Linrad r1082

[Ramiro Aceves]

Hello Leif and Linraders,


I am having trouble with Linrad r1082 (and before). I am using Afedri
SDR-Net receiver as input and USB M Audio Fast Trak Pro sound card for
output directly, not using Portaudio. Operating system is Debian
GNU/Linux Stable.

I get random freezes and audio stops working. Also "X" does not work for
scaping, only "Esc" twice works. Sometimes it exits alone.


ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 388421
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
lir error 966572Abortado

ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 388421

lir error 966572ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 776123
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

lir error 966572Abortado
ramiro@debian-nuc8i7:~/linrad-alfredo$

ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 388421
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
lir error 966572Abortado

ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 776123
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
lir error 966572Abortado


ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 776123
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
lir error 966572Abortado


I can reproduce the problem like this:

./xlinrad64

"D" for SSB

left click on main graph. Audio stars working fine.

press the arrows up or down to change Afedri Gain (or change frequency,
it does not matter).

left click again o main graph.

right clic again o main graph.


Doing the same sequence using RTL dongle as input device does not
trigger the error.


Thanks

Ramiro.

[Leif Asbrink]

Hello Ramiro,

The error 388421 is because the thread blocking_rxout did not respond in time.
I have allowed much more time and put the error in errors.lir

The error 966572 isbecause of an error in errors.lir which I have corrected.

I have uploaded r1083, but I forgot to change in vernr.h so it still presents
itself as r1082.

Maybe this helps, but at least it should point more clearly tothe problem
because 388421 could be caused by two different wait loops. Now they have
separate error codes.

73

Leif

> --
> There is an excellent Linrad User Guide by Gaetan, ON4KHG, at:
> http://w3sz.com/Linrad%20Installation%20&%20Configuration%20User%20Guide%20-%20V1-0.pdf
> ---
> You received this message because you are subscribed to the Google Groups "Linrad" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to linrad+un...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/linrad/0cdb9962-1906-41c4-bb61-8c6a07c00e7e%40gmail.com.

[Oleg]

Hello Leif!

Yesterday I compiled version 1082 from the repository for Windows 64 bit.
But the autorun function does not work in it.
I usually start automatic reception in SSB mode.
After a couple of seconds the error appears: Internal error 116711
and Linrad closes.
Manual start works fine.

Best regards,
 Oleg.
 oleg.r...@gmail.com

нд, 15 груд. 2024 р. о 12:30 Leif Asbrink <le...@sm5bsz.com> пише:

To view this discussion visit https://groups.google.com/d/msgid/linrad/20241215112956.e452be34a164ff7ffcdcfe7c%40sm5bsz.com.

[Ramiro Aceves]

Hello Leif,

Thanks for code change.

I compiled r1083:

ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 776123

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Abortado

ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 776123

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Abortado

ramiro@debian-nuc8i7:~/linrad-alfredo$ ./xlinrad64

lir error 776123

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Abortado


Regards.
Ramiro.



El 15/12/24 a las 11:29, Leif Asbrink escribió:

[Leif Asbrink]

Hello Oleg,

The autorun function is now fixed. r1084 on the repo.

73

Leif

[Leif Asbrink]

Hello Ramiro,

The blocking rxout thread did not respond to the idle command.
I made a small change that I do not think really will help,
but please try r1084 to see if you still get th error.
Now hopefully with a message.

If the problem is still there I will send a debug version of
rxout.c to run with #define DUMPFILE TRUE in options.h

If the problem is still present, please try with portaudio.

73

Leif

[Ramiro Aceves]

Hello Leif,

Many thanks for trying to fix the problem. Sadly the problem is not
fixed as you guessed:

ramiro@debian-x260:~/linrad$ ./xlinrad64

lir error 1074

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Abortado

ramiro@debian-x260:~/linrad$ ./xlinrad64

lir error 1074

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Abortado

ramiro@debian-x260:~/linrad$


Forgot to tell you the behaviour using Portaudio: Doing the same steps
that trigger the error when not using Portaudio, audio no longer works.
Waterfalls continue working but no audio.  "X" does not work for exit. I
must press "Esc" to exit. No error number is shown:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Abortado

ramiro@debian-x260:~/linrad$

Thanks

Ramiro.







El 16/12/24 a las 14:38, Leif Asbrink escribió:

[Leif Asbrink]

Hello Ramiro,

Can you please verify that old Linrad, latest released
version http://sm5bsz.com/linuxdsp/linrad.htm works properly.
Something could have gone wrong with drive routines.
It is unexpected to me that you have problems with and without
portaudio. The software is very different, the code for
portaudio does not depend on what soundcard is in use.

Do you get any sound output before the 1074 error occurs
when you run native ALSA?

