Well, I've been following pop music for 45 years now. Styles come and go, but signal processing has been ever-increasing in music technology.
It seems hard to believe that guitarists used to plug their guitar straight into an amplifier. Now it seems the average guitarist has half a dozen effect pedals, and some have way more. More processing is also moving into guitars, and into guitar amps. Digital simulations of vacuum tube guitar amps are at the point where some people prefer them.
Keyboards have mostly turned into digital synthesizers. They can generate simple waveforms into digital filters, trying to simulate various analog synthesizers, or they can play back recordings of actual instruments, or they can do a physical model of an actual instrument.
The recording studio has now migrated into a home computer. You have more capabilities at your finger tips than the Beatles ever had - you are only missing a good sounding room, some excellent microphones, and perhaps a spot of talent. OK, talent, that's tough, but there are multi-element microphones using signal processing coming in a few more years that should be pretty amazing.
Vocal pitch correction is everywhere - illustrating the fact that some inventions may be used for evil. Ultra-light weight power amplifiers enable any band to be far too loud without having to carry around hundreds of kilograms.
Electrical engineering has done all this. It's mostly digital signal processing these days. A Bachelor's in EE, or perhaps a Master's, specializing in signal processing would be my recommendation if you are seriously interested in these kinds of things. You also need ears, because frequently there is no way to measure results. Oh, you can compute a signal-to-noise ratio - but that doesn't necessarily tell you if something sounds good or bad. A minor in the Music Department at college would be a good idea. Or just spending a lot of time in a home studio, where one constantly faces the question of - is it a good sound? or a bad sound?
So what sorts of things might you be asked do at a job? Perhaps write some C code to run on a PC for an audio mixer, maybe bass, treble, and midrange controls. Maybe design a signal processor architecture to be implemented on a gate array, using enough bits to sound good, but not so many that the result is too expensive or too slow. How about a pitch estimator that has to work in real time. Or a digital model of a Maestro FZ-1A germanium fuzz box, that sounds spot on to the original.
There is some math involved - although the actual code is mostly adds, subtracts, and multiplies. (I've seldom been able to do a division - far too expensive.) Linear systems theory is what's behind all this. It had a reputation as an incomprehensible course at my school. Maybe it was, I didn't take it; I just picked it up on my own. You have to understand it well enough to just use it as another tool; that does take some time. I suppose I have been fortunate to have been given the time to study new things during the course of a project. I would also advise against excessive reliance on canned software packages, whose workings you don't really understand. If you have a serious math phobia, then this might not be such a good idea.
The bottom line is that you will spend most of your time writing code on a computer, occasionally playing some music to test things out, but sometimes getting bored enough to go online and try to encourage others to do the same. Seriously, though, I have found this a very nice career.
