22 Nov
2004
22 Nov
'04
10:57 p.m.
Hi,
I released fmit (free music instrument tuner) v 0.9.0
JACK is now supported
https://gna.org/projects/fmit/
I'm looking for testers to be sure to release a bug-free 1.0 version :)
Enjoy,
Gilles
22 Nov
22 Nov
11:56 p.m.
Gilles Degottex <gilles.degottex(a)net2000.ch> writes:
I'm looking for testers to be sure to release a
bug-free 1.0 version :)
The INSTALL instructions are not very clear. Trying to follow them,
it doesn't build on my Debian woody system...
$ make check
make: *** No rule to make target `check'. Stop.
$ make prefix=/INSTALL/fmit configure
make: *** No rule to make target `configure'. Stop.
--
joq
23 Nov
23 Nov
12:08 a.m.
On Mon, 2004-11-22 at 21:56 -0600, Jack O'Quin wrote:
Gilles Degottex <gilles.degottex(a)net2000.ch>
writes:
For your 1.0 release, you should consider it a bug if configure; make &&
make install does not work. Your users should not have to read an
INSTALL file.
Lee
I'm looking for testers to be sure to release
a bug-free 1.0 version :)
The INSTALL instructions are not very clear. Trying to follow them,
it doesn't build on my Debian woody system...
8:28 a.m.
On Tuesday 23 November 2004 05:08, Lee Revell wrote:
On Mon, 2004-11-22 at 21:56 -0600, Jack O'Quin
wrote:
okay okay, I'll do it.
let's dive in the autotools dream for today .........
Gilles Degottex
<gilles.degottex(a)net2000.ch> writes:
For your 1.0 release, you should consider it a bug if configure; make &&
make install does not work. Your users should not have to read an
INSTALL file. I'm looking for testers to be sure to release
a bug-free 1.0 version :)
The INSTALL instructions are not very clear. Trying to follow them,
it doesn't build on my Debian woody system...
3 p.m.
On Tue, 2004-11-23 at 13:33 +0100, Gilles Degottex wrote:
On Tuesday 23 November 2004 05:08, Lee Revell wrote:
Read the above command carefully. It does not require you to use
autotools. But, if you don't then "make && make install" damn well
better work.
--
Lee Revell <rlrevell(a)joe-job.com>
For your 1.0 release, you should consider it a
bug if configure; make &&
make install does not work. Your users should not have to read an
INSTALL file.
okay okay, I'll do it.
let's dive in the autotools dream for today .........
8:24 a.m.
On Tuesday 23 November 2004 04:56, Jack O'Quin wrote:
Gilles Degottex <gilles.degottex(a)net2000.ch>
writes:
*ouch*, I cant beleve I forgot this.
fixed now in:
http://download.gna.org/fmit/fmit-0.9.1.tar.bz2
I'm looking for testers to be sure to release
a bug-free 1.0 version :)
The INSTALL instructions are not very clear. Trying to follow them,
it doesn't build on my Debian woody system...
$ make check
make: *** No rule to make target `check'. Stop.
$ make prefix=/INSTALL/fmit configure
make: *** No rule to make target `configure'. Stop.
2:20 p.m.
Gilles Degottex <gilles.degottex(a)net2000.ch> writes:
On Tuesday 23 November 2004 04:56, Jack O'Quin
wrote:
That's better, but that YAMakefile stuff doesn't work yet. What does
that "ccgnu2-config script not found" message indicate?
using Qt library with thread support (Qt directory: /usr/share/qt3)
checking Qt user interface compiler: ... ok ( /usr/share/qt3/bin/uic )
checking Qt meta object support code generator: ... ok ( /usr/share/qt3/bin/moc )
checking for requested libraries ...