73

Leif

> --
> There is an excellent Linrad User Guide by Gaetan, ON4KHG, at:
> http://w3sz.com/Linrad%20Installation%20&%20Configuration%20User%20Guide%20-%20V1-0.pdf
> ---
> You received this message because you are subscribed to the Google Groups "Linrad" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to linrad+un...@googlegroups.com.

> To view this discussion visit https://groups.google.com/d/msgid/linrad/0bc9d824-afcc-4b4e-8ece-3e87238d4024%40gmail.com.

[Ramiro Aceves]

Hello Leif, thanks for answering and for all your suggestions. Please
read below:

El 17/12/24 a las 17:47, Leif Asbrink escribió:
> Hello Ramiro,
>

> Can you please verify that old Linrad, latest released
> version http://sm5bsz.com/linuxdsp/linrad.htm works properly.
> Something could have gone wrong with drive routines.
> It is unexpected to me that you have problems with and without
> portaudio. The software is very different, the code for
> portaudio does not depend on what soundcard is in use.

I have installed 5.02 and cannot trigger the error (in both native ALSA
and pulseaudio). Linrad sounds well. The only thing it does not work is
pressing "X", it does not exit. After pressing "X" sound never works
when clicking on the waterfall. Then I must press "Esc" to go out. After
pressing "Esc" the following is written on the console:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99
100 Abortado
ramiro@debian-x260:~/linrad-05.02$

>
> Do you get any sound output before the 1074 error occurs
> when you run native ALSA?

In native Alsa Linrad r1078: I trigger the error, sound stops and after
5 seconds "lir error 1074" appears in the terminal console.

Thanks.


73, Ramiro.

[Leif Asbrink]

Hello Ramiro,

This is unclear to me:

> In native Alsa Linrad r1078: I trigger the error, sound stops and after
> 5 seconds "lir error 1074" appears in the terminal console.

Does it mean that linrad runs with spectrum and waterfall until you
click on a signal and that you then get sound for about 5 seconds
after which sound stops and the error message appears?

73

Leif

[Oleg]

Hello Leif.

The autorun feature in release 1084 works well.

Thank you very much!

Best regards,
 Oleg.
 oleg.r...@gmail.com

пн, 16 груд. 2024 р. о 15:29 Leif Asbrink <le...@sm5bsz.com> пише:

--
There is an excellent Linrad User Guide by Gaetan, ON4KHG, at:
http://w3sz.com/Linrad%20Installation%20&%20Configuration%20User%20Guide%20-%20V1-0.pdf
---
You received this message because you are subscribed to the Google Groups "Linrad" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linrad+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/linrad/20241216142944.44b2ddac98c8fd5ae40b8dd5%40sm5bsz.com.

[Leif Asbrink]

Hello Ramiro,

I was able to configure my NET-Afedri, the culpit was to unpower the unit
before starting the control program again.

I run on a somewhat elderly computer and I se the errors misbehave more clearly.
I have tried older versions back to 04.00 and none works properly.

The comand to set frequency is repeated continously, the flag that should
set that it has been done does not work. The code involves the function:

INLINE bool lir_set_thread_flag_if(int32_t threadId, int32_t oldState, int32_t newState)
{
return lir_interlocked_compare_exchange(thread_status_flag + threadId, oldState, newState) == oldState;
}
It uses __sync_val_compare_and_swap under Linux

I do not understand what it does and why it no longer works - not even with
linrad-04.00 on an elderly computer. Maybe it never worked properly.

I did not write this code myself, it was implemented by someone else to solve
a problem when running linrad remotely.

I think one solution would be to issue the frequency control commands from
the same thread that reads the data from the Afedri. Then I could simply
let the input wait for the frequency control to return. Some lossof data
while changing the frequency would not be a big problem.

Maybe someone else with better programming skils can suggest a better solution?

73

Leif

[Ramiro Aceves]

Hi Leif, I have just SVN updated to r1085 and appears to work fine! Have
you found the fix?

73, Ramiro.


El 22/12/24 a las 0:35, Leif Asbrink escribió:

[Leif Asbrink]

Hello all,

The problem with frequency control for the NET-Afedri turned out
to have been in place for a very long time. It is present already
in linrad-04.00 and most probably much earlier.

The problem causes Linrad so send the updated frequency and/or
gain continously after a single mouse click. The effect of
these useless updates becoes different on different computers.
I have never noticed it before Ramiro reported it.

The problem is the same with NET-Afedri and SDR-IP. It should have
been the same for CloudIQ, but since some code was missing the
frequency and gain settings for the Cloud IQ did not work at all.
(On had to use a separate control program for that)

I will wait for error reports about a week and if nothing is reported
I will release linrad-06.00

Please report any old error that is still present.

