[linux-audio-announce] Mic-In-Room Simulator

Greetings to all,

Announcing grad_Boxster for creating impulse response functions
(IR's) for rectangular parallelepipeds --- for example concert
halls modelled as giant boxes --- but using the *gradient* of the
solution field projected on a direction specified by the user,
similar to the operation of a pressure-gradient microphone --- in
a room.

At the URL below is a link to this package for downloading as well
as a NON-optimized, alternating demo of grad_Boxster versus
Boxster (i.e. "ordinary stereo versus binaural" or even "speakers
versus headphones"):

http://home.earthlink.net/~davidrclark

grad2_Boxster is also available for simulating second-order mics
in rooms. Extending grad_Boxster, the gradient of the first-order
projection is computed and projected onto a second direction
specified by the user.

By combining IR's from Boxster (pressure mic), grad_Boxster
(first-order mic), and grad2_Boxster (second-order), it should be
possible to simulate any type of mic setup and combination of
setups in any room that could be considered to be a box. Starting
with dry, monophonic recordings made in a dead room, one should be
able to simulate very sophisticated mic setups in a concert hall,
then assemble them into a mix, simulating an ensemble recording,
all with a physically consistent model.

The demo contrasts simulation of nearly coincident pressure-
gradient mics (grad_Boxster) in a 12,500 m3 concert hall versus
binaural omni pressure mics (Boxster, no HRTF) in the same hall
with the same separation between mics for both. The only
difference for the first half of the demo is whether or not the
gradient is computed, yet the sounds are very different both with
speakers and with headphones (esp. the latter). In the second
half, the positions of the guitars are slightly different (maximal
separation for either).

---------------------------

In short, grad_Boxster can be used to simulate ordinary stereo
recordings (Blumlein, etc.) in the same manner that Boxster can be
used for simulating binaural recordings, complete with reverb, echo,
stereo separation, and so on --- but with physical consistency. As
a side effect, both new programs provide other "knobs" for extending
sample generation capability. All programs can be run either from
a PyGtk GUI or from the command line.

Regards to all,
Dave

Dave Clark

_______________________________________________
linux-audio-announce mailing list
linux-audio-announce@email-addr-hidden
http://music.columbia.edu/mailman/listinfo/linux-audio-announce
Received on Tue Feb 28 04:15:06 2006