convolution tool and reverbs (semi-OT)

[Carlos Vila Deutschbein]

Hello Rene and all

On 20-Jan-00, you wrote:

> 
> Hi Martin and all,
> 
> I know that there is a technique which would be not too difficult to
> implement.
> Its similar to raytracing (or geometrical optics), you work with "mirror"
> sound sources.

ATTENTION: Semi-OT, but you asked for it...

That's right, Rene. There are basically three methods for analyzing acoustic
properties of rooms. Each one is good for a certain purpose, here they go:

Statistical theory:
This assumes that the acoustic field in the room is diffuse (i.e. There are
no sound pressure concentrations). This method gives a good approximation of
RT60 (Reverberation Time, -60dB).And calculations can be done with "paper
and pen".
According to Sabine Formula:

RT60 = 60*V/(1.086*c*S*am)

where: 
V: Volume of the room [m3]
c: sound speed (~345 m/s)
S: Area of all Walls, floors,ceilings [m2]
am: (alpha-sub-m) mean absorption coefficient of materials

A good RT60 for music could be 1.5 to 2 seconds,but it depends on the type
of music. 

Geometric Theory:
Considers sound as a series of rays which propagate in straight lines.Uses
the same principles as optics (reflection, delays,etc). Advantages: Useful
for rooms with very different materials (ex: curtains+brick walls). You can
calculate the pressure level at a certain point and analyze problematic
room shapes (concave, convex..)
You should only use this method when the size of the room is way bigger than
the analyzed wavelenght! This means you cannot model diffusion effects. 
This method includes the mirror (phantom) sources explained by Rene and the
raytracing approach.
You can draw rays by hand in 2D, but things start to mess up in 3D and with
second generation reflections, so let a computer do the job...

Wave Theory:
This method takes sound as what it really is physically: a propagating wave.
The formulas are far more complicated to develop but the results can give
information about standig wave frequencies, modal density, floating echoes,
etc. It is also useful to design rooms from scratch, using "perfect"
dimensions. You can use this theory in practice when you have a rectangular
room, otherwise, the expressions get way to complicated. 
for example: measure the distance between 2 parallel walls in your room (d)

force:  d=(lambda/2) ;search for first mode        
  
 =>     lambda=2*d    lambda(wavelenght)=c/f  c=345, f=frequency

f=c/(2*d)  frequency of the first mode. If d=4m  f=43.125 Hz: You have a
peak at this freq.


Feel free to ask for more details on that or other acoustic issues
(privately, if appropiate)

Regards
-- 
==========================================================
Carlos Vila Deutschbein        si04697 at salleURL.edu       
Enginyeria La Salle            www.salleURL.edu/~si04697/
----------------- Barcelona, Spain -----------------------
==========================================================