | Jamie Bullock | |
| 2013-06-19 | Print which FFT implementation is being used at startup | Jamie Bullock | |
| 2013-06-18 | Copy out input values for Ooura so the in-place transform doesn't overwrite ↵ | Jamie Bullock | |
| *data | |||
| 2013-06-18 | Update definition with const to reflect prototype | Jamie Bullock | |
| 2013-06-18 | Comment out unused variable | Jamie Bullock | |
| 2013-06-18 | Update function prototype to use const for passed in sample buffer | Jamie Bullock | |
| 2013-06-02 | Pass block size to xtract_init_vdsp_data() instead of block size / 2. Fixes ↵ | Jamie Bullock | |
| bug with erroneous values in second half of spectrum when using vDSP FFT | |||
| 2013-05-31 | Reinstate path fix | Jamie Bullock | |
| 2013-05-31 | Remove obsolete reference to xtract_scalar_private.h | Jamie Bullock | |
| 2013-05-31 | Add wavelet-based pitch tracker | Jamie Bullock | |
| 2013-04-24 | Provide relative include path for libxtract.h for compatibility with ↵ | Jamie Bullock | |
| non-autotools builds | |||
| 2013-01-10 | fixed Linux bugs | Jamie Bullock | |
| 2013-01-09 | implemented optimised FFT via the Accelerate framework. closes #5 | Jamie Bullock | |
| 2013-01-09 | add option to explicitly specify Ooura library at configure time | Jamie Bullock | |
| 2013-01-09 | fixed bug in xtract_dct where data wasn't being copied to result vector | Jamie Bullock | |
| 2013-01-09 | run FFT in place on the data vector | Jamie Bullock | |
| 2013-01-09 | switched from single to double precision througout. closes #9 | Jamie Bullock | |
| 2013-01-07 | removed references to FFTW | Jamie Bullock | |
| 2013-01-07 | changed license from GPL to MIT. closes #6 | Jamie Bullock | |
| 2013-01-07 | added Ooura implementation to repository | Jamie Bullock | |
| 2012-03-29 | - fixed build-time bug. in some cases average_deviation was commented out ↵ | Jamie Bullock | |
| instead of spectral_average_deviation | |||
| 2011-03-31 | - fixed xtract_spread() -- now just calls xtract_spectral_variance() | Jamie Bullock | |
| 2011-03-30 | - reimplemented xtract_spectral_variance() xtract_spectral_skewness() and ↵ | Jamie Bullock | |
| xtract_spectral_kurtosis() using correct maths - removed xtract_spectral_average_deviation() | |||
| 2011-03-30 | - fixed typos in *result initialisation potentially fixing horrible bug | Jamie Bullock | |
| 2010-02-03 | - fixed DC/Nyquist inclusion bug in xtract_spectrum() and refactored a bit | Jamie Bullock | |
| 2008-11-11 | - Committed patch by Chris Cannam to avoid reading past the end of an array ↵ | Jamie Bullock | |
| in peak detection routine in xtract_peak_spectrum() | |||
| 2008-09-12 | - Committing ancient change left on HD. Relates to FFTW fft type | Jamie Bullock | |
| 2008-05-04 | Fixed bug in peak interpolation algorithm in xtract_peak_spectrum() | Jamie Bullock | |
| 2008-03-26 | - Fixed build fail if --enable-fft not specified | Jamie Bullock | |
| - Fixed doxygen build so that it includes libxtract.h - Doxygen tweaks | |||
| 2008-02-16 | - Added to pd example the ability to differentiate between different | Jamie Bullock | |
| argv types (XTRACT_FLOAT, XTRACT_INT) and pass the correct data type to the xtract[]() function - Added xtract_flatness_db() details to descriptors.c - Fixes to tonality and xtract_subbands descriptors - Added Pd examples for 'subband mean' and tonality calculated using subbands | |||
| 2008-02-15 | Added new extraction function: xtract_subbands() | Jamie Bullock | |
| 2008-02-15 | - Fixed bugs in xtract_flatness(), or at least added necessary | Jamie Bullock | |
| documentation and error checking to avoid problems - Added xtract_is_denormal() helper function and XTRACT_DENORMAL_FOUND return code - Replaced all instances of log, sqrt, exp etc. with respective floating point counterparts (logf etc.) - Added check for architecture endianness to configure script - Bug fix to PD example, now no longer crashes if no arguments are given - Minor documentation updates | |||