73

Leif

[Ramiro Aceves]

Superb Leif! I works just fine, both native alsa and pulseaudio (well, I
think it is pipewire now), I will play a bit more just in case, but I
think you have fixed it very well.

Regards!
Ramiro.


El 23/12/24 a las 17:04, Leif Asbrink escribió:

[SP2BPD]

Leif,

In 1085 (Windows) I've noticed  that after hitting F5 (pause processing) I cannot move or resize windows, mark the measurement range in Meter window etc. No reaction to mouse clicks at all.

Just to let you know before 6.0 release 

Piotr

[SP2BPD]

after going back to versions installed on my 2 computers:  this error appeared somewhere between versions 1068 and 1074

Piotr, SP2BPD

[Leif Asbrink]

Hello Piotr,

F5 is primarily for studies of recorded data. It freezes everything
and you have topress it again to resume processing. Whenused in
normal processing, the data correspoding to the frozen time is
lost and it may take some time until processing is norma again.

I will change:
F5 = Pause processing. Useful for wideband recordings on slow computers.

F5 = Pause/Resume processing. Particularly useful for wideband recordings.

Please check that it behaves as expected.

73

Leif

> --
> There is an excellent Linrad User Guide by Gaetan, ON4KHG, at:
> http://w3sz.com/Linrad%20Installation%20&%20Configuration%20User%20Guide%20-%20V1-0.pdf
> ---
> You received this message because you are subscribed to the Google Groups "Linrad" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to linrad+un...@googlegroups.com.

> To view this discussion visit https://groups.google.com/d/msgid/linrad/e6ba5bff-f25c-409f-8c1f-f4b7fb3ffd46n%40googlegroups.com.

[SP2BPD]

Leif,

Yes: F5 Pause/Resume in the 1085 works as it always has. But, when processing is stopped, it is not possible to select the measurement range in the Meter window, because Linrad does not respond to mouse clicks. This is the case both when playing a previously recorded file and when working with a connected receiver. After hitting F5 again, Linrad returns to signal processing (either from the receiver or from the recording) and responds correctly to the mouse.

Piotr

[john....@gmail.com]

Hello Leif,

I do not think this needs to be resolved before release of 6.00.  There are workarounds, such as manually editing parameter files.

When using portaudio under windows (Windows10 here):

For windows version (r1085, both 32 bit and 64 bit windows executables), the issue with changing the parameter "Output sampling speed (Hz)"  with portaudio is still present.  The program crashes soon after you click on the parameter to change output sampling speed.

This is true when using WDM-KS, WASAPI and ASIO sound systems.  I checked portaudio/MME and portaudio/DirectSound (which I never use) and those crash too.

I tried to follow the source code.  I think the problem may be when pa_get_valid_samplerate is called from menu.c

I don't use portaudio on Linux.  I can't tell you if this issue is present on Linux.

73,

John n2nc

[Leif Asbrink]

Hi Piotr,

The mouse was implemented long ago when Linrad was single-threaded.
Timing is controlled by the data flow from the input so there is no
mouse when input is stopped. Changing that would be a major undertaking,
not something I want to try now. Linrad is still far from stable
although I uploaded 1086 ntoday with a couple of corrections.

73

Leif

[Leif Asbrink]

Hello John,

I think I have cured the problem you reported plus some more things.

r1086 is now on the repo. There are still serious errors that I do not
(yet?) understand. Win7 on a desktop seems OK, but on a Win10 laptop
only the 2 bit version works. Probably a drive routin issue, but
I am not sure. The 64 bit linrad-05.02 also does not work on the
Win10 laptop.

Please, all of you, try it out and report new and old problems and errors.

73

Leif

[john....@gmail.com]

Hello Leif,

A short note -- I ran r1086 on both a 64 bit WIN10 desktop and 64 bit WIN10 laptop.  Both 32 bit and 64 bit executables ran OK on both machines.  The "Output sampling speed (Hz)" issue is corrected.

I'll keep testing.

73,

John n2nc

[john....@gmail.com]

Hello Leif,

I am reporting another r1086 issue.

This is regarding the parameter "Default output mode" that saves the initial configuration state of the baseband window.

1)  Whatever value is stored in a file such as par_cw or par_hsms for "Default output mode", when you start the receiver, Linrad starts with a value 3 higher than what is in the file.  For example, if I set "Default output mode [3]", the baseband window shows 6 in the lower right corner when starting.

2)  When you press X, then P to change parameters, the value for "Default output mode" is the current number showing in the lower right corner of the baseband window.  It does not show the current value stored in the parameter file.

73,

John N2NC

[john....@gmail.com]

Hello Leif,

I just noticed more baseband graph issues window with r1086:

Arrows in lower left (expand/compress dB scale):

1) do not update par_xxx_bg file when you click on the up/down arrows

2) the shape of the filter response plots/dots does not change when you change this scale

Arrows on right side of spectrum (move dB scale up/down)

1)  do not update par_xxx_bg file when you click on these arrows.