linker libraries paths: -L/usr/X11R6/lib -L/usr/share/qt3/lib
linker options:
checking Xft ... ok (/usr/X11R6/lib/libXft.so)
checking X11 ... ok (/usr/X11R6/lib/libX11.so)
checking freetype ... ok (/usr/bin/../lib/libfreetype.so)
checking z ... ok (/usr/bin/../lib/libz.so)
checking fontconfig ... ok (/usr/bin/../lib/libfontconfig.so)
checking Xrender ... ok (/usr/bin/../lib/libXrender.so)
checking jack ... ok (/usr/local/lib/libjack.so)
checking GL ... ok (/usr/X11R6/lib/libGL.so)
checking GLU ... ok (/usr/X11R6/lib/libGLU.so)
checking glut ... ok (/usr/bin/../lib/libglut.so)
checking qt-mt ... ok (/usr/share/qt3/lib/libqt-mt.so)
[31mError[0m appears while checking:
requested ccgnu2-config script not found
generating config header file config.h (requested by )
[joq@sulphur] fmit-0.9.1/ $ make
Qt: generate header ConfigForm.h for interface ConfigForm.ui
uic: File generated with too recent version of Qt Designer (3.3 vs. 3.1.2)
make: *** [ConfigForm.h] Error 1
That last message looks like the typical problem with Qt designer
always writing files that claim to depend on the latest version of the
library (whether it does or not). If it really does depend on that
version, you need to check for it at configure time. If not, you can
probably edit those files to lower the claimed requirement.
--
joq
Gilles Degottex
<gilles.degottex(a)net2000.ch> writes:
*ouch*, I cant beleve I forgot this.
fixed now in:
http://download.gna.org/fmit/fmit-0.9.1.tar.bz2 I'm looking for testers to be sure to release
a bug-free 1.0 version :)
The INSTALL instructions are not very clear. Trying to follow them,
it doesn't build on my Debian woody system...
$ make check
make: *** No rule to make target `check'. Stop.
$ make prefix=/INSTALL/fmit configure
make: *** No rule to make target `configure'. Stop. 
6:28 p.m.
<snip>
That's better, but that YAMakefile stuff
doesn't work yet. What does
that "ccgnu2-config script not found" message indicate?
<snip>
it means you don't have the gnu Commons c++ 2 package installed (figured
that out after some apt-cache searching around). A list of dependencies
in the README/INSTALL would be nice... After that, and some fiddling
with the QTDIR environment variable it built fine here (and works nicely
too!)
Cheers,
Walco
10:17 p.m.
Walco <walco-linux(a)n--tree.net> writes:
<snip>
What is that DEB package called? I seem to need it. At one time I
must have had a libcommonc++-dev, but it's not installed and nothing
like that shows up in my Debian search patch.
--
joq
That's better, but that YAMakefile stuff
doesn't work yet. What does
that "ccgnu2-config script not found" message indicate?
<snip>
it means you don't have the gnu Commons c++ 2 package installed
(figured that out after some apt-cache searching around). A list of
dependencies in the README/INSTALL would be nice... After that, and
some fiddling with the QTDIR environment variable it built fine here
(and works nicely too!)
10:27 p.m.
On Wednesday 24 November 2004 03:16, Jack O'Quin wrote:
Walco <walco-linux(a)n--tree.net> writes:
like seen on the fmit website: http://home.gna.org/fmit/doc.html
for Debian: libcommoncpp2-dev libqt3-mt-dev xlibmesa-gl-dev xlibmesa-glu-dev
freeglut3-dev libasound2-dev
but dont lose your time with Common C++, I removed this depedency :P
a release with theses changes should come in a few minutes.
<snip>
What is that DEB package called? I seem to need it. At one time I
must have had a libcommonc++-dev, but it's not installed and nothing
like that shows up in my Debian search patch. That's better, but that YAMakefile stuff
doesn't work yet. What does
that "ccgnu2-config script not found" message indicate?
<snip>
it means you don't have the gnu Commons c++ 2 package installed
(figured that out after some apt-cache searching around). A list of
dependencies in the README/INSTALL would be nice... After that, and
some fiddling with the QTDIR environment variable it built fine here
(and works nicely too!) 
10:48 p.m.
On Tue, 2004-11-23 at 21:31, Gilles Degottex wrote:
but dont lose your time with Common C++, I removed
this depedency :P
a release with theses changes should come in a few minutes.
Cool. Have Mandrake spec file waiting. Will upload new version tonight
to the mirrors.
Any chance of making the binary both jack and alsa compatible in the
future?
Austin
11:04 p.m.
On Wednesday 24 November 2004 03:50, Austin wrote:
On Tue, 2004-11-23 at 21:31, Gilles Degottex wrote:
thx :)
but dont lose your time with Common C++, I
removed this depedency :P
a release with theses changes should come in a few minutes.
Cool. Have Mandrake spec file waiting. Will upload new version tonight
to the mirrors. Any chance of making the binary both jack and alsa
compatible in the
future?
what you exactly mean by compatible ?
merged in only one prog with an option in the settings to choose between
them ?
