[Coder(365)]

I'm building a streaming and recording server for a local volunteer group. I've been trying to find a way to either...

... record audio from the sound card, process it and stream it

OR

... record audio on one card, process and save. and use the other sound card for streaming (2 cards needed).


I've come to the conclusion that option 2 is the best route, and I've began to tailor my shell script towards that. But... I've run into a snag.


Here's the issue.

I can't find a way to specify the device for SoX to use when recording. Any Ideas?

rec -t wav -q - silence 1 1 10% -1 7 2% | lame -S --resample 11025 -b 24 $file_prefix-$date.mp3 &
sleep 2
rec -t wav -q - | lame -S --resample 11025 -b 24 - - | ezstream -c /hamradio/scripts/stream.xml &

[ntheory]

Any success doing things the Unixy way and using "cat" to fetch the audio from the soundcard and then to write it back?

Although when you do something like this:

cat /dev/sound/dsp0 | magic-sound-processing-scripts > /dev/sound/dsp1

I'm not sure how you set the sampling rate or if there's an implied rate of the sound card's maximum rate (44.1KHz or 48KHz). You may have to write a simple C program that can set the rate and then pump STDIN to/STDOUT from the sound card.

[Coder(365)]

Actually it helped out a lot.

If I change the code to this (below) it works great!

Thanks NotTheory!

rec -t wav -q - silence 1 1 10% -1 7 2% | lame -S --resample 11025 -b 24 - /hamradio/audio/$file_prefix-$date.mp3 
cat /dev/dsp1 | rec -t wav -q - |lame -S --resample 11025 -b 24 - - | ezstream -c /hamradio/scripts/stream.xml

I took the "sleep 2" out because it wasn't needed.

Edit: I removed the ampersand from the end of each command so the forum would parse it correctly. For those that don't know the ampersand ( & ) puts the command into the background.

This post has been edited by Coder(365): 17 April 2008 - 03:54 PM