View Issue Details
|ID||Category||Date Submitted||Last Update|
|0006882||bugs||2016-05-21 15:47||2020-05-16 15:36|
|Platform||Ubuntu Studio 16.04 Xenial Xerus||OS||Linux||OS Version||Maverick|
|Fixed in Version|
|Summary||0006882: Midi note continue playing|
|Description||With the latest official ardour's update through apt-get update/upgrade a bug appears making MIDI note continue while playing is stopped.|
I didn't notice this behaviour before.
Nevermind if start then stop, MIDI note still continue.
Mute the related track stop the sound.
Re-starting Ardour clears the sound and allows continue composing music.
But it is not comfortable.
|Steps To Reproduce||- play multiples MIDI and Audio tracks|
- while playing (and maybe just when a new MIDI note start but before the end of the current measure) move the time cursor back
- if you stop playing you should hear the MIDI notes playing
- could concerns multiples MIDI tracks simultaneously
|Tags||Midi, notes, stop, transport|
I have faced the same problem since the first time I tried Ardour (v3.?). The problem hasn't gone away. I have confirmed this with 4.6 (distributed with Ubuntu Studio 16, also completely removed and reinstalled) and 4.7 (demo). Working with MIDI is impossible.
Today I used MIDI Tracer to see what is going on with the MIDI output from my MIDI track. MIDI Tracer confirms what I hear: no note off's (nothing) is sent at STOP. Notes that were on, stay on.
I also confirm that track MUTE does send note off's (for the notes playing) and sustain pedal off for all MIDI channels, but that may not work 100% when toggling MUTE on/off, some notes are stuck on, sometimes.
I have no understanding why even Panic button does not help turning all notes off. I have heard it turning some of them off.
Audio MIDI basic setup 2016 tests (no audio).zip (21,337 bytes)
I uploaded my session with just one MIDI chord. Start playing and press STOP during the chord. I have verified this on hardware MIDI synths, I haven't used software synths at all.
My computer: Ubuntu Studio 16, Intel Core 2 Duo (2.0 GHz, performance governor - though ondemand does exactly the same). USB MIDI adapter. MIDI plays out fine and in tight timing (no noticeable jitter). MIDI plays fine with audio track as well.
||Note: I will capture MIDI messages with gmidimon as soon as I have the time for it.|
I verified the problem with JACK setup as well.
Here is a video showing that even a single note will remain on at pressing STOP (no NOTE OFF):
Here is a corresponding video showing that TRACK MUTE triggers NOTE OFF:
I realised a "twist" in NOTE OFFs, why they are so important, why the notes get indefinitely stuck. It's because polyphonic synthesizers (at least hardware, at least Yamaha TG33, Yamaha MU128) can play the same note (overlapping) a number of times.
Turning all the voices of the key (MIDI NOTE) off requires sending as many NOTE OFF events for the same note as there were NOTE ON events.
Simple example of why this becomes an issue: have one long note in sequence, play the note, stop during the note, go back and play the note again -> that's a total of two NOTE ONs for the same key and no NOTE OFF. When playing to the end of the note, one NOTE OFF is sent. That still leaves one note playing.
As it's been said TRACK MUTE sends note offs but may miss them, seemingly randomly. I guess it has something to do with timing. I witnessed the same problem with MIDI event list when modifying the MIDI NOTE by scroll wheel on the mouse. Scrolling quickly enough some notes were left playing on the synthesizer, which means that note offs are missing.
||I can reproduce the issue described.|
I think this issue is fixed in Ardour 5.0rc1. The notes played are tracked correctly, and note off are sent
a) When starting/stopping the transport
b) When moving the play head (jumping)
On track mute, "ALL NOTE OFF" signals are sent, so if notes are missed it's probably the synthethiser's fault.
Also, midi note trackers in ardour *do* count the number of note-on and will send as many note off as there have been note ons.
Please check in Ardour 5rc1 if the issue is still there
||The same issue is reported in id 0002139, and it mentions the same problem in Ardour 5.4.|
FirthofFifth.ardour (120,585 bytes)
Genesis - Firth Of Fifth-t0-1.mid (282 bytes)
Genesis - Firth Of Fifth-t1-1.mid (20,220 bytes)
Genesis - Firth Of Fifth-t2-1.mid (18,657 bytes)
Genesis - Firth Of Fifth-t2-1-4.mid (18,619 bytes)
Genesis - Firth Of Fifth-t3-1.mid (17,231 bytes)
Genesis - Firth Of Fifth-t4-1.mid (24,173 bytes)
Genesis - Firth Of Fifth-t5-1.mid (24,254 bytes)
Genesis - Firth Of Fifth-t6-1.mid (75,602 bytes)
Genesis - Firth Of Fifth-t7-1.mid (13,694 bytes)
And the last two files. It appears that my notes for the above got clobbered due to the spam detector. This is a file that I was working on that has stay on note problems. I am unsure how to fix it. I tried splitting the file up between two external synths in attempt to balance the number of notes going to each device. Rosegarden and Tuxguitar both have no problems playing the same files.
