[LAU] correct jconvolver matrix gain for convolution reverb?

Jörn Nettingsmeier nettings at stackingdwarves.net
Thu Jan 31 01:02:26 UTC 2013


hi fons, hi everyone!


playing with aliki and jconvolver again. i've recorded some sweeps from 
a single neumann kh120 loudspeaker into a tetramic. the IRs have been 
deconvolved and normalized, and after that their direct part and floor 
reflection have been trimmed off.

now when i put those files in jconvolver with a gain of 1 (which is 
inserted into a post-fader aux fx bus), the reverb level is way too 
high. i see that you always use a gain coefficient of .5 in your 
b-format examples - why?

i would like to start with the "correct" reverb level and go from there.
the initial (peak) normalisation was done in ardour. maybe i should have 
normalized the initial impulse area to unit energy rather than the 
highest peak to unit amplitude, but i'm not quite sure how to accomplish 
that. the normalization feature in aliki doesn't seem to do anything, or 
i'm misunderstanding something.

below is the config file. ignore the five inputs for now, the level 
problem is evident when using just a single input.

any hints would be much appreciated, and yes, as soon as i've listened 
through these IRs properly and fixed the config, i'll be happy to share 
them with everyone.


best,


jörn


/cd .
#
#
#                in  out   partition    maxsize    density
# --------------------------------------------------------
/convolver/new    5    4         256     400000        1.0
#
#
#              num   port name     connect to
# -----------------------------------------------
/input/name     1     Input.LL
/input/name     2     Input.L
/input/name     3     Input.M
/input/name     4     Input.R
/input/name     5     Input.RR

#
/output/name    1     Output.W
/output/name    2     Output.X
/output/name    3     Output.Y
/output/name    4     Output.Z
#
#
#               in out  gain  delay  offset  length  chan      file
# 
-----------------------------------------------------------------------------------
/impulse/read    1   1   1      0       0       0     1     LL.wav
/impulse/read    1   2   1      0       0       0     2     LL.wav
/impulse/read    1   3   1      0       0       0     3     LL.wav
/impulse/read    1   4   1      0       0       0     4     LL.wav

/impulse/read    2   1   1      0       0       0     1     L.wav
/impulse/read    2   2   1      0       0       0     2     L.wav
/impulse/read    2   3   1      0       0       0     3     L.wav
/impulse/read    2   4   1      0       0       0     4     L.wav

/impulse/read    3   1   1      0       0       0     1     M.wav
/impulse/read    3   2   1      0       0       0     2     M.wav
/impulse/read    3   3   1      0       0       0     3     M.wav
/impulse/read    3   4   1      0       0       0     4     M.wav

/impulse/read    4   1   1      0       0       0     1     R.wav
/impulse/read    4   2   1      0       0       0     2     R.wav
/impulse/read    4   3   1      0       0       0     3     R.wav
/impulse/read    4   4   1      0       0       0     4     R.wav

/impulse/read    5   1   1      0       0       0     1     RR.wav
/impulse/read    5   2   1      0       0       0     2     RR.wav
/impulse/read    5   3   1      0       0       0     3     RR.wav
/impulse/read    5   4   1      0       0       0     4     RR.wav

-- 
Jörn Nettingsmeier
Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487

Meister für Veranstaltungstechnik (Bühne/Studio)
Tonmeister VDT

http://stackingdwarves.net



More information about the Linux-audio-user mailing list