Thanks a lot Guido. This confused me many times. I thought there is certain formula for this. Like 1k--1.5k Hz is the vocal area and curve out those in Guitars would give vocals room. But you're right, it always goes down to specific cases.
My ears are still very beginner and if I look at the audio spectrum, they are all similar to me.. Vocals has more 1.5 k, acoustic has a bit more 1k and amps has a bit more 2k.
For the multiband-eq/compressor technique you mentions, should I sidechain it to the other track, or just compress the track itself to make it more even?