| Age | Commit message (Collapse) | Author | 
|---|
|  | Group classes by type (peak detection, partial
tracking, etc) rather than by algorithm name. This
is because Cython needs the full inheritance
hierarchy to be in the same file. | 
|  | PEP8 cleanup of test_sms.py. | 
|  |  | 
|  |  | 
|  |  | 
|  | PeakDetection.find_peaks | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  | Move main license file to project root. | 
|  |  | 
|  |  | 
|  |  | 
|  | Ignore valgrind errors due to a memory leak when
throwing a std::exception | 
|  |  | 
|  | arrays instead of having their own std::vector
of samples | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  | Currently only builds the simplsndobj module | 
|  | beginning of analysis which was causing synthesised audio to be out of sync with the original. | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  | peaks, and a pointer assignment in find_peaks could lead to a double free in freeAnalysis/freeSpectralPeaks. | 
|  |  | 
|  | The new libsms memory management code in v1.15 fixes this issue | 
|  | power of 2 would stop execution. Fixed another SMS test. | 
|  |  | 
|  |  | 
|  |  | 
|  | improvements/simplifications but the basic API is the same. Also started updating the unit tests, which will now use the nose framework | 
|  |  | 
|  |  | 
|  |  | 
|  |  | 
|  | bit easier. Fixed a bug that would cause SndObjSynthesis to crash if peak values were floats. | 
|  |  |