it's also not a very clever one. I think what the author wants to do is actually segmentation which is much more complex program. the best algorithm Ive used for doing any kind of amplitude following is a detrender with two envelopes running at different rates. next to that is Foote's audio novelty audio algorithm.
I enjoyed this and am going to take a look at the rest of the series, always wanted to learn a bit more about audio processing.
Thanks! Glad you like it :)
Great series. I always wanted to learn about audio processing.
Any learning material for this? Besides this series obviously
It depends on what programming language you'd like to work in and what exactly processing means to you. Are you interested in audio analysis or writing programs that generate audio like a synthesizer for a DAW? For audio analysis I'd recommend essentia a C++ library with Python bindings. For writing plugins, I'd recommend you take a look at JUCE, which has a decent set of tutorials on their website. You'll notice that C++ is pretty standard in industry because speed and hardware access are important. If JUCE is the sort of thing you're looking for, I'd also recommend the Audio Developer Conference talks on Youtube.
I've found rust to be very capable as well, if lacking in runtime libraries with native bindings. It's not too difficult to bind into coreaudio and jack, though.
I've actually been doing a bit of research into using Rust for an upcoming project. The JUCE library is just so fully featured that it would be an undue burden to use any other language to write audio software in. Still, I think Rust is an attractive language for this space and there's definitely some talk about expanding libraries for a JUCE alternative.