73,

John n2nc

[Ramiro Aceves]

Hello,

Just a curious thing that at the moment I have not been able to solve is
that in NetBSD operating system "F1" does not work as help key. Only "!"
key works. In FreeBSD operating system "F1" key works fine as in Linux,
no problem at all.

I have tested both XFCE4 and CTWM window managers, just in case F1 key
was captured as a keyboard shortcut.

I will investigate it and ask to the gurus of the operating system.

Thanks.
Ramiro.




El 2/1/25 a las 23:06, john....@gmail.com escribió:

> --
> There is an excellent Linrad User Guide by Gaetan, ON4KHG, at:
> http://w3sz.com/
> Linrad%20Installation%20&%20Configuration%20User%20Guide%20-%20V1-0.pdf

> <http://w3sz.com/
> Linrad%20Installation%20&%20Configuration%20User%20Guide%20-%20V1-0.pdf>

> ---
> You received this message because you are subscribed to the Google
> Groups "Linrad" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to linrad+un...@googlegroups.com

> <mailto:linrad+un...@googlegroups.com>.

> To view this discussion visit https://groups.google.com/d/msgid/

> linrad/9a77693f-4012-471f-857a-830f4cca3051n%40googlegroups.com
> <https://groups.google.com/d/msgid/
> linrad/9a77693f-4012-471f-857a-830f4cca3051n%40googlegroups.com?
> utm_medium=email&utm_source=footer>.

[Leif Asbrink]

Hello Ramiro!

SM5BFK had his problem very long ago. I think it was under Windows and
because of something in the BIOS.

That is why there is the possibillity to use '!' instead.

73

Leif

> To unsubscribe from this group and stop receiving emails from it, send an email to linrad+un...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/linrad/479f1565-85b0-4cd4-8916-6bcc3705efdc%40gmail.com.

[Ramiro Aceves]

El 6/1/25 a las 5:31, Leif Asbrink escribió:

> Hello Ramiro!
>
> SM5BFK had his problem very long ago. I think it was under Windows and
> because of something in the BIOS.
>
> That is why there is the possibillity to use '!' instead.
>
> 73
>
> Leif
>

Thanks, Leif.

It's curious that in FreeBSD and Linux it works fine on the same amd64
computer, but in NetBSD it does not work—neither on amd64 with the Intel
NUC nor on arm64 with the Raspberry Pi.

I asked some NetBSD experts, and they pointed me to the "xev" program to
investigate and capture key presses. They concluded that the F1 key
event is passed correctly to the Linrad program. I initially suspected
that some key shortcuts in the Xorg desktop environment might be
intercepting the keys, but that does not seem to be the case.

Anyway, it doesn't matter at all; we have the alternative "!" key.

Regards,
Ramiro

[Leif Asbrink]

Hello John and all,

r1087 is now uploaded. The "Default output mode" error is corrected.

At the moment I am not aware of any reported but not corrected
errors.

Please test and report any errors.

73

Leif





> Hello Leif,
>

[john....@gmail.com]

Hello Leif,

There were some other issues related to controlling the scale of the spectrum in baseband window and the shape of the filter shown over the spectrum.  I believe these exist in r1087.  See below,

73,

John n2nc

Arrows in lower left (expand/compress dB scale):

1) do not update par_xxx_bg file when you click on the up/down arrows

2) the shape of the filter response plots/dots does not change when you change this scale

Arrows on right side of spectrum (move dB scale up/down)

1)  do not update par_xxx_bg file when you click on these arrows.

[Leif Asbrink]

Hello John,

Thank you:-)

This error was easy, r1088 is now uploaded.

73

Leif

[john....@gmail.com]

Hello Leif,

r1088 looks good.

One last thing I have noticed with the baseband graph:  The red volume control bar is sometimes covered by the numbers for the spectrum db scale.  If you set the level of volume control with the mouse, the numbers look like they did in v5 Linrad.  If you then click on one of the borders of the baseband graph, when the graph is redrawn there are gaps in the volume control red bar where the db scale numbers are located.

73,

John n2nc

[Leif Asbrink]

Hello John and all,

r1089 is now available. The volume control bar is corrected and a
modification in the support for Perseus is introduced to make it
easier to several instances of linrad in separate directories
dedicated to hardware with a particular serial number.

Open linrad, if the error/warning that the serial number is incorrect,
open more instances until the right one is found. Then close the ones
you do not want in that directory.

73

Leif

[Leif Asbrink]

Hi All,

Are there issues that I have forgotten ?
I think I have corrected all reported errors, but I may
have forgotten something.

And, are there still any problems that I introduced with the major
structural change in r1069 ?

73

Leif