PyHarmonize: Adding harmony lines to recorded melodies in Python

Last update: May 20, 2022

Related tags

Machine Learning PyHarmonize

Overview

PyHarmonize: Adding harmony lines to recorded melodies in Python

About

To use this module, the user provides

a wav file containing a melody,
the key in which the melody is, and
the scale degree(s) of the desired harmony.

The module then outputs a wav file which contains the original melody, together with the added harmony line(s).

We first give some examples, the installation instructions are further below.

Examples (with audio files)

We here provide three audio examples together with the code used to generate them. See the folder examples/ for more detailed example notebooks.

Note that the embedded mp4 video files that contain the audio in the following are by default muted.

Example 1: Added third on a distorted electric guitar

In this example we add a harmony line a third above the input melody, which is played on a distorted electric guitar. Here are the input signal used, as well as the final result:

guitar_distorted_E_major_ex1.mp4

guitar_distorted_E_major_ex1_added_3.mp4

And here is the code used to generate this output:

import PyHarmonize

# Create dictionary with parameters
parameters = {'input_filename':'./guitar_distorted_E_major_ex1.wav', # input audio is in the key of E major
              'output_filename':'./guitar_distorted_E_major_ex1_with_harmony.wav',
              'key':'E',
              'mode':'major'}

# Generate instance of the class harmony_generator
harmony_generator = PyHarmonize.harmony_generator(parameters=parameters)

# Add harmony
# Note that scale_degrees = [3] means we add one melody line,
# which is always three notes higher within the scale. Depending on the note
# played, "three notes higher within the scale" is either 3 or 4 semitones up.
output_dictionary = harmony_generator.add_harmonies(scale_degrees = [3])

Example 2: Added third and fifth on a distorted electric guitar

In this example we add two harmony lines to an input signal. Here are the input signal and the result:

guitar_distorted_E_major_ex2.mp4

guitar_distorted_E_major_ex2_added_3_5.mp4

The code for this example is essentially the same as in the first example, except that now the list scale_degrees contains more than one element:

import PyHarmonize

# Create dictionary with parameters
parameters = {'input_filename':'./guitar_distorted_E_major_ex2.wav', # input audio is in the key of E major
              'output_filename':'./guitar_distorted_E_major_ex2_with_harmony.wav',
              'key':'E',
              'mode':'major'}

# Generate instance of the class harmony_generator
harmony_generator = PyHarmonize.harmony_generator(parameters=parameters)

# Add harmony
output_dictionary = harmony_generator.add_harmonies(scale_degrees = [3, 5]) # add third and fifth

If we add some more octaves and thirds, we can generate a more synthesizer-like sound. Here is an example for that:

guitar_distorted_E_major_ex2_added_3_5_octaves.mp4

To generate this output, we pass scale_degrees = [-8, -6, 3, 5, 8, 10], which adds pitch shifted signals an octave lower (-8), the third one octave lower (-6), a third up (3), a fifth up (5), an octave up (8), and a third an octave higher (10).

Example 3: Added third, fifth, and octave on a clean electric guitar

In this example we add thirds, fifths, and octaves to a melody in A major, which is played on a clean electric guitar. Here are input and output files:

guitar_clean_A_major.mp4

guitar_clean_A_major_added_3_5_8.mp4

The code for generating this harmony is:

import PyHarmonize

# Create dictionary with parameters
parameters = {'input_filename':'./guitar_clean_A_major.wav', # input audio is in the key of A major
              'output_filename':'./guitar_clean_A_major_with_harmony.wav',
              'key':'A',
              'mode':'major'}

# Generate instance of the class harmony_generator
harmony_generator = PyHarmonize.harmony_generator(parameters=parameters)

# Add harmony
output_dictionary = harmony_generator.add_harmonies(scale_degrees = [3,5,8])
# The list
#       scale_degrees = [3, 5, 8]
# means that we add four melody lines:
# 1. a third up
# 2. a fifth up
# 3. one octave up

Installation

To install the module PyHarmonize, as well as its requirements (NumPy, SciPy, librosa, and SoundFile), clone this repository and run the installation script:

>> git clone https://github.com/juliankappler/PyHarmonize.git
>> cd PyHarmonize
>> pip install -r requirements.txt
>> python setup.py install

PyHarmonize: Adding harmony lines to recorded melodies in Python

Related tags

Overview

PyHarmonize: Adding harmony lines to recorded melodies in Python

About

Examples (with audio files)

Example 1: Added third on a distorted electric guitar

Example 2: Added third and fifth on a distorted electric guitar

Example 3: Added third, fifth, and octave on a clean electric guitar

Installation

Owner

Julian Kappler

Land Cover Classification Random Forest

Dive into Machine Learning

This is my implementation on the K-nearest neighbors algorithm from scratch using Python

Bottleneck a collection of fast, NaN-aware NumPy array functions written in C.

Kats is a toolkit to analyze time series data, a lightweight, easy-to-use, and generalizable framework to perform time series analysis.

Apache (Py)Spark type annotations (stub files).

Tutorial for Decision Threshold In Machine Learning.

Flask app to predict daily radiation from the time series of Solcast from Islamabad, Pakistan

An open source framework that provides a simple, universal API for building distributed applications. Ray is packaged with RLlib, a scalable reinforcement learning library, and Tune, a scalable hyperparameter tuning library.

Python library which makes it possible to dynamically mask/anonymize data using JSON string or python dict rules in a PySpark environment.

Model Validation Toolkit is a collection of tools to assist with validating machine learning models prior to deploying them to production and monitoring them after deployment to production.

A linear regression model for house price prediction

ThunderGBM: Fast GBDTs and Random Forests on GPUs

A webpage that utilizes machine learning to extract sentiments from tweets.

Case studies with Bayesian methods

database for artificial intelligence/machine learning data

Repository for DCA0305, an undergraduate course about Machine Learning Workflows and Pipelines

A Lightweight Hyperparameter Optimization Tool 🚀

A simple and lightweight genetic algorithm for optimization of any machine learning model

A visual dataflow programming language for sklearn