About Neural Audio Lab
Who is behind Neural Audio Lab, how its articles are made, and how to reach it.
What Neural Audio Lab is
How neural networks hear, separate, transcribe and make sound: the research explained, the open models run on the same inputs.
It sets out to explain audio and music AI from the papers and to run the open models on shared inputs, publishing what they measured and what broke.
It is written for developers, students and curious listeners who want to know how audio AI works and whether the models hold up.
What it covers
How it works: spectrograms, onsets, codecs, diffusion; open models run on the same inputs; does the paper's code still run?; from 1988 to now: the field's turning points; separation, transcription, enhancement, generation.
How the articles are made
Articles are researched and drafted with AI, checked against their sources, and reviewed before they are published.
Each one ends with "How this was made": the sources it used, the model that drafted it and what it cost. Nothing about how a page was made is hidden.
Who signs them
Articles are signed with a pen name for Neural Audio Lab's editorial team, so the site keeps one voice. The note under each article says when a name is a pen name.
A pen name is not a person with credentials, and no article claims experience nobody had.
Mistakes and corrections
If something is wrong or out of date, write through the contact page. It is checked, and the article is corrected.
Every correction made after an article was published is noted under the article and listed, with its date, on one page:
How the site is paid for
Some articles may contain affiliate links: if you buy through one, the site may earn a commission, at no extra cost to you. Recommendations are not for sale.