Hey!
I just noticed something : what comes out the track outputs sounds different from what comes out the main outputs.
Say you make a very simple pattern, e.g. using a ride sample regularly placed in the pattern, close enough in time so that the ride sample's tail overlaps with the next sample occurence when the pattern is played.
Listen to the result from the main hydrogen outputs, then listen again but from the ride track outputs directly (bypassing the main hydrogen outputs, you can redo the routing in e.g. qjackctl). There's definitely a big difference : when listening from the main outputs, you can hear that each sample occurence plays fully, meaning the sample's tail is not cut when the next occurence is played. This does not sound to be the case when listening directly from the track outputs, instead, it sounds rather staccato-like.
Why ?
EDIT: I am looking at the code right now and I can see that you don't apply the low pass resonance filter on the direct track outputs. I'll modify the code for myself and check that it's working fine. If anybody is interested, I'll post a patch.
EDIT2: I am not sure any longer it has anything to do with the issue described above ...
Hey! thanks for looking into this as well! I came to the same conclusion last night after some wandering inside the C++ guts (by the way, I use snavigator for that, not too bad).
It will require some rework to achieve what we want ... damnit, I have little time for coding anything ... but let's see, if I have anything, I will post back. Yeah, that's the open-source spirit I guess ;)