Genesis - Firth Of Fifth-t8-1.mid (2,591 bytes)
Genesis - Firth Of Fifth-t9-1.mid (10,889 bytes)
||Just noticed I forgot to update the OS version. I am running on Ubuntu Studio 19.10|
I am not sure what your intended purpose is. This bug was reported for early version of 5.x of Ardour. We are now close to a release of 6.0 and 5.12 is the current version.
What version of Ardour are you using? Instead of attaching a whole bunch of MIDI files and the session file, please use Session > Archive and attach the result.
Ardour is version 5.12 installed from Ubuntu Studio 19.10 repository. I attached the files so you could see if it was reproducible as the issue still exists (I see from your pics that you have some MIDI capable keyboards that you can try it against. If you wanted some files to test against version 6.0, these may be useful. As a software developer, I always appreciated when clients sent me files that would produce errors for me to work with. If the session archive will bundle everything together better, no problem, I have attached the file as requested.
I really like the Ardour workflow, but it is currently not a usable solution for me. I look forward to the 6.0 release in hopes the new MIDI code has fixed this issue.
FirthofFifth_2019-12-14_184237.tar.xz (37,130 bytes)
||Thanks. I'll do some testing. I know this piece very very well :)|
||any particular things i should try to trigger the stuck notes?|
||They tend to appear around the 5min mark and progressively get worse until it sounds really bad. Stopping will silence them, but muting track 4 or 11 (where they happen) does not appear to stop them. 66 F#4 seems to be one of the notes that stays playing.|
||I have been progressively removing notes from tracks 4 and 11 in hopes that I would be able to stop the problem. The current file has quite a few notes removed now. I can try bringing back the full files to see if it will fix it. The original file is from https://tabs.ultimate-guitar.com/tab/genesis/firth-of-fifth-guitar-pro-219878 which I exploded into midi files using tuxguitar. I was working on fixing the guitar solo as the tab for it is not even close to what Steve is playing.|
||I did a straight run through with zero stopped notes. I also did a run through with occasional start/stop commands. no stuck notes.|
||I am using ardour instead of jack if that makes a difference.|
||Doh!! ALSA instaead of Jack...|
||What SR and buffer size(s) have you been using?|
Ok, there is a stuck note for sure that startss somewehere near 5 mins on track 10 for sure (you may be calling this track 11)
The bug appears to be that when the region ends on precisely the same sample as the note, we do not handle the noteOff properly. Extending the length of the region by 1 sample fixes this.
Fix in the current code base expected in a few hours.
||Cool, thanks :-)|
||and, as of now, i can no longer reproduce it. the good news: investigating this led me to track down an inadvertently introduced bug that had the opposite effect of not playing notes near zero.|
||Nice! Always good to get a two for one :-)|
||I will try to download and compile a new version week|
||If you've paid for ardour in the past, nightly builds are always available at nightly.ardour.org|
||I built it from source. Just have to figure out why ALSA is no longer an option when starting it as I tend not to use Jack.|
||Any chance for the fix to be back ported to 5.12?|
||No chance at all. There will be no further 5.x releases.|
||When does 6 get released?|
Rule Number 1 about 6.0 release date: do not discuss 6.0 release date
||Any chance you would tell me the file name that got the edit so I can go fix it until 6 gets released?|
||not possible. MIDI rendering has been completley changed in the current master branch,|
I kept running into this problem with the Windows 64 bit build and my Yamaha S80 as a controller & sound generator. It would be very inconsistent though: sometimes playing back a MIDI track, the Note Offs would get sent, but then replaying moments later with Record enabled, they wouldn't (all confirmed with the MIDI logger enabled).
I tried out the Ardour-6.0.rc1.232 nightly and the problem doesn't seem to be reproducible now. Thanks!
this particular bug report is now resolved.
similar/related issues should be filed in new reports.
