Hi everyone there!
First of all I wanna' say thank you all for your awesome work!
Said that, I was wondering if somebody could help me finding where in McVicar's dissertation I can find the filterbank implementation referenced in filters.constant_q().
Here is a copy of comment in the correspodent code:
_def constant_q(sr, fmin=None, n_bins=84, bins_per_octave=12, tuning=0.0,
window='hann', filter_scale=1, pad_fft=True, norm=1,
**kwargs):
r'''Construct a constant-Q basis.
This uses the filter bank described by [1]_.
.. [1] McVicar, Matthew.
"A machine learning approach to automatic chord extraction."
Dissertation, University of Bristol. 2013._
Thank you guys!
I guess I'm best placed to answer this one...
The paper you are referencing uses the chromagram features from my
dissertation - details can be found in Section 3.5 here:
http://www.mattmcvicar.com/wp-content/uploads/2014/02/thesis.pdf
The original implementation was in MATLAB and can be found here:
https://www.dropbox.com/s/tk5bn2azqp4a734/Harmony_Progression_Analyzer_Toolbox_ver1.2.zip?dl=0
On Wed, 2 Aug 2017 at 21:05, jbraga@github notifications@github.com wrote:
Hi everyone there!
First of all I wanna' say thank you all for your awesome work!
Said that, I was wondering if somebody could help me finding where in
McVicar's dissertation I can find the filterbank implementation referenced
in filters.constant_q.Here is a copy of comment in the correspodent code:
def constant_q(sr, fmin=None, n_bins=84, bins_per_octave=12, tuning=0.0,
window='hann', filter_scale=1, pad_fft=True, norm=1,
**kwargs):
r'''Construct a constant-Q basis.
This uses the filter bank described by [1]_.
.. [1] McVicar, Matthew.
"A machine learning approach to automatic chord extraction."
Dissertation, University of Bristol. 2013.Thank you guys!
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
https://github.com/librosa/librosa/issues/607, or mute the thread
https://github.com/notifications/unsubscribe-auth/ACfGsukSbuBtCNBDF4jvQN4aJtRm8idiks5sUNaDgaJpZM4Orloc
.>
Matt McVicar
Machine Learning Researcher
Jukedeck
contact: [email protected]
blog/website: www.mattmcvicar.com
A bit more context here: I'm not sure the current implementation bears much resemblance to Matt's, since it's been revised heavily between the 0.3.x and 0.5.x series. A more accurate ref might be
Schörkhuber, Christian, and Anssi Klapuri. "Constant-Q transform toolbox for music processing." 7th Sound and Music Computing Conference, Barcelona, Spain. 2010.
But note that we did not use that as a reference implementation, merely reimplemented the basic idea of the method.
Thanks guys for your response!