Data Driven Audio Signal Processing - A Tutorial with Computational Examples
This tutorial accompanies the lecture Data Driven Audio Signal Processing. The lecture and the tutorial are designed for International Standard Classification of Education (ISCED) level 7 (Master, in total 6 ECTS credits).
Jupyter notebooks can be accessed via the services
- dynamic version using mybinder: https://mybinder.org/v2/gh/spatialaudio/data-driven-audio-signal-processing-exercise/dev?labpath=index.ipynb
- static version using nbviewer: https://nbviewer.org/github/spatialaudio/data-driven-audio-signal-processing-exercise/blob/dev/index.ipynb
- sources (tex, ipynb) at: https://github.com/spatialaudio/data-driven-audio-signal-processing-exercise
Jupyter notebooks with rendered outputs can be viewed at https://nbviewer.org/github/spatialaudio/data-driven-audio-signal-processing-exercise/blob/main/index.ipynb
- v0.1 for winter term 2021/22, initial version
- v0.2 for winter term 2022/23
- v0.3 for winter term 2023/24, many beamer tex slides added, CI
- TBD for winter term 2024/25
- the default branch of the repository is
dev
used for development - all notebook outputs in
dev
branch are cleared for convenient diff handling main
branch contains notebooks with rendered outputs, which is maintained from time to time- do not rely on
main
branch as this is hard reset from time to time - probably in future we rename
main
to somewhat less confusing
The Anaconda distribution is a convenient solution to install a required environment, i.e. to have access to a Jupyter Notebook renderer with a Python interpreter on a personal computer. It is very likely that a very recent installation of Anaconda already delivers most of the required standard packages just using the base
environment. It is however good practice to create a dedicated environment for each project. So, for this tutorial we might use a myddasp
(or whatever name works for us) environment. We might consider the following install routine:
- clone the repo to local machine (if not already available)
git clone [email protected]:spatialaudio/data-driven-audio-signal-processing-exercise.git
(via SSH) orgit clone https://github.com/spatialaudio/data-driven-audio-signal-processing-exercise.git
(via https) or- get a zip file from current
dev
commit via https://github.com/spatialaudio/data-driven-audio-signal-processing-exercise/archive/refs/heads/dev.zip
- get into the folder where the exercises are located, e.g.
cd my_ddasp_folder
- in the subfolder
.binder
theenvironment.yml
can be used to create a dedicated condamyddasp
environment asconda env create -f environment.yml --force
- (we can remove this environment with
conda env remove --name myddasp
)
- activate this environment with
conda activate myddasp
- this should also have installed sound / audio related libraries using pip
pip install pyloudnorm==0.1.0
- we might check this with
pip list
- Jupyter notebook renderer needs to know our dedicated environment:
python -m ipykernel install --user --name myddasp --display-name "myddasp"
- we might want to archive the actually installed package versions by
python -m pip list > detailed_packages_list_pip.txt
andconda env export --no-builds > detailed_packages_list_conda.txt
- start a Jupyter lab environment via a local server instance by
jupyter lab
- start the landing page
index.ipynb
of the tutorial - make sure that the notebooks we want to work with are using our dedicated kernel
myddasp
- University of Rostock:
- Frank Schultz, concept, coding
- Sascha Spors, concept
Please cite this open educational resource (OER) project as
Frank Schultz, Data Driven Audio Signal Processing - A Tutorial Featuring Computational Examples, University of Rostock ideally with relevant file(s), github URL, commit number and/or version tag, year
.
- Creative Commons Attribution 4.0 International License (CC BY 4.0) for text/graphics
- MIT License for software