11:17 p.m.
On Wednesday 24 November 2004 03:50, Austin wrote:
On Tue, 2004-11-23 at 21:31, Gilles Degottex wrote:
0.9.2 released !
http://download.gna.org/fmit/
Enjoy,
Gilles
but dont lose your time with Common C++, I
removed this depedency :P
a release with theses changes should come in a few minutes.
Cool. Have Mandrake spec file waiting. Will upload new version tonight
to the mirrors.
8:25 p.m.
On Tuesday 23 November 2004 19:19, Jack O'Quin wrote:
[joq@sulphur] fmit-0.9.1/ $ make
Qt: generate header ConfigForm.h for interface ConfigForm.ui
uic: File generated with too recent version of Qt Designer (3.3 vs. 3.1.2)
make: *** [ConfigForm.h] Error 1
That last message looks like the typical problem with Qt designer
always writing files that claim to depend on the latest version of the
library (whether it does or not). If it really does depend on that
version, you need to check for it at configure time. If not, you can
probably edit those files to lower the claimed requirement.
but I got no clue it depend on which minimal version ??
and I got no system with 15 installed version of Qt :P
could I ask you to modify it and try ?
for sure, it will not work on version lower than 3.0, but to be more
precise ...
thanks for your interest,
Gilles
9:23 a.m.
On Tuesday 23 November 2004 11:43, Wolfgang Woehl wrote:
* http://download.gna.org/fmit/fmit-0.9.0.tar.bz2
seems to be missing
"Makefile.unroll" so there are no make targets. Can't configure or
build it.
fixed now in http://download.gna.org/fmit/fmit-0.9.1.tar.bz2
I'll switch to "auto"tools as fast as I can ..... in the joy and the good
mood .......
* Isn't "Capture system" actually a
debugging element? I think there is
no real need for it from a user's/tuner's point of view.
setting,
connecting and mixing to get capture working is not always easy for
everybody. sometimes sound servers are not working or simply not launched,
alsa services are not started, modules not loaded or what ever and the user
stay behind a blank window and dont know what he should do. This element is
just there to say: "everything is ready, just play your instrument now".
It's
like the small led in the corner of your handheld tuner which light on when
you push the power button.
I'll keep this information visible, but not important as it is, okay.
The levelmeter
above it is about 6 pixels high on my box. It's yellow and
therefore a bit hard to see.
that's the volume meter, it's not really
usefull, that's why I did a small
one :) but I raised it now.
* Maybe think about changing the name "Error
history". If it wasn't for
talking about it it wouldn't even need a name because you immediately
grasp the concept when you see it in action. "Error history" sounds so
-- hm, oppressive to the poor tuner's ear who's trying really hard to
master the art of tuning :)
lol, I dont usualy speak english, that's why some
"translations" comes from
nowhere :)
could you suggest me another name ? (maybe it dont need any name, but I want
one to refere it in the settings dialog)
* I'd expect "Pause" to turn off all
rendering/updating in the gui and
processing of audio. After having tuned an instrument I would hit
"Pause" to completely free the CPU from fmit and later come back with
another instrument
hum, already exist and shortcuted to the space bar.
-- fmit being ready with one mouse-click. Instead it
disconnects any jack-connections and unregisters from jack. Unexpected
behaviour I think.
I currently stop then restart the sound capture to simulate a
pause state.
It's usefull when you have to reset the capture system, just hit twice the
pause button.
to avoid register at each unpause, you can add your prefered jack source in
the settings dialog.
Should I concider add a real reset button and do a real pause button ?
* To my taste "Dominant note" could be
bigger
[...]
yes, this part of the interface is not really clear, I'll thing about
it.
ps:
[...]
maybe we should all start to use reply-to-list in order to keep answers
and questions together. People might not be getting it and they'd be
missing a true gem.
done :)
Gilles Degottex <gilles.degottex(a)net2000.ch>ch>:
> Hi,
>
> I released fmit (free music instrument tuner) v 0.9.0
>
> JACK is now supported
>
> https://gna.org/projects/fmit/
>
> I'm looking for testers to be sure to release a bug-free 1.0 version
> :)
>
> Enjoy,
> Gilles
7:56 p.m.
