Dear Corky,
I think the most important thing to have is a reference-sound which you do not touch at any time regarding gain / volume. On my rig it’s the piano-sound. It is at 0db at anytime and it is far away from clipping. I made the experience, that the outputs of an interface (Babyface here) is so hot that the engineers want me to lower it anyways. So you don’t loose anything of the dynamic range.
So if I start creating a new song the first thing I do is adding my reference sound and adjust the volumes of the other sounds in relation to the piano-sound.
Another trick that really helps is adding a button, that sets mainvolume -20db. With that trick you can hear very easily which sounds are way too loud. I think this has to do with the natural compression of the ear. If you have an overall loud-volume your ear will compress the sounds and it’s hard to tell which sounds are way to loud. So turn down the volume significant and play your song. If you still here an instrument very well you know: It’s too loud and will blow your ears on stage.
I don’t use compression as VST too much. But on most of my natural sounds (especially strings) I limit the dynamic range of the velocity-input. To be honest: As keyboarder you might have the widest dynamic range on stage and that is a problem. I learned, that I have to limit myself regarding dynamic when playing in a band. And I really don’t need a range of 0-127 on a strings-sound… especially if it’s just a pad.
And of course it’s very useful to have a song-volume-rack where all other instrument racks are routed into audio-wise. This allows you to easily adjust the volume of the whole song with one slider.
I hope that helps! 