On 01/08/2014 12:28, jeff millar wrote:
Hi Bill...
Hi Jeff,
Thanks for the ideas. I'm off to vacation for a week and will not
have access to the desktop or radio for that time. I will try your
debugging build next weekend.
OK, enjoy the break, we can pick this up later.
The input level is reported and control with pulse audio volume
control or the pull down audio control on the top bar. I'm watching
the slider position for Input Audio and stepping the code. one more
click on "step into" causes the input level go from about 20% to 100%.
The audio controls do work on my VM so I will see if I can reproduce
that behaviour.
That line of code is emitting a Qt signal and in this case it is
crossing a thread boundary so it is not a simple function call that is
going on. QtCreator may not follow that signal to any connected slots in
the target thread(s), I don't use QtCreator for debugging so I'm not
100% sure on the mechanics. If I can reproduce it; I will set break
points on the target slots and see if I can track down what is happening
here.
If it were the signal emission in the MainWindow constructor; then it is
quite possible that the previous line is the one that is causing the
issue you see. The previous line is emitting another signal which is
starting the audio input stream, again via an inter-thread signal to
slot connection. It is quite possible that the target thread doesn't run
and dequeue the message until the line you cite runs. You can trace this
in debug by setting a break point on SoundInput::start which is the
target slot of the MainWindow::startAudioInputStream signal. Then step
through that to see what line causes the level reset.
This probably boils down to the fact that Qt creates a new stream (and
stream name) for every audio connection and the we have no control of
the stream name. This means that levels set via the system or
pavucontrol are not remembered between runs. You will probably have to
set the RX level via the WSJT-X level slider rather than the system.
IIRC the slider in WSJT-X doesn't change the stream slider, we do our
own gain processing on the audio received. I would think that given two
different digital level controls (pulseaudio and ours) in series; it is
probably best to have one or the other at 0dB anyway as each stage of
attenuation is likely to introduce artefacts so only using one of them
is best.
We have raised an issue report to Qt on this but AFAIK no one has picked
it up and submitted a fix. On Windows since 7 the levels are associated
with the executable name that set them. In pulseaudio the levels are
associated with the stream name so not being able to set the stream name
is IMHO a Qt defect.
jeff, wa1hco
Hope this helps & 73
Bill
G4WJS.
On 08/01/2014 05:10 AM, Bill Somerville wrote:
On 01/08/2014 04:42, jeff millar wrote:
Hi Jeff,
Sorry you are having some issues with the current WSJT-X code.
First problem: Most of the time, when starting a transmission or
using the Tune button, WSJTX throws a "Rig Control Error Hamlib
error: Invalid parameter while getting current VFO frequency". This
occurs about 1 sec after starting a transmission. But sometimes it
runs ok for a while without throwing errors. The Test CAT button
doesn't show a problem. I can browse the bands with the band select
button without seeing the problem.
So I can see what is going on here can you build a version that has
CAT diagnostics enabled. Details of a way to do that are in this posting:
https://sourceforge.net/p/wsjt/mailman/message/32650244/
Once built; run the application to reproduce the issues then close
it. The file wsjtx_trace.log is created in the directory that the
binary was in. Compress and send me that file (directly is probably
best) and I'll have a look at what is going on.
Second problem: Whenever WSJTX starts up it somehow resets the input
level to 100%. I traced this down to this line in mainwindow.cpp
Q_EMIT initializeAudioOutputStream (m_config.audio_output_device (),
AudioDevice::Mono == m_config.audio_output_channel () ? 1 : 2,
m_msAudioOutputBuffered);
any ideas?
That line is the initialisation of the audio output stream, that has
to be done before it can be used. That should only effect TX audio so
I am surprised that it is having any impact on RX audio.
Unfortunately I only have Linux on a VM and the audio drivers for the
VM do not support input audio so I cannot debug this issue here.
When you say "resets input level to 100%" where are you seeing that?
jeff, wa1hco
I'm running WSJTX, r1.4, r4224 on Ubuntu 14.04. Hamlib is V3.0.
IC-746PRO. PTT method by RTS.
Application output from qtcreator...
ig_set_conf: serial_handshake='None'
rig_set_conf: ptt_pathname='/dev/ttyUSB1'
rig_set_conf: ptt_type='RTS'
rig:rig_open called
read_string(): Timed out 0.200237 seconds without reading a
character.
read_string(): Timed out 0.200235 seconds without reading a
character.
read_string(): Timed out 0.200244 seconds without reading a
character.
read_string(): Timed out 0.200218 seconds without reading a
character.
rig:rig_close called
rig:rig_close called
rig:rig_cleanup called
QDialog::exec: Recursive call detected
Got a buffer underflow!
Got a buffer underflow!
Got a buffer underflow!
Got a buffer underflow!
The only issue I see with that output is the read_string() time outs
which implies that the CAT module is not getting any response from
the rig from a command sent to it at some point, the diagnostic trace
should help with that too.
73
Bill
G4WJS.
------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
