Ted:

I did more testing in an attempt to narrow down the problem.  In doing
so, I discovered some significant information.  

First, I found that the problem exists in the prior version of
Rosegarden (11.06) as well.  In testing on Lubuntu 11.10, I observed the
same problem.  

In testing, I found that if I did not use the recording filter, but
recorded both MIDI channels to a single Rosegarden track, the problem
usually did not occur (though it sometimes did).  

I did a recording of the piece that causes the crash, without using the
recording filters (and therefore recording both MIDI channels to track
1).  

I have attached the Rosegarden (.rg) file of that recording to this
e-mail.  It is the first attachment.  

I then set it up to record MIDI channel 5 of my keyboard to track 1, and
MIDI channel 12 to track 2, by using the recording filters.  

In recording, I observed the proper notes going to the two different
tracks.  Everything went okay until I clicked the Stop button after
stopping playing, and then Rosegarden froze-up, having to be terminated
via a kill-signal.  

I took the recording made to a single track, and split it by source
(MIDI channel), and I was able to separate the melody and chords notes
into two separate tracks.  I have attached that Rosegarden file as the
2nd attachment.  

Using the "split segment by source" functionality at least gives me a
way to work around the problem.  

I discovered I can listen to what I am playing while recording (both
tracks with their intended instruments) by connecting (using qjackctl)
my keyboard device to a different Emu10k1 port than the one Rosegarden
uses, and also by turning-off the MIDI-Thru-Routing (so I only hear what
is played on that other Emu10k1 port).  

In the testing I did, I discovered that sometimes notes of the chords
get missed in the recording process.  Chord notes are more likely to be
dropped because all notes of the chord are sent at the same time (one
immediately after the other).  

I discovered that if I selected (in Rosegarden MIDI preferences) the
"Sequencer timing source" as "PCM playback 0-0-0", it seemed less likely
to miss any notes.  

The computer I ran the test on is a Dell Dimension-1100, having a
gigabyte of RAM, and a 2.53 gigahertz processor speed.  I am running
Lubuntu 12.04, with Rosegarden 11.11.42.  

My MIDI keyboard is communicating with Rosegarden (and JACK and the
emu10k1) using the Soundblaster's MPU-401 UART.  I have the two MIDI
cable ends connected using a male-to-male connector plug.  

I use this arrangement at present because there is no connection in
Linux Java MIDI to JACK, but it does see the hardware interfaces (such
as the USB-UNO, M-Audio 2x2, and EMU10K1 MPU-401).  I have used all of
these hardware interfaces (cabled to loop-back) successfully for playing
various devices with my keyboard application currently under
development.  

This is a temporary arrangement, since I plan to write a Java Native
Interface (JNI) library to access the MIDI devices using JACK.  I have
not quite gotten to that stage of the development.  

The reason I'm mentioning this, is I wonder if there are problems
involved in plugging the MIDI input cable into the MIDI output plug.  It
does seem to work in playing various devices, but recording MIDI may be
a different animal.  

Anyway, I hope this additional information helps shed light on the
problem, and with the attached files, you have an example of what
records successfully to a single track, but hangs when recording the
same to two tracks using the recording filters.  

I suspect the hang is Rosegarden looping, sending out the same error
message I passed to you in a prior e-mail.  In some of the testing I
ran, that was indeed what happened (when I ran Rosegarden from a
terminal).  

- Aere


On Sun, 2012-06-10 at 23:01 -0400, Ted Felix wrote:

> On 6/5/2012 4:41 PM, Aere Greenway wrote:
> > WARNING: SegmentNotationHelper::makeNoteViable(): No valid split for
> > event of duration 896 at 107520 (split duration 960), ignoring remainder
> > WARNING: This is probably a bug; fix required
> 
>    Boy, that looks familiar.  I've been in that code and fixed something 
> like that.  Anyway, I'm still looking at your problem, but a bit swamped 
> right now.  I'll get back to this hopefully in a week or so.  I did play 
> with the recording filters (which had no effect, still no crash for me) 
> and I have a response that I need to forward to you (don't have email on 
> my Linux machine, so it gets complicated).  Stay tuned....
> 
> Ted.
> 
> 
> ------------------------------------------------------------------------------
> 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
> [email protected] - use the link below to unsubscribe
> https://lists.sourceforge.net/lists/listinfo/rosegarden-devel


-- 

Sincerely,
Aere

Attachment: ErrorTesting-1.rg
Description: audio/rosegarden

Attachment: ErrorTesting-1a.rg
Description: audio/rosegarden

------------------------------------------------------------------------------
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
[email protected] - use the link below to unsubscribe
https://lists.sourceforge.net/lists/listinfo/rosegarden-devel

Reply via email to