Gilles Degottex <gilles.degottex(a)net2000.ch>ch>:
The whole point is: In a recording session you set up a certain
connection jack <-> fmit. You use fmit -- which has an appetite for cpu
cycles -- and then you want to release the cpu from fmit. So you want
to halt it's operation.
But you don't want to manually setup the connection again. You want to
hit a button and have fmit back connected to where you left it.
"Preferred jack source" is not always applicable, it's session
specific. fmit could remember it's connections and restore them
whatever they were, right?
Something else: Feeding for example 880 hz out of the audio analyzer
"jaaa" (http://users.skynet.be/solaris/linuxaudio) into fmit the
measured frequency in fmit says ~886 hz. The sine jaaa generates has
measurable harmonics. Do they account for the error?
Wolfgang
On Tuesday 23 November 2004 11:43, Wolfgang Woehl
wrote:
-- fmit being
ready with one mouse-click. Instead it
disconnects any jack-connections and unregisters from jack.
Unexpected behaviour I think.
I currently stop then restart the sound capture to simulate a pause
state. It's usefull when you have to reset the capture system, just
hit twice the pause button.
to avoid register at each unpause, you can add your prefered jack
source in the settings dialog.
Should I concider add a real reset button and do a real pause button
?
8:42 p.m.
On Wednesday 24 November 2004 00:55, Wolfgang Woehl wrote:
Gilles Degottex <gilles.degottex(a)net2000.ch>ch>:
The whole point is: In a recording session you set up a certain
connection jack <-> fmit. You use fmit -- which has an appetite for cpu
cycles -- and then you want to release the cpu from fmit. So you want
to halt it's operation.
But you don't want to manually setup the connection again. You want to
hit a button and have fmit back connected to where you left it.
"Preferred jack source" is not always applicable, it's session
specific. fmit could remember it's connections and restore them
whatever they were, right?
In a fast research, JACK does not seems to support a
"suspend" state or
something like this, but I'll simulate one. There will be always a little bit
CPU usage, just enough to raise the jack process function and throw away the
data.
Gilles
On Tuesday 23 November 2004 11:43, Wolfgang Woehl
wrote:
-- fmit
being ready with one mouse-click. Instead it
disconnects any jack-connections and unregisters from jack.
Unexpected behaviour I think.
I currently stop then restart the sound capture to simulate a pause
state. It's usefull when you have to reset the capture system, just
hit twice the pause button.
to avoid register at each unpause, you can add your prefered jack
source in the settings dialog.
Should I concider add a real reset button and do a real pause button
?
9:29 p.m.
Gilles Degottex <gilles.degottex(a)net2000.ch> writes:
> The whole point is: In a recording session you
set up a certain
> connection jack <-> fmit. You use fmit -- which has an appetite for cpu
> cycles -- and then you want to release the cpu from fmit. So you want
> to halt it's operation.
>
> But you don't want to manually setup the connection again. You want to
> hit a button and have fmit back connected to where you left it.
> "Preferred jack source" is not always applicable, it's session
> specific. fmit could remember it's connections and restore them
> whatever they were, right?
In a fast research, JACK does not seems to support a
"suspend" state or
something like this, but I'll simulate one. There will be always a little bit
CPU usage, just enough to raise the jack process function and throw away the
data.
You could use jack_deactivate(), but then you'd have to restore your
port connections when it's time to jack_activate() again.
http://jackit.sourceforge.net/docs/reference/html/jack_8h.html#a19
Probably easier and good enough just to return immediately from
process() when "suspended".
--
joq
9:52 p.m.
On Wednesday 24 November 2004 02:27, Jack O'Quin wrote:
Yes, exactly what
Wolfgang wont :) (to restore manualy the connection)
Gilles Degottex <gilles.degottex(a)net2000.ch>
writes:
bit
> The whole point is: In a recording session
you set up a certain
> connection jack <-> fmit. You use fmit -- which has an appetite for cpu
> cycles -- and then you want to release the cpu from fmit. So you want
> to halt it's operation.
>
> But you don't want to manually setup the connection again. You want to
> hit a button and have fmit back connected to where you left it.
> "Preferred jack source" is not always applicable, it's session
> specific. fmit could remember it's connections and restore them
> whatever they were, right?
> In a fast research, JACK does not seems to support a "suspend" state or
> something like this, but I'll simulate one. There will be always a little
> CPU usage, just enough to raise the jack process
function and throw away
the
data.
