Re: [LAU] bristol synths hang

From: Nick Copeland <nickycopeland@email-addr-hidden>
Date: Sun Dec 12 2010 - 11:15:21 EET

In release 0.40 bristol would not configure Jack MIDI interface when you
used the -jack option. I had not tested this interface that much so when you
requested -jack then bristol would still default to ALSA MIDI. Along the way
as more testing was done I took the obvious choice to default audio and MIDI
to Jack when you give the -jack flag.

That means bristol will now appear in the Jack MIDI tab of qjackctl, not in the
ALSA tab. Here is how you work around that:

startBristol -jack -midi alsa -mini

That will give you Jack audio and ALSA MIDI connections. If you want this to
be the default behaviour then you can edit your bristolrc and put in one line:

vi $HOME/.bristol/bristolrc

-jack -midi alsa

In my opinion, if you are using bristol 0.60.7 with jack drivers you should
also use the following options:

-jackstats -count <period size> -rate <samplerate>

What this means it that you tell bristol, a priori, what buffer sizes and sample
rates it should use to be consistent with Jack. If you leave out these options
then bristol will attempt to find the correct rates using a poll to the daemon.
It works 99.999% of the time, but when it doesn't then things get a bit ugly:

-jack -jackstats -count 256 -rate 48000 -jack -midi alsa

This will add these options every time you start bristol. Somebody recently
commented that the bristol command line options are a real inspiration
killer. No shit! Either bury these options once in the bristolrc or try using
monoBristol to launch the app, it will hide a lot of this stuff.

Let me know how it goes with the sticky notes and clicks/pops. In 0.60 there
is one known case where they still happen: monophonics without note
precedence (-nnp) however you have to ask for that to be configured, it is
not a default. The next release is intended to have a fix for this last issue.

As I said before, I can't rule out you still having the issue, there may be
other cases I have not considered but either send reports on this list or
send me an email and I will look into them.

In one of your other email you asked about Jack Session Management:

> | Build with JACK Session support ................ : false
>
> What does this mean?

Bristol has an interface to the Jack Session Manager (JSM) so that you can
save application states and recover them in their entirety. I only include
the drivers in bristol if I find the correct Jack header files. The Jack dev
team will have to comment on which version that would be however last
time I checked it was not yet in general release. I would say that the
header files are not in the jack version you have installed but it shouldn't
give you any problems (unless you were going to use JSM of course).

Kind regards, nick.

"we have to make sure the old choice [Windows] doesn't disappear”.
Jim Wong, president of IT products, Acer

Date: Sat, 11 Dec 2010 22:59:49 +0000
From: f.rech@email-addr-hidden
Subject: RE: [LAU] bristol synths hang
To: rennabh@email-addr-hidden; nickycopeland@email-addr-hidden
CC: linux-audio-user@email-addr-hidden

--- En date de : Sam 11.12.10, Nick Copeland <nickycopeland@email-addr-hidden> a écrit :

De: Nick Copeland <nickycopeland@email-addr-hidden>
Objet: RE: [LAU] bristol synths hang
À: f.rech@email-addr-hidden, rennabh@email-addr-hidden
Cc: linux-audio-user@email-addr-hidden
Date: Samedi 11 décembre 2010, 21h43

> > particular note "hangs" on the synth, and sending note off events
> > doesn't stop it... I have to restart the program, which makes it
> > unusable since it happens quite often. Untill now this happenede with
> > b3, minimoog and dx7.

> I have the same problem here, playing with an usb keyboard.
> Sometimes you probably have "clicks" (digital saturation) too.

Bristol has had these issues although they were supposed to have been
iron out in the later releases. The cause of the clicks was largely due to
voice reassignment with mono synths fixed by implementing a different
note assignment method (note preference with droning) and the sticky
notes were due to scheduling issues between threads fixed with a jack
ringbuffer message passing method.

Now a lot of the distributions are still packaging bristol-0.40 which is
a couple of years out of date. If you have newer versions
 then I can
work on some debuging with you, if you don't have 0.60 then it would
make sense to try that first.

I cannot rule out other issues but lets start with the versions you have
and take it from there.

Kind regards, nick.

Have compile 0.60 with no problem, and no clicks anymore :)
But now Bristol lives -alone- in Qjacks'MIDI tab, and the rest of MIDI stuff are in ALSA tab...
What did I missed ?
                                               

                                               

_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@email-addr-hidden
http://lists.linuxaudio.org/listinfo/linux-audio-user
Received on Sun Dec 12 12:15:02 2010

This archive was generated by hypermail 2.1.8 : Sun Dec 12 2010 - 12:15:02 EET