Julie, and all:

As I recall, from months ago, Ted (I think it was Ted) said something about the note-on event of the subsequent note being sorted so it inadvertently came before the note-off event of the prior note, which would produce a zero-length note. 

My recollection of this is uncertain. 

- Aere

On 04/21/2013 03:46 PM, Julie S wrote:
Hello Aera,

Sounds like when we save or load a .rg file we may be mangling things.

I remember -- many moons ago -- that I had to put in a Engineering Compromise (otherwise known as HACK) for notes of zero duration of file load.

Maybe something in that logic is messing things up.  I'm not in a position currently to dig around that code, but maybe for someone that is, this could be a helpfuls starting point.

Julie S

--- On Sun, 4/21/13, Aere Greenway <Aere@Dvorak-Keyboards.com> wrote:

From: Aere Greenway <Aere@Dvorak-Keyboards.com>
Subject: Re: [Rosegarden-devel] Rosegarden Multiple Keyboard Problem (1356)
To: "Ted Felix" <ted@tedfelix.com>
Cc: "Developers, Rosegarden" <rosegarden-devel@lists.sourceforge.net>
Date: Sunday, April 21, 2013, 1:22 PM

On 04/18/2013 04:10 PM, Ted Felix wrote:
> Sounds like the usual MIDI recording problems related to high CPU
> usage by rg.  I really wouldn't trust rg for MIDI recording at all.
> Particularly on low-spec hardware.  It seems to work ok on my Core i3,
> but still uses way too much CPU.  I'm working on this now, but it's
> going to take many months.

This conclusion puzzles me, because I don't think it fits the symptoms.

Remember that I said that the first time I played it back, it sounded fine?

One can't be 100% sure of their recollection of an event, but this is
how I remember it.

I started recording, and was pleased to hear the different instruments
(chords vs. melody), thinking "Ah - it's working."

After stopping the recording, I was pleased that it didn't crash on
hitting the "Stop" button (which it used to do before you fixed that
problem months ago).

Then I played-back the entire recording, and it was perfect - no missing
chord notes (like it used to do months earlier).

So I saved the file for reference, feeling good that the problem was
fixed.  My memory of events here is not clear - I may have re-loaded the
".rg" file.

For some reason, after saving it, I decided to play it back again,
upon-which I noticed many missing chord notes, to which my thoughts were
"What the (expletive deleted)?"

The next day I examined the results in more detail, seeing the missing
notes in the matrix (and notation) editors.  That puzzled me.

I exported the file as a MIDI file (.mid), so my onboard sequencer could
play it.  On playing it with my onboard sequencer, I noticed the missing
chords were there, but they were not continuous notes - if they were
tied, the ties were lost.

I remember there was a prior problem you (or someone) fixed with
exporting MIDI files, where somehow the note-off event was getting
exported immediately after the note-on event of the following note. I
conclude that the fix for this earlier problem was causing it to export
the MIDI file without the missing notes.

But the significant thing in this long-winded e-mail, is that given the
sequence of events I observed, It appears the chord notes were messed up
in the process of saving the ".rg" file - not in recording it.


Precog is a next-generation analytics platform capable of advanced
analytics on semi-structured data. The platform includes APIs for building
apps and a phenomenal toolset for data science. Developers can use
our toolset for easy data analysis & visualization. Get a free account!
Rosegarden-devel mailing list
Rosegarden-devel@lists.sourceforge.net - use the link below to unsubscribe