gqrx + wxtoimg = distortion?

[stefan....@gmail.com]

Hi,

I'm tinkering with gpredict, gqrx and wxtoimg at the moment. So far I can have gpredict control gqrx pretty ok. For getting the sound to wxtoimg, I used pavuctl to route the output of gqrx to the HDMI audio channel of my GPU and wxtoimg to use "monitor of HDMI" as input via the Alsa pulse plugin. wxtoimg does get some input - I can see the image changing based on the signal and I can also change the Gain in gqrx to see the volume indicator in wxtoimg change. However, the images wxtoimg produces from it's own recording are badly distorted and basically illegible. If I load the directly recorded audio file from gqrx (after conversion to 11025) into wxtoimg, the output is reasonably good.

My guess is, it might have something to do with sample rates. gqrx offers only 48khz while wxtoimg offers only 11k025 and 44k1.

I could, of course, just plug a short cable between line out and line in of the computer but I do use it for other stuff too :D. And a direct pipe from program to program would be kinda more elegant :).

Any idea how to get this done?

-Stefan

[Alexandru Csete]

Hi Stefan,

Personally, I have given up on doing things like this on linux desktop
with pulseaudio because it is broken like hell. But for what it's
worth...

I think pulseaudio uses 44.1 kHz sample rate by default and so it will
resample the audio from gqrx. The default resampler settings could be
bad for data. You can check the settings in /etc/pulse/default.pa

You could resample the output from gqrx using something like UDP
interface -> sox -> pulse to avoid resampling in pulseaudio.

Or, since now there is the option to build using portaudio, you could
use this to pipe through an ALSA loopback device
(http://www.alsa-project.org/main/index.php/Matrix:Module-aloop) or
even through Jack if you can figure that one out.

Just some ideas...

Alex