Add python-soxr

[dofuuz]

(Note: I worte python-soxr)

https://pypi.org/project/soxr/
https://github.com/dofuuz/python-soxr

Python-SoXR is fast and high quality sample-rate conversion library for Python
(It's Python wrapper of libsoxr.)

Ran on Google Colab:
https://colab.research.google.com/drive/1NkJ6n_e08cAjH92Zrlx7nO7xKKNZhCNf?usp=sharing

Jupyter Notebook Viewer:
https://nbviewer.jupyter.org/github/dofuuz/audio-resampling-in-python/blob/soxr/Audio%20Resampling%20in%20Python.ipynb

[jonashaag]

This is great, thank you! Am I correct to assume that #2 is included in this PR? If so, we can close #2.

From #2:

lilfilter should go to 'very good' tier. It's very fast and accurate.

The sweep plots look very good but the impulse plots look only "good" to me. But not sure how important that is.

resampy goes to 'good' tier. It's is still good for most of case, but not best solution.

kaiser_best is pretty good I think. But also not sure here, I'm not an expert.

I think it's better to remove opinion-based text section on ipynb notebook if you are not going to update it.

Definitely agree, feel free to remove that. Maybe we can just add a link on the top of the notebook to the original blog post and remove all/most of the text.

Also I feel it's safe to add soxr to the "very good" list, and feel free to add it to the README list.

If you want you can also work on #4 and #5 :)

[dofuuz]

This is great, thank you! Am I correct to assume that #2 is included in this PR? If so, we can close #2.

Yes you are.

About quality tier things, I'm not 100% sure. I'm not an expert expert about that.

Sorry, I'm not going to improve this further. I already spent too much time on it! 😆