AcAna_use of scattering coefficient

In order to understand wheter the scattering coefficient (http://www.rpginc.com/research/r002a.htm) could be helpful for the analisys, i set up an "experiment".

 

I wish i checked the effects of the simplified component in different configuration in the same situation. The component is located in an standard ambient (20sqm per 270cm heigh) with standard furniture, windows and doors; in the room there is just one source (omnidirectional- it's a virtual source rather than a real approximation) and 8 relief points, all of them at the same height, the same of the source, the same of the listener's hear; the relief points are simmetrical as the component is.

I made the analisys 6 times, one for each configuration of the component. What is changing is the fact that i assigned the scattering coefficient to different part(yellow ones in picture).

  1. none
  2. left only
  3. center only
  4. right only
  5. left+right
  6. full

In this phase we are not testing the component, but the software, in order to understand wheter it could be useful.

Giancarlo Bianchi mantains that:

  • we need to find an homogeneization in the Equivalent Level of Sound Pressure within the component since it is a diffusor (for frequencies 1000-4000 Hz for which I assigned the diffusing behaviour)
  • the sound should decay sooner as effect of the loss of energy in the diffusion

I'm publishing two results that i find significative.

This grafic is a comparison of the T30: we clearly see that the effect of the diffusion as predicted. In the configuration center_only, there is no T30 lowering: this could be for the small quantity of diffusing surface. But, in this case, it's difficuolt to explain the beaviour of most of the others configuration (since there is no correspondece in T30 lowering as the diffusing surface grows).

I find even more interesting the results of the "homogenization". I won't meet G.Bianchi until monday, so i'm giving my opinion.

To check out an homogenization in the Leq iwthin the component, i'm publishing a graf of the Standard Deviation of the values collected. We expect a lowering of the St.Dev for the frequencies I assigned as scattered, as regards the none configuration. The only configuration which satisfies the "homogenization" is the center_only, while the other do not seem to work properly (due to componet's simmetry + ambient's asimmetry?  i reckon).

Most interesting is the configuration left+ right: we clearly see an inversion in the trend, we find the opposite of the results we were expecting. But it is my opinion (i spoke Stefano about this hypotesis long time ago) that, tough the lowering is quite negligible this time, this could be the proof of the fact that two diffusing surfaces interferes in the near field, which could open new (and welcome) perspectives to the research .

I will speak about all of this with G.Bianchi monday.

commenti

G.Bianchi agrees with most of my conclusions. Indeed we're ready now to work on the real component configuration. I hope to see significative results in one week.

Marco,

you are doing a great job of documentation of your progresses.

It is a great news  to see you are now able to evaluate different configurations of the chair.

I look forward to see those attempts.

What I can suggest you in the meantime, is that it would be great if you could link both form configuration and analysis results to the values of some parameters.

 

Something Like (parameters chosen randomly!):  

Depth: 3, Marker:5 --> Scattering 5 

Depth: 2.8, Marker:4 --> Scattering 3.1

...etc

This would help a lot the possible automation of your evaluation process. Of course this means you reached a clear understanding of the dynamics that link form and behaviour, that you are still investigating.

Anyway, great job

see you tomorrow.

 

Stefano