[Minor] CPU usage

[FM33]

When timer is active, even if there is no sound detection (and even if all lines are commented), the program uses 6% CPU permanently.

[moma (osmoma)]

Hello,
Very good observation.
The listener process is still running even the timer has no conditions (the timer has no commands, it's empty).

The listener (in gst-listener.c) is started by the timer (in timer.c). And the listener collects signal level (audio level decibel) values.

I have now modified the code so it switches the listener off when the timer is empty.

Notice that the listener creates a recorder process for the selected devices. It is an ordinary GStreamer recorder that outputs to a "fakesink" element. GStreamer delivers the signal values (decibel level values) via listener_message_handler(...) callback.

The listener process is a quite heavy and CPU intensive. However I haven't found a better way to gather the signal/level values.

Please see:
http://bazaar.launchpad.net/~osmoma/audio-recorder/trunk/view/head:/src/gst-listener.c

Functions:
static GstElement *listener_create_pipeline(gchar *audio_source, GList *dev_list) {
...
}

And
static gboolean listener_message_handler(GstBus * bus, GstMessage * message, gpointer data) {
...
}

[moma (osmoma)]

Re-hello,
A small addition to my previous post.
The listener (in gst-listener.c) is only needed when the timer has audio-level related commands. For example these commands will need level-data from the listener:
start if audio 5s 7%
start if voice -17 dB
stop if silence

While these commands do not need it. The timer.c can handle these by itself.
start at 10:30 pm
stop at 20:00 | 1.2GB | 1 h 20 min

I may add a test to the timer that checks the timer conditions, and starts the listener (CPU hungry dummy recorder) only when needed by the "silence", "sound" and "audio" commands.