I’ve used those random ambient noise generators,such as Noisli, for a while, especially to help me focus when programming but they’ve lost their effectiveness over time since they’re the same noise repeated over and over again. My brain can easily tell that it’s unnatural and inorganic which makes those useless, at least after listening to it after a while.
I propose using Deep Learning to train on actual samples of ambient noise such as in a cafe, in a train station, out in the woods, etc. then have it generate new noises (the same way DL has been used to generate new images of Pokemon, new Game of Thrones scripts, new music, etc.)
What do you think?