Ted and all:

Unfortunately, I am the bearer of bad news this morning. 

I was not comfortable with changing my device not to use timestamps being the answer.  In particular, why would it record successfully (using the bad timestamps) when I recorded into a single track, rather than into two tracks simultaneously? 

So just to be sure, I repeated my test this morning, and to my dismay, it failed two out of two times.  The failure experienced was hanging when the Stop button was pressed, as before. 

So the problem still remains, and the way of duplicating the problem I sent you earlier will still work. 

I don't know why it works sometimes, and fails other times.  It may be the result of missing a note-off event in the recording process, which happens only some of the time. 

This behavior could explain why I wrongly concluded that the problem was not in the earlier version of Rosegarden. 

- Aere

On Wed, 2012-06-13 at 17:43 -0400, Ted Felix wrote:
On 6/13/2012 1:27 PM, Aere Greenway wrote:
I have come up with a way for you to reproduce the problem on your

  Wow.  Thanks for all this.  I'll see if I can squeeze some time in on this soon.  That repeating error message was related to the quantizer getting confused with notes of certain durations.  (At least the last time I worked on it, this is what the issue was.)  So I think the quantizer is a bit fragile and may be the culprit here.

  Here's my email that I drafted a while ago detailing the tests that I performed.  No crash, but I did verify the problem with recording multiple channels and hearing the wrong instruments:

While testing with two keyboards on different channels, I was able to verify one of Aere's observations.  In rg 11.11.42, if you have two keyboards, each set to a different channel, and recording on different tracks.  And if you have a different instrument on each track (e.g. piano and organ), you will hear a different instrument from each keyboard *while recording*.  While in "stop" mode, you will hear the instrument of the currently selected track for both keyboards.

Fast-forward to the latest svn, and the behavior is different while recording.  Instead of each keyboard playing its track's instrument, they both play the currently selected instrument (like when stopped).

Here are the steps to reproduce for those with only a single hardware MIDI keyboard:

1. Launch jack and fluidsynth.  (I use a script that I wrote for this.
   See "My Audio Script" at http://tedfelix.com/linux/linux-midi.html)
2. Connect hardware MIDI keyboard as usual.  Set it for MIDI channel 1.
3. Launch vmpk (the software MIDI keyboard) and set to channel 2.
4. Launch rg.
5. Make sure both MIDI keyboards are in Studio > Manage MIDI Devices.
6. In the "Recording Filters" in the Track Parameters box, set track 1
   to record from channel 1 and track 2 from channel 2.
7. Set instrument 2 to an organ to make it obviously different from
   the piano in track 1.
8. Arm tracks 1 and 2.
9. Press record.
10. Play on both keyboards.

With 11.11.42, the hardware keyboard on channel 1 will make piano sounds.  The software keyboard on channel 2 will make organ sounds. With the latest svn, both keyboards sound like either piano or organ depending on which track is currently selected.

Seems to me that the old behavior is more useful and we should try and get it back.

  Additionally, when recording to two channels, notes were dropped on channel 2.  This sounds similar to what you were describing.

  I think we've got three or four rather serious bugs to track down.  I'll move these to the top of my list for when I have some time.  We should probably open some bug reports to track these.  Feel free to do that if you want.

  If Tom has some time, he should be able to fix the problem with multiple channels all making the same sound when recording.  Although it might be a bit tricky.


Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________ Rosegarden-devel mailing list Rosegarden-devel@lists.sourceforge.net - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-devel