So is the question essentially how to visualise sound? If so the process is essentially:
- Decide what you want to achieve (something abstract like these music-based sculptures? Something with practical useful value? Something with a dual purpose?)
- Decide which sound variables (frequency, magnitude, etc) are most useful to that goal - will involve some research, direct questions to technical places like Audio-Video Production Stack Exchange
- Find a way of turning your sound source into digital data of the target variables - a technical question probably better asked at Audio-Video Production Stack Exchange
Make the thing. This could be literally anything from a chart to a 3D printing extravaganza.
You'll find useful stuff by searching for generative art based on sound. Here's a good intro to generative art, including similar recommended tools to the above. It also links to a couple of projects based on sound:
As you'll see, it's all about customising some kind of script-based tool to interpret a stream of data (audio data over time) in a way that suits the needs of your project.
Also look into data visualisations based on audio. Here's some examples based (loosely...) on audio from popular films that has an aim of raising a smile of recognition rather than being an analysis tool...
(Jurassic Park)
Here's an image of some abstract sculptures based on alt rock albums, just because.
In this particular case, the making process happened to look like this:
It's extremely varied. The common components are: clear goal, data collection, data processing, then output, which can be literally anything.
Even then, there are exceptions. Here's an example that involves no code at all:
They're made by stretching a balloon over a speaker, then dropping
paint in the center, before blasting it with a single sharp note. The
paint splashes, and a camera set on what looks like a gravity ride
spins around the speaker, snapping away at 5,400 frames a second.
It's all about assembling a mechanism - probably uniquely built, arranged or coded for the project - to make variables in the sound sculpt the final output - then finish it by hand.