Issue has been closed automatically, by Trigger Close Plugin.
Feel free to re-open with additional information if you think the issue is not resolved.
|2016-05-21 15:47||cyril||New Issue|
|2016-06-07 15:41||Some-E||Note Added: 0018237|
|2016-06-07 16:01||Some-E||File Added: Audio MIDI basic setup 2016 tests (no audio).zip|
|2016-06-07 16:04||Some-E||Note Added: 0018238|
|2016-06-07 16:05||Some-E||Tag Attached: Midi|
|2016-06-07 16:05||Some-E||Tag Attached: notes|
|2016-06-07 16:05||Some-E||Tag Attached: stop|
|2016-06-07 16:05||Some-E||Tag Attached: transport|
|2016-06-09 06:57||Some-E||Note Added: 0018244|
|2016-06-11 13:05||Some-E||Note Added: 0018248|
|2016-06-12 08:05||Some-E||Note Added: 0018249|
|2016-06-12 08:08||Some-E||Note Added: 0018250|
|2016-06-12 08:08||Some-E||Status||new => confirmed|
|2016-07-30 11:01||_FrnchFrgg_||Note Added: 0018331|
|2016-07-30 11:01||_FrnchFrgg_||Status||confirmed => feedback|
|2017-09-26 07:27||Some-E||Note Added: 0020041|
|2019-05-30 11:10||x42||Relationship added||duplicate of 0002139|
|2019-11-26 16:52||ixd704||File Added: FirthofFifth.ardour|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t0-1.mid|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t1-1.mid|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t2-1.mid|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t2-1-4.mid|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t3-1.mid|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t4-1.mid|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t5-1.mid|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t6-1.mid|
|2019-11-26 16:52||ixd704||File Added: Genesis - Firth Of Fifth-t7-1.mid|
|2019-11-26 17:57||ixd704||File Added: Genesis - Firth Of Fifth-t8-1.mid|
|2019-11-26 17:57||ixd704||File Added: Genesis - Firth Of Fifth-t9-1.mid|
|2019-11-26 17:57||ixd704||Note Added: 0020830|
|2019-12-13 18:01||ixd704||Note Added: 0020843|
|2019-12-14 15:52||paul||Note Added: 0020846|
|2019-12-14 23:44||ixd704||File Added: FirthofFifth_2019-12-14_184237.tar.xz|
|2019-12-14 23:44||ixd704||Note Added: 0020847|
|2019-12-15 17:09||paul||Note Added: 0020848|
|2019-12-15 18:40||paul||Note Added: 0020849|
|2019-12-15 21:15||ixd704||Note Added: 0020854|
|2019-12-15 21:21||ixd704||Note Added: 0020855|
|2019-12-15 21:22||paul||Note Added: 0020856|
|2019-12-15 21:26||ixd704||Note Added: 0020857|
|2019-12-15 21:28||ixd704||Note Added: 0020858|
|2019-12-15 21:30||paul||Note Added: 0020859|
|2019-12-15 23:56||paul||Note Added: 0020860|
|2019-12-16 01:11||ixd704||Note Added: 0020861|
|2019-12-17 00:43||paul||Note Added: 0020863|
|2019-12-17 13:06||ixd704||Note Added: 0020864|
|2019-12-17 13:07||ixd704||Note Added: 0020865|
|2019-12-17 15:10||paul||Note Added: 0020866|
|2019-12-17 18:14||ixd704||Note Added: 0020867|
|2020-02-01 22:02||ixd704||Note Added: 0020944|
|2020-02-01 23:51||paul||Note Added: 0020945|
|2020-02-02 18:46||ixd704||Note Added: 0020947|
|2020-02-02 18:56||paul||Note Added: 0020948|
|2020-02-02 18:56||paul||Note Edited: 0020948||View Revisions|
|2020-02-02 18:57||paul||Note Edited: 0020948||View Revisions|
|2020-02-02 19:03||ixd704||Note Added: 0020949|
|2020-02-02 19:06||paul||Note Added: 0020950|
|2020-05-09 16:48||rimaraf||Note Added: 0024113|
|2020-05-16 15:35||paul||Assigned To||=> paul|
|2020-05-16 15:35||paul||Status||feedback => resolved|
|2020-05-16 15:35||paul||Resolution||open => duplicate|
|2020-05-16 15:35||paul||Note Added: 0024172|
|2020-05-16 15:36||anonymous||Note Added: 0024173|
|2020-05-16 15:36||anonymous||Status||resolved => closed|