I had an RS7000 for years and I think it would be very well suited to what you want to accomplish.
One thing I really liked about the sampler on the RS7000 that I don't think has been touched upon here is that you can use it like a digital multitrack recorder (it's called KIT+NOTE mode in Yamaha samplerspeak).
Basically, it lets you record audio during song playback, and it remembers where the audio is in the song by placing a MIDI note at that point, sustaining it for the duration of the recorded sample. It's a similar method to how you can "trick" Reason into supporting audio by loading NN-19's with audio phrases triggered by MIDI notes.
Here is a blurb about this from the manual, which may or may not make things clearer for you:
When the KIT+NOTE sampling type is selected, the recorded sample is not only saved as a sample voice, but the note data required to play the sample back is automatically created as well.
For example, if a vocal sample is recorded while listening to playback of other tracks, the recorded sample will be played back exactly as it was recorded in relation to the other tracks — just like a multitrack recorder. Since what was actually recorded was the vocal sample and the note data required for playback, the data can be edited later to modify the sample and the way it plays back in a variety of ways.
I'm not familiar with your approach to live stuff, but I am very familiar with your musical output (as well as most of the Suction catalog) and I think that either the RS7000 or RM1x would work well for you in a live situation. You could use the "multi-track" feature I mentioned to provide further audio backing in the form of 8-16 bar loops.
I've used all kind of hardware sequencers in the past, Q-80's, MMT-8s, MCs, Ensoniq, Akai, etc...to me the Yamaha sequencers are the most difficult to wrap your head around (manual REQUIRED), but very very powerful.
Good luck with your sequencer quest!