| 2008-01-02 | - Minor fix in xtract_lpc... or at least think it's a fix | Jamie Bullock | |
| 2008-01-02 | - Re-factoring in xtract_spectrum and fixed normalisation bug | Jamie Bullock | |
| - Fixed bug in xtract_lnorm | |||
| 2008-01-01 | - Improvements to SWIG bindings generation script | Jamie Bullock | |
| - Fixed omission in xtract_bark_coefficients that was causing the output to be complete b/s! This fixed bark_coeffs and loudness feature which depends on it - Changes to descriptor API: added is_delta and id. id corresponds to value in xtract_features_ enum and is useful for programmatic conversions between id and name string. | |||
| 2007-12-29 | - PD example brought in line with new delta features and subframe function | Jamie Bullock | |
| - subframe-test.pd added - fix to a_blockswap~.pd | |||
| 2007-12-28 | - Added new helper functions: xtract_windowed() and ↵ | Jamie Bullock | |
| xtract_features_from_subframes() - Added windowing functions (window.c) | |||
| 2007-12-27 | Added l-norm/spectral flux | Jamie Bullock | |
| 2007-12-27 | - Added extra argument to xtract_spectrum to give the option of normalising ↵ | Jamie Bullock | |
| the magnitude/power coeffificients - Removed duplicate code block (argc assignment) from descriptors.c - Added some extra documentation to libxtract.h | |||
| 2007-12-24 | - Fixes to descriptors.c where no break statement was given for certain ↵ | Jamie Bullock | |
| cases is switch conditionals - Added LPC and LPCC extraction functions. LPC implements Durbin method as described in Rabiner and Juang and implemented in Dr. Dobbs 1994 edition by Jutta Degener | |||
| 2007-10-16 | Fix for 'multiple symbol definitions' vs 'symbol not defined' dilemna. I ↵ | Jamie Bullock | |
| think the solution is to wrap the globals in a struct, declare it in a header, and wrap with a definition guard, then define _once_ at library init time. (Sounds like a recipe for something...) | |||
| 2007-10-09 | Added missing extern storage class specifier to private globals | Jamie Bullock | |
| 2007-10-09 | Fixed bug in xtract_init_mfcc() - iteration to decide filter peak positions ↵ | Dan Stowell | |
| only filled in freq_bands+1 values, when it should fill in freq_bands+2 | |||
| 2007-10-08 | Committed missing files needed for new out-of-place fftw plan creation | Jamie Bullock | |
| 2007-10-06 | Removed fftw_plan from xtraction functions. Created new init function ↵ | Jamie Bullock | |
| xtract_init_fft() for creating plans, which have global scope. Updated examples to reflect the change. New configure option: --with-fft_optimisation (0 = FFTW_ESTIMATE, 1 = FFTW_MEASURE, 2 = FFTW_PATIENT). | |||
| 2007-10-03 | Dan Stowell: Removed strange "zeroing" part of xtract_mfcc() which was ↵ | Dan Stowell | |
| zeroing a load of elements despite the fact that they're ignored by the DCT process called next, and never used for anything. This was writing to an assumed large result array (same size as number of FFT bins) despite the fact that only a small number of MFCCs (typically less than 50) are required, therefore either wasting memory or writing to memory it shouldn't do! | |||
| 2007-10-03 | Minor modifications to MFCC init, mainly just code formatting | Dan Stowell | |
| 2007-09-24 | Made xtract_loudness() more permissive, so it doesn't crash if N > BARK_BANDS | Jamie Bullock | |
| 2007-09-06 | Java bindings (with simple test) now working. Unified Python and Java ↵ | Jamie Bullock | |
| bindings tests. | |||
| 2007-09-05 | Added java bindings. Everything seems to work OK, except test.java won't ↵ | Jamie Bullock | |
| compile. Some kind of path error, but I can't work out how to fix it. | |||