[jazzplusplus-devel] Expanding concept of 'default' MIDI test-kit

[D.B. M. <db...@ho...>]

Greetings,

If you've been reading my doc updates, you already
know this isn't 'set in stone'. If you read my bio posting,
you'll recall I mention having a QA/QC cert. 

As best I can figure it, people who experience trouble
using jazz++ and that trouble is related to linux kernel
latencies, such a problem can VERY much depend on
-their- actual setup.

The highlight of unknown variables include how fast
their CPU is, how much ram they have (not because
you need lots of ram, but you need adequate ram to
carry the system footprint without it going out to
swap), how many (other) applications they have
running, how complex the .mid file they're replaying
actually is, whether or not the structure of that file
is valid and correct...and in the softsynth & MIDI
synth soundcard case, just what soundfont is used.

IF their CPU coller is choked with dust, this could
well impact realtime performance as well...so when
I was typing up the bugreports content, it occurred
to me that we can at least eliminate some variables
here, and at the same time facilitate pinponting
what might be going on at the user's end.

What I'm thinking of looks something like this; 

Jazz++ is a constant
The included jazz.mid file is the file constant
The fluidsynth/qsynth is the softsynth constant
The jack-connection-kit is the current constant

What we need now, is a soundfont constant.

Once we have that, we can more or less
be sure that using jazz++ to play jazz.mid using
jack/fluidsynth/qsynth as output with a specific
soundfont file loaded is going to represent the
same -system- load in all cases. Close anyhow.


What do we think?


Regards,

Donald B
_________________________________________________________________
Never miss another e-mail with Hotmail on your mobile.
http://www.livelife.ninemsn.com.au/article.aspx?id=343869