> If you don't need/use those settings, they can be removed
> by editing the Controllers for the Device:

This sounds exactly what I need. Excellent!

>    There are definitely bugs WRT rewinding CCs.  We just don't do that
> at all.  The only way to reliably get what was in the sequence is to
> start from time 0.  Bug #1560 covers this issue.  Not sure I'll ever
> have the necessary time to address this properly, but you might want to
> try back-leveling to 16.02 and see if you like that better.  It does
> make an attempt at sending the proper CCs at playback.  And with the
> Controller removed from the Device, you might see exactly what you want.

Okay. My problem with volume CC's is a bit strange, as I have not touched
track parameters, mixers or recorded any CC (ok, should still double-check
the events AND log the MIDI output).

>    Let me know how it goes as this will be a second vote for me to at
> least putting in an option to provide 16.02's behavior in the current
> version.  That shouldn't be too hard to do.

I feel like the bad guy, but I just cannot afford (time wise) trying
different versions right now. I stay up late one night to get something
done, then I may not do anything the next night because of staying up too
late. So no vote for now.

A few words about previous talk: panic button. Should need no specific
reason to have that implemented, as it's standard feature of every
sequencer I've seen. It's not a nice thing to hit when doing live playing,
but surely in "just browsing" mode it will be ok. I did not try it for one
stuck note mainly because I don't believe in it. And that is not anybody's
fault here. I've been trying to use Ardour for MIDI work :D I had stuck
notes on STOP or just play! And the panic button did not help. Ardour stops
notes immediately on track mute for example, so somewhere the logic works.
Sometimes I did see note off being sent before note on, which obviously
caused stuck notes (if pressing stop amidst the note(s) playing), but let's
not get into Ardour, it's now being rewritten largely so maybe bug hunt is
waste of time over there. Surely we can learn from other software.

Stuck notes can be useful, yes :)

A thought: notes can also be cut by means of active sensing. That can be
done in the intended way (unplugging a cable or disconnecting software MIDI
flow for 300 ms). May not be ok for every situation, not all devices /
plugins may not support active sensing and I don't know if rg supports
active sensing, but that's one way. If there is active sensing from rg,
include a 300+ ms break when using the panic button.

Which notes should be playing at any given moment? I've thought about the
note bookkeeping and how it should be done: integer array(0..127) for each
channel of each MIDI port counting how many times a certain note on has
been sent. When need to stop MIDI notes (stop, track mute, change track
focus or other reason), send as many note offs as required to cancel the
note on  "on count" > 0. Synthesizers, at least the hardware ones I've
used, can play multiple notes for the same key. It does not always help
sending note off just once. These array values would be incremented and
decremented regardless of the "source" (external input or rg segment data),
so it should not be a burden. Admitting I have no idea how things are
implemented already.

(I really try to keep this short...)

Sami Jumppanen
Rosegarden-user mailing list
Rosegarden-user@lists.sourceforge.net - use the link below to unsubscribe

Reply via email to