Sorry about the incessant chatter
@bluebell
My sequencer is Rosegarden - I've been using that for years, and in spite of some quirks and problems it just perfectly suits my workflow.
I use no effects on Yoshimi apart from the built-in ones, which cover just about everything I need. However, the sound patches I make are very often multi-layer and multi-part.
Guitar goes into my soundcard clean via a simple pre-amp, then I use Timemachine to record the 'dry' signal, while listening via Rakarrack (usually with just some compression and reverb). I find this gives hassle-free clean recording while I get the vibe of playing to the effects.
A totally clean recording means it's possible to do some editing which will then be made quite invisible by adding effects on afterwards.
All finalised guitar parts are then loaded into Rosegarden so they play along with the MIDI tracks, which means I can match up levels and (if I want to) get really tight timing, making fine adjustments to the MIDI - usually I don't so much, as I want the music to 'breathe'.
Incidentally, on the rare occasions I do vocal work, I pass that through Jamin for both EQ and compression.
Mastering, as such, is a final clean-up, top and tail, done in Audacity, usually with a light skim with the fast look-ahead limiter, then level increase to -1dB. This ensures There is a good signal level, but it will never clip even with poorly designed decoders. I use LAME to produce VBR mp3 files, and oggenc for the ogg ones.
