Mercurial > hg > qm-dsp
changeset 73:dcb555b90924
* Key detector: when returning key strengths, use the peak value of the
three underlying chromagram correlations (from 36-bin chromagram)
corresponding to each key, instead of the mean.
Rationale: This is the same method as used when returning the key value,
and it's nice to have the same results in both returned value and plot.
The peak performed better than the sum with a simple test set of triads,
so it seems reasonable to change the plot to match the key output rather
than the other way around.
* FFT: kiss_fftr returns only the non-conjugate bins, synthesise the rest
rather than leaving them (perhaps dangerously) undefined. Fixes an
uninitialised data error in chromagram that could cause garbage results
from key detector.
* Constant Q: remove precalculated values again, I reckon they're not
proving such a good tradeoff.
| author | cannam |
|---|---|
| date | Fri, 05 Jun 2009 15:12:39 +0000 |
| parents | d0b35b1e3a98 |
| children | 769da847732b |
| files | dsp/chromagram/CQprecalc.cpp dsp/chromagram/Chromagram.cpp dsp/chromagram/ConstantQ.cpp dsp/keydetection/GetKeyMode.cpp dsp/transforms/FFT.cpp qm-dsp.pro |
| diffstat | 6 files changed, 63 insertions(+), 16 deletions(-) [+] |
line wrap: on
line diff
--- a/dsp/chromagram/CQprecalc.cpp Tue Jun 02 11:25:58 2009 +0000 +++ b/dsp/chromagram/CQprecalc.cpp Fri Jun 05 15:12:39 2009 +0000 @@ -1,3 +1,10 @@ + +#ifdef NOT_DEFINED + +// Well, I now think this doesn't save enough time to justify the +// extra bulk in the plugin executable. But keep it around and only +// ifdef it out, just in case... + #include <vector> using std::vector; static unsigned int sk_i_76_16384[5818] = { @@ -49838,3 +49845,6 @@ imag.push_back(sk_imag_84_65536[i]); } } + +#endif +
--- a/dsp/chromagram/Chromagram.cpp Tue Jun 02 11:25:58 2009 +0000 +++ b/dsp/chromagram/Chromagram.cpp Fri Jun 05 15:12:39 2009 +0000 @@ -135,7 +135,7 @@ m_window->cut(m_windowbuf); // FFT of current frame - m_FFT->process(0, m_windowbuf, m_FFTRe, m_FFTIm); + m_FFT->process(false, m_windowbuf, m_FFTRe, m_FFTIm); return process(m_FFTRe, m_FFTIm); }
--- a/dsp/chromagram/ConstantQ.cpp Tue Jun 02 11:25:58 2009 +0000 +++ b/dsp/chromagram/ConstantQ.cpp Fri Jun 05 15:12:39 2009 +0000 @@ -12,6 +12,9 @@ #include <iostream> +#ifdef NOT_DEFINED +// see note in CQprecalc + #include "CQprecalc.cpp" static bool push_precalculated(int uk, int fftlength, @@ -38,6 +41,7 @@ } return false; } +#endif //--------------------------------------------------------------------------- // nextpow2 returns the smallest integer n such that 2^n >= x. @@ -70,12 +74,15 @@ SparseKernel *sk = new SparseKernel(); +#ifdef NOT_DEFINED if (push_precalculated(m_uK, m_FFTLength, sk->is, sk->js, sk->real, sk->imag)) { +// std::cerr << "using precalculated kernel" << std::endl; m_sparseKernel = sk; return; } - +#endif + //generates spectral kernel matrix (upside down?) // initialise temporal kernel with zeros, twice length to deal w. complex numbers
--- a/dsp/keydetection/GetKeyMode.cpp Tue Jun 02 11:25:58 2009 +0000 +++ b/dsp/keydetection/GetKeyMode.cpp Fri Jun 05 15:12:39 2009 +0000 @@ -174,13 +174,13 @@ // Added 21.11.07 by Chris Sutton based on debugging with Katy // Noland + comparison with Matlab equivalent. MathUtilities::circShift( m_ChrPointer, m_BPO, 1); - /* - std::cout << "raw chroma: "; - for (int ii = 0; ii < m_BPO; ++ii) { - std::cout << m_ChrPointer[ii] << " "; - } - std::cout << std::endl; + std::cout << "raw chroma: "; + for (int ii = 0; ii < m_BPO; ++ii) { + if (ii % (m_BPO/12) == 0) std::cout << "\n"; + std::cout << m_ChrPointer[ii] << " "; + } + std::cout << std::endl; */ // populate hpcp values; int cbidx; @@ -232,13 +232,27 @@ for( k = 0; k < m_BPO*2; k++ ) { - m_keyStrengths[k/(m_BPO/12)] += m_Keys[k]; + int idx = k / (m_BPO/12); + int rem = k % (m_BPO/12); + if (rem == 0 || m_Keys[k] > m_keyStrengths[idx]) { + m_keyStrengths[idx] = m_Keys[k]; + } + +// m_keyStrengths[k/(m_BPO/12)] += m_Keys[k]; } /* std::cout << "raw keys: "; for (int ii = 0; ii < 2*m_BPO; ++ii) { - std::cout << m_Keys[ii] << " "; + if (ii % (m_BPO/12) == 0) std::cout << "\n"; + std::cout << m_Keys[ii] << " "; + } + std::cout << std::endl; + + std::cout << "key strengths: "; + for (int ii = 0; ii < 24; ++ii) { + if (ii % 6 == 0) std::cout << "\n"; + std::cout << m_keyStrengths[ii] << " "; } std::cout << std::endl; */ @@ -246,7 +260,7 @@ // '1 +' because we number keys 1-24, not 0-23. key = 1 + (int)ceil( (double)MathUtilities::getMax( m_Keys, 2* m_BPO, &dummy )/3 ); -// std::cout << "key pre-sorting: " << key << std::endl; +// std::cout << "key pre-sorting: " << key << std::endl; //Median filtering @@ -285,11 +299,15 @@ int sortlength = m_MedianBufferFilling; int midpoint = (int)ceil((double)sortlength/2); +// std::cout << "midpoint = " << midpoint << endl; + if( midpoint <= 0 ) midpoint = 1; key = m_SortedBuffer[midpoint-1]; +// std::cout << "returning key = " << key << endl; + return key; }
--- a/dsp/transforms/FFT.cpp Tue Jun 02 11:25:58 2009 +0000 +++ b/dsp/transforms/FFT.cpp Fri Jun 05 15:12:39 2009 +0000 @@ -278,12 +278,23 @@ KissFFTRealRec *rec = (KissFFTRealRec *)m_private; if (inverse) { kiss_fftr(rec->inverse, rin, rec->out); + for (int i = 0; i < m_n; ++i) { + rout[i] = rec->out[i].r; + iout[i] = rec->out[i].i; + } } else { kiss_fftr(rec->forward, rin, rec->out); - } - for (int i = 0; i < m_n; ++i) { - rout[i] = rec->out[i].r; - iout[i] = rec->out[i].i; + rout[0] = rec->out[0].r; + iout[0] = rec->out[0].i; + for (int i = 1; i < m_n/2; ++i) { + rout[m_n-i] = rout[i] = rec->out[i].r; + } + for (int i = 1; i < m_n/2; ++i) { + iout[i] = rec->out[i].i; + iout[m_n-i] = -iout[i]; + } + rout[m_n/2] = rec->out[m_n/2].r; + iout[m_n/2] = rec->out[m_n/2].i; } }