alpha-0.3

Update alpha-0.3

• Added exploration_of_timbre directory with various experimental ML-based sound generation approaches using various global optimization algorithms. The primary goal is to synthesize a DFT frame within a restricted FM/AM setup, optionally aimed to be used within Ableton/Operator.

• spectral_fm3.ipynb - Notebook for FM-based sound approximation using a single DFT frame as the source

• spectral_am3.ipynb - Notebook for AM-based sound approximation using a single DFT frame as the source

• operator_fm.ipynb - Notebook for adjusting the calculated FM values to Ableton's Operator preset format

• operator_am.ipynb - Notebook for adjusting the calculated AM values to Ableton's Operator preset format

• operator_preset_editor_fm.ipynb - Notebook for extracting and saving the Operator presets in .adv format (native format for Ableton's Operator)

• optimization_gif.ipynb - Notebook for creating learning process visualizations for different optimization algorithms

• distances_demo.ipynb - Notebook for visualizing distances of different objective functions

Full Changelog: alpha-0.2...alpha-0.3

alpha-0.2

Update alpha-0.2

Refactoring

• All code has been refactored.
• All Python scripts are now moved to the py_scripts folder.

Tutorials Added

All tutorials can be found in the tutorials folder:

• mfcc_bank.ipynb - Brief introduction to MFCC-based sound representations.
• peaks_scipy_showcase.ipynb - Quick introduction to find_peaks function from scipy.signal used for DFT-based peak filtering.
• showcase_bayle.ipynb, showcase_noanoa.ipynb, showcase_parm.ipynb - Various examples of DFT-based peak detection and resynthesis aimed to extract symbolic representations from various sounds and resynthesize the DFT frames for aural judgement and exploration of analyzed sounds.
• above_nyquist.ipynb - Brief introduction to the effects of aliasing.
• dft_resolution.ipynb - Brief introduction to the effects of sampling rate and sample length.

Notebooks Added/Revised

• symbolic_mei.ipynb - Completely rewritten implementation of Verovio-based MEI rendering (check mei.py in py_scripts folder as well). Now supports various modes of rendering, including MIDI cent deviation notation above the note (useful for microtonal analysis).
• ssm.ipynb - Plotly-based self-similarity matrix visualization of selected audio files, includes 'chroma', 'mfcc', or 'chroma+mfcc' analysis methods.