[LAD] [Ann] Folve: Fuse file system that convolves flac files on-the-fly

Henner Zeller h.zeller at acm.org
Sun Sep 23 20:48:44 UTC 2012


Hi,
On 23 September 2012 12:40, Fons Adriaensen <fons at linuxaudio.org> wrote:
> On Sun, Sep 23, 2012 at 10:17:10AM -0700, Henner Zeller wrote:
>
>> The latest version now contains a simple highpass- and lowpass-filter
>> to test out the configuration. And a description how to get the Santa
>> Lucia Basilica reverb example from the jconvolver demo distribution.
>
> Be careful when using the Santa Lucia IR. It's the recording of a
> gunshot, so the direct sound is a) clipped and b) low-fi. The file
> that comes with jconvolver cuts of the direct (dry) sound for that
> reason, assuming the normal signal path through a mixer will add it.
>
> In this case it has to be added to the IR. You need something like
>
> #               in out  gain  delay  offset  length  chan      file
> #------------------------------------------------------------------------
> /impulse/read    1   1   0.1  1400    1400       0    1   santalucia.wav
> /impulse/read    2   2   0.1  1400    1400       0    2   santalucia.wav
> /impulse/dirac   1   1   1    0
> /impulse/dirac   2   2   1    0
>
> The ratio of the 'gain' parameters determines the wet/dry ratio.

Thanks Fons, this is a good idea. Just the santalucia.wav is not very
high quality indeed (and needs heavy attenuation).

But if I understand the filter configuration right, I can't really
just use a gain of 1.0 with the dirac: I need to attenuate the
original signal so that there is enough 'room' for the SantaLucia
reverb to be added ... otherwise the output signal is clipped if
already close to 0dB.

So this is what I did now, essentially adding the original signal to
it with -8dB, the rest 'filled out' with the SantaLucia:

   https://github.com/hzeller/folve/blob/master/demo-filters/SantaLucia/filter-44100.conf

(Change from before:
https://github.com/hzeller/folve/commit/d629c8899d39b96650708fc6b46e5369681b464d
)

ciao,
 -henner

>
> You can use this at either 44.1 or 48 kHz, the only difference
> will be a slightly different (10%) reverb time.
>
>
> Ciao,
>
> --
> FA
>
> A world of exhaustive, reliable metadata would be an utopia.
> It's also a pipe-dream, founded on self-delusion, nerd hubris
> and hysterically inflated market opportunities. (Cory Doctorow)
>



More information about the Linux-audio-dev mailing list