You could use jack_deactivate(), but then you'd have to restore your
port connections when it's time to jack_activate() again.
10:07 p.m.
On Wednesday 24 November 2004 02:27, Jack O'Quin
wrote:
> You could use jack_deactivate(), but then you'd have to restore your
> port connections when it's time to jack_activate() again.
Gilles Degottex <gilles.degottex(a)net2000.ch> writes:
Yes, exactly what Wolfgang wont :) (to restore manualy
the connection)
I didn't say "manually". It could be done automatically by the program.
--
joq
10:20 p.m.
On Wednesday 24 November 2004 03:06, Jack O'Quin wrote:
there is already a such mechanism in the program
but it's not sufficient if you choose to change the connection just for a
session.
On Wednesday
24 November 2004 02:27, Jack O'Quin wrote:
> You could use jack_deactivate(), but then you'd have to restore your
> port connections when it's time to jack_activate() again.
Gilles Degottex <gilles.degottex(a)net2000.ch> writes:
Yes, exactly what Wolfgang wont :) (to restore
manualy the connection)
I didn't say "manually". It could be done automatically by the program.
11:19 p.m.
On Wednesday 24 November 2004 00:55, Wolfgang Woehl wrote:
Something else: Feeding for example 880 hz out of the
audio analyzer
"jaaa" (http://users.skynet.be/solaris/linuxaudio) into fmit the
measured frequency in fmit says ~886 hz. The sine jaaa generates has
measurable harmonics. Do they account for the error?
very interesting, I got nearly the same error with a tone generated by
Audacity. I'll track it as first priority.
Before the 1.0, I have to test fmit with a pure extern generator and compare
it with a handheld tuner, too.
24 Nov
24 Nov
3:23 a.m.
New subject: [linux-audio-user] Re: Instrument Tuner: fmit 0.9.0
Gilles Degottex <gilles.degottex(a)net2000.ch> writes:
Before the 1.0, I have to test fmit with a pure extern
generator and compare
it with a handheld tuner, too.
I have compared a couple of handheld tuners and must tell you that they all
showed different results when measuring the same note...
Robert Epprecht
23 Nov
23 Nov
2:03 p.m.
>>>> "Gilles" == Gilles Degottex
<gilles.degottex(a)net2000.ch> writes:
Gilles> I released fmit (free music instrument tuner) v 0.9.0
Gilles> JACK is now supported
Gilles> https://gna.org/projects/fmit/
Gilles> I'm looking for testers to be sure to release a bug-free 1.0 version
:)
I was able to build it after I figured out where QTDIR was. I agree
with whoever it was that said "./configure;make;make install" is a
better build process.
When I run it, it says:
Free Music Instrument Tuner version 0.9.1 built at Nov 23 2004 12:20:00
CaptureThread: ERROR: JACK: cannot create client and so cannot get server sampling rate
CaptureThread: ERROR: JACK: cannot create client, jack deamon is running ?
CaptureThread: ERROR: JACK: cannot create client, jack deamon is running ?
This is whether I have the jack daemon running or not.
I have unmuted my mic, so I would think it should be seeing something
but it isn't.
--
Laura (mailto:lconrad@laymusic.org , http://www.laymusic.org/ )
(617) 661-8097 fax: (501) 641-5011
233 Broadway, Cambridge, MA 02139
7:21 p.m.
On Tuesday 23 November 2004 19:03, Laura Conrad wrote:
I was able to build it after I figured out where QTDIR
was. I agree
with whoever it was that said "./configure;make;make install" is a
better build process.
I'll do a cleaner "make;make install"
When I run it, it says:
Free Music Instrument Tuner version 0.9.1 built at Nov 23 2004 12:20:00
CaptureThread: ERROR: JACK: cannot create client and so cannot get server
sampling
rate
CaptureThread: ERROR: JACK: cannot create client, jack
deamon is running ?
CaptureThread: ERROR: JACK: cannot create client, jack deamon is running ?
This is whether I have the jack daemon running or not.
?!?!? is your jack server
too old ??
my jack version is 0.94.0
versions are visible on:
http://home.gna.org/fmit/doc.html
7333
days inactive
7335
days old
linux-audio-user@lists.linuxaudio.org
24 comments
8 participants
participants (8)
-
Austin -
Gilles Degottex -
Jack O'Quin -
Laura Conrad -
Lee Revell -
Robert Epprecht -
Walco -
Wolfgang Woehl