1. IMPORTANT:
    We launched a new online community and this space is now closed. This community will be available as a read-only resources until further notice.
    JOIN US HERE

Bandstand doesn't acknowledge note off events

Discussion in 'BANDSTAND' started by JF Paris, 16/7/06.

Thread Status:
Not open for further replies.
  1. JF Paris

    JF Paris Member

    Messages:
    51
    When Logic 7.2.1 starts playing [a Bandstand track], it does send note-on/off events to Bandstand, but the triggered notes won't stop... until the sequencer encounters a new region on the same track, or loops on the very same region for the 2nd time. This happens whatever the midi channel or the loaded instrument — except drums on channel 10, obviously. Yet this never happens when using the Kontakt (2) plug-in instead, albeit with the same nkm file, hence the same instruments and samples...

    Another possible workaround is to create a dummy [empty] midi region before any Bandstand region...

    Q1: is this a known bug, and is it being addressed?

    Q2: has anyone found another efficient workaround for this problem?
     
  2. broc

    broc Forum Member

    Messages:
    144
    Do you have sustain events (cc#64) somewhere on these tracks?
    (just check the regions in event list window)
     
  3. JF Paris

    JF Paris Member

    Messages:
    51
    note off events not acknowledged

    Of course, no!... And if I had, wouldn't Kontakt acknowledge them just the same?

    Isn't my description of the bug accurate enough? Note-off events aren't recognized by Bandstand before they originate from a new [or looping] region. No other plug-in behave like this, using the very same track data.

    Steps to reproduce:
    • create a new midi track, and create a one-bar region on it
    • put a short 1/8th note somewhere in that bar
    • assign it to Bandstand and select a melodic instrument
    • play the track: the 1/8th note sustains till you stop the sequencer
    • loop the region and play the track again: notice the difference?
     
  4. Markus @ NI

    Markus @ NI New Member

    Messages:
    989
    Hi jf@ai.univ-paris8.fr,

    I thought we had fixed all these issues with the Bandatand 1.0.1 version.

    we did change some more in this area with the latest Kontakt versions. should be fixed in Bandstand in the next software update then.


    best,
    Markus
     
  5. broc

    broc Forum Member

    Messages:
    144
    Oops. I didn't expect such an elementary bug in version 1.x software...
     
  6. Markus @ NI

    Markus @ NI New Member

    Messages:
    989
    Hi all,

    I have not really confirmed this bug - we need to look into it at a later point.

    the fixes we did in Kontakt were related to interpreting CC#120 and CC#123 correctly when some hosts send it instead of note offs when the user presses stop. the below might or might not be releated.

    is anybody else having the same problem in Logic?


    best,
    Markus
     
  7. broc

    broc Forum Member

    Messages:
    144
    Markus,

    perhaps the problem is due to a general subtlety of note off events.
    As far as I know Logic sends 'note off' as 'note on' with velocity 0.
     
  8. Markus @ NI

    Markus @ NI New Member

    Messages:
    989
    Hi all,

    I am a PC guy so I have no quick way of checking the scenario below. sent it on to the support to enter a bug if it repros.

    it sounds all quite familiar to me. something with the stamps of the note on events and note off events being the same so that only the second note off stops the first note on? pretty sure we fixed that in the Kontakt engine after Bandstand was released.


    best,
    Markus
     
  9. JF Paris

    JF Paris Member

    Messages:
    51
    event time stamps

    But this isn't the case: the steps to reproduce — see above — clearly state that we're not confusing note on/off time stamps, since in a 1/8th note, chances are the onset and offset are many hundred ticks apart. Moreover, the second note off is the very same event as the first one, since the region is looped.
     
  10. broc

    broc Forum Member

    Messages:
    144
    I've just monitored the midi events sent from Logic into the plugin (using Bidule AU). At sequencer start Logic sends CC#123 (all notes off), at note begin it sends note on, at note end it sends note on with velocity 0. This is correct and conforms to the midi standard.

    However, Logic does NOT send any event when the sequencer encounters a new region! So it's complete mystery to me why Bandstand would stop the note at this point...
     
  11. Markus @ NI

    Markus @ NI New Member

    Messages:
    989
    weird indeed!

    we have to deal with a lot of unuique behaviors in every single host.

    does it make a difference if you lengthen your original note? what if you change the note on 0 to a note off event?


    best,
    Markus
     
  12. JF Paris

    JF Paris Member

    Messages:
    51
    No difference: whatever the number of notes, their length, their midi channel, or their position in the bar, they won't shut off until the next region. At first, I suspected some congestion, like too many events at the same time, then I narrowed the problem to only one note and no control event: that's how I came up with the protocol listed above.

    Besides: I can't reproduce this with any other plug-in, including Kontakt 2.

    PS: I haven't tried to convert note on with velocity 0 to actual note off: even if it worked, it wouldn't be a practical solution anyhow.
     
Thread Status:
Not open for further replies.