Messing with audio is a peripheral task so expecting me to take my eyes off the road etc is still unacceptable. Especially to change the track or adjust volume.
That said, we're all used to small touchscreens. These tend to require the user to maintain the current context on what is on screen. The whole screen is subject to change and this complicates / weakens muscle memory. I've used larger touchscreens where controls don't change around so much and muscle memory can easily kick in if the design is sane. The difference is tangible. Eg you can blind press using peripheral vision and it can work well.
The real answer might just be that we're in the early days of touchscreen use. The technology and application thereof is still immature. The design language, for want of a better term, is still evolving.