[pulseaudio-discuss] selective sound amplification by frequency @ PulseAudio.

[Tanu Kaskinen]

On Thu, Jan 10, 2008 at 02:22:16PM +0100, Adam Sulmicki wrote:
> Of all three modules the mbeq seems to give me best results,. However I did
> not try the other modules too hard so maybe I just used them wrongly (for
> example wrong order of parameters). Also, I had to edit the mbeq source to
> remove the +30 UpperBound limit.
> 
> Now, I wonder is there some way to visually verify that it works ( sort of
> like the audacity visual), besides having the subjective impression that it
> works.

Every sink has a source associated it, named <sink
name>.monitor, which can be used for recording. I couldn't
get audacity to use pulse as the recording device, but this
way you can create a noise file:

First, play some white noise (turn the volume down before
running this...):

    pacat -p --device=ladspa_out < /dev/urandom

Then, record the output:

    pacat -r --device=ladspa_out.monitor > noise.raw

Then stop both commands.

The resulting file is just raw samples, with the following
sample format: signed 16bit native endian (LE here) integer,
44100 Hz sample rate, 2 channels (stereo). It can be loaded into audacity using
File->Import...->Raw Data...

>From here on I guess you know this stuff better than me, I
just tried audacity's spectrogram for the first time.

-- 
Tanu Kaskinen