Mercurial > hg > nnls-chroma
changeset 66:611a3cfab5fc matthiasm-plugin
added Sonic Visualiser screenshot
| author | Matthias Mauch <mail@matthiasmauch.net> |
|---|---|
| date | Tue, 26 Oct 2010 20:37:54 +0900 |
| parents | f3afde2a8ced |
| children | 38d2be4413a1 |
| files | README sv-screenshot-nnls-chroma.png |
| diffstat | 2 files changed, 16 insertions(+), 16 deletions(-) [+] |
line wrap: on
line diff
--- a/README Tue Oct 26 11:11:04 2010 +0200 +++ b/README Tue Oct 26 20:37:54 2010 +0900 @@ -1,9 +1,9 @@ -## NNLS Chroma ## +### NNLS Chroma ### System identifier – vamp:nnls-chroma:nnls-chroma -RDF URI – http://vamp-plugins.org/rdf/plugins/nnls-chroma#nnls-chroma (not yet available) +RDF URI – http://vamp-plugins.org/rdf/plugins/nnls-chroma#nnls-chroma -### General Description ### +#### General Description #### NNLS Chroma analyses a single channel of audio using frame-wise spectral input from the Vamp host. The plugin was originally developed to extract treble and bass chromagrams for subsequent use in chord extraction methods. The spectrum is transformed to a log-frequency spectrum (constant-Q) with three bins per semitone. On this representation, two processing steps are performed: * tuning, after which each centre bin (i.e. bin 2, 5, 8, ...) corresponds to a semitone, even if the tuning of the piece deviates from 440 Hz standard pitch. @@ -11,7 +11,7 @@ The processed log-frequency spectrum is then used as an input for NNLS approximate transcription (using a dictionary of harmonic notes with geometrically decaying harmonics magnitudes). The output of the NNLS approximate transcription is semitone-spaced. To get the chroma, this semitone spectrum is multiplied (element-wise) with the desired profile (chroma or bass chroma) and then mapped to 12 bins. The resulting chroma frames can be normalised by (dividing by) their norm (L1, L2 and maximum norm available). -### Parameters ### +#### Parameters #### The default settings (in brackets, below) are those used for Matthias Mauch's 2010 MIREX submissions. @@ -22,7 +22,7 @@ * spectral shape (0.5 -- 0.9; default: 0.7): the shape of the notes in the NNLS dictionary. Their harmonic amplitude follows a geometrically decreasing pattern, in which the i-th harmonic has an amplitude of [spectral shape]^[i-1], where "^" means "to the power of". * chroma normalisation (none, maximum norm, L1 norm, L2 norm; default: none): determines whether or how the chromagrams are normalised. If the setting is not 'none', then each chroma frame separately is divided by the chosen vector norm. Note that normalisation implies that the joint 24-dim. "Chroma and Bass Chromagram" output will be different from the individual 12-dim. "Chromagram" and "Bass Chromagram" outputs. -### Outputs ### +#### Outputs #### * Log-frequency Spectrum: a spectrum similar to the well-known constant Q spectrum, in which bins are linear in log-frequency. Three bins per semitone. * Tuned Log-frequency Spectrum: has the same format as Log-frequency Spectrum, but has been processed by the following processes: tuning, subtraction of background spectrum, spectral whitening. @@ -31,16 +31,16 @@ * Chromagram: a 12-dimensional chromagram, restricted with mid-range emphasis. At each frame the Semitone Spectrum is multiplied by a mid-range pattern and then mapped to the 12 chroma bins. * Chromagram and Bass Chromagram: a 24-dimensional chromagram, consisting of the both Bass Chromgram and Chromagram, see above. When normalisation is used, this representation will however be scaled differently, and hence be different from the individual chromagrams. -## Chordino ## +### Chordino ### System identifier – vamp:nnls-chroma:chordino -RDF URI – http://vamp-plugins.org/rdf/plugins/nnls-chroma#chordino (not yet available) +RDF URI – http://vamp-plugins.org/rdf/plugins/nnls-chroma#chordino -### General Description ### +#### General Description #### Chordino provides a simple chord transcription based on NNLS Chroma (described above). Chord profiles given by the user in the file "chord.dict" are used to calculate frame-wise chord similarities. Two simple (non-state-of-the-art!) algorithms are available that smooth these to provide a chord transcription: a simple chord change method, and a standard HMM/Viterbi approach. -### Parameters ### +#### Parameters #### The default settings (in brackets, below) are those used for Matthias Mauch's 2010 MIREX submissions. @@ -52,30 +52,30 @@ * spectral shape (0.5 -- 0.9; default: 0.7): the shape of the notes in the NNLS dictionary. Their harmonic amplitude follows a geometrically decreasing pattern, in which the i-th harmonic has an amplitude of [spectral shape]^[i-1], where "^" means "to the power of". * chroma normalisation (none, maximum norm, L1 norm, L2 norm; default: none): determines whether or how the chromagrams are normalised. If the setting is not 'none', then each chroma frame separately is divided by the chosen vector norm. Note that normalisation implies that the joint 24-dim. "Chroma and Bass Chromagram" output will be different from the individual 12-dim. "Chromagram" and "Bass Chromagram" outputs. -### Outputs ### +#### Outputs #### * Chord Estimate: estimated chord times and labels. * Harmonic Change Value: an indication of the likelihood of harmonic change. Depends on the chord dictionary. Calculation is different depending on whether the Viterbi algorithm is used for chord estimation, or the simple chord estimate. -## Tuning ## +### Tuning ### System identifier – vamp:nnls-chroma:tuning -RDF URI – http://vamp-plugins.org/rdf/plugins/nnls-chroma#tuning (not yet available) +RDF URI – http://vamp-plugins.org/rdf/plugins/nnls-chroma#tuning -### General Description ### +#### General Description #### The tuning plugin can estimate the local and global tuning of piece. The same tuning method is used for the NNLS Chroma and Chordino plugins. -### Parameter ### +#### Parameter #### * spectral roll on spectral roll on (0 % -- 5 %; default: 0 %): consider the cumulative energy spectrum (from low to high frequencies). All bins below the first bin whose cumulative energy exceeds the quantile [spectral roll on] x [total energy] will be set to 0. A value of 0 means that no bins will be changed. -### Outputs ### +#### Outputs #### * Tuning: returns a single label (at time 0 seconds) containing an estimate of the concert pitch in Hz. * Local Tuning: returns a tuning estimate at every analysis frame, an average of the (recent) previous frame-wise estimates of the concert pitch in Hz. -## References and Credits ## +### References and Credits ### Mauch, Matthias and Dixon, Simon: [*Approximate Note Transcription for the Improved Identification of Difficult Chords*](http://schall-und-mauch.de/artificialmusicality/?p=89), Proceedings of the 11th International Society for Music Information Retrieval Conference (ISMIR 2010), 2010.