Surface de parabole non reflétante

[Yannick Tedeschi]

Bonsoir et bonnes fetes, 

Dans le but d'un projet audio, j'essaye de caractériser la réponse d'une forme particulière de parabole audio à l'air libre. Pour cela, j'ai importé ma géométrie (une boite théorique contenant la parabole) depuis logiciel CAO. J'arrive à obtenir des réponses cohérentes avec autres géométries (sphères, plats...) mais pas avec la parabole. Dans la simulation des particules, je n'en vois aucune qui rebondisse sur la surface, alors que dans les autres cas cela est très visible. 

Les parois théoriques de la boite étant à 100% absorption et la surface des objects étudiés étant à 0% d'absorption.

Dans l'image "parabole", la configuration de la parabole ne marchant pas. Le récepteur 1 étant celui où les ondes sonores devraient être concentrés par la parabole. Aucun signal réfléchi n'est capté par les 3 récepteurs, seulement celui en émission initiale.

Dans l'image "arc_cercle", une géométrie tout aussi complexe qui arrive a faire réfléchir les ondes sonores sans problèmes:

Les paramètres de simulation sont exactement pareils, l'importation des scènes se fait avec les valeurs de défaut. Je n'arrive vraiment pas à comprendre quelle différence pourrait causer ce genre de problème..

Je peux vous mettre à disposition les fichier projets si cela est nécessaire.

Merci beaucoup de votre aide, ce logiciel est une vraie pépite!

Cordialement,

Yannick T.

[Yannick Tedeschi]

For non french speaking people:

Good evening and happy holidays,

For an audio study project, I'm trying to characterize the audio response of a particular form of audio parabola in open air. To do this, I imported my geometry (a theoretical box containing the parabola) from CAD software. I manage to get coherent answers with other geometries (spheres, dishes ...) but not with the parabola. In the parabola simulation, I don't see any sound particle bouncing off the surface, while in other cases they are very visible.

The theoretical walls of the box are at 100% absorption and the surface of the studied objects are at 0% absorption.

In the "parabola" image, the configuration does not work. Receiver 1 is the one where the sound waves should be concentrated by the parabola dish. No reflected signal is picked up by the 3 receivers, only the original direct signal from the original transmission.

In the image "arc_cercle", there's an equally complex geometry but which manages to reflect the sound waves without problems:

The simulation parameters are exactly the same, the scenes are imported with the default settings. I really can't see what differences could cause this problem.

I can make the project files available if necessary.

Thank you very much for your help, this software is a real treasure!

Regards,

Yannick T.

[JP]

May be that due to reflection there is no particle that cross the receivers. Have you try to display the particles during propagation ?

See https://i-simpa-wiki.readthedocs.io/en/latest/code_configuration_SPPS.html#calculation-parameters and parameter ‘Number of sound particles per source (display)’

Best regards

[Yannick Tedeschi]

Yes, I've tried displaying the particles. I've used the Unidirectional sound source, so I could point the particle exactly were I wanted. Even if the surface is 0% absorbant, no particle ever bounced on its surface, so I'm pretty lost at this point. I've tried pointing the sound source to a flat surface and it bounces just fine. I've put a screenshot detailing the tests I'm doing, with the several unidirectional sources pointing at the parabola.

[JP]

Very curious.

Could you share your I-Simpa project file ?

JP

[Yannick Tedeschi]

Sure! Thanks for your kind help!

[JP]

All particles are lost due to numerical problems (see the 'statsSPPS' file/data in the Results tab, and the error message "Error:Calculation code error:Warning 964537l particles has been in error on 1000000l particles. The computation result may be wrong, please check the particles statistics file for more details." in the Messages Console.

It may be due to the curvature of the surface elements of the parabola.

We check this problem ASAP.

JP

[Yannick Tedeschi]

Ah, I didn't think about checking that log, my console wasn't showing me that issue, just the calculation time. It's weird though, I import my geometry directly from a Catia-generated STL, so there shouldn't be many issue regarding the geometry. It must be a meshing issue. I've discovered that using the "repair surface" while importing the scene results in the parabola disappearing from the scene, so I don't use that feature anymore. That might generate an unusable mesh, not sure about that.. 

[Nicolas Nico]

Hi,

SPPS is made for room scale simulation. TetGen software is not able to recognize your reflector as a separate volume. You should try http://acousto.sourceforge.net for advanced wave simulation at this scale.

Best regards,

[Yannick Tedeschi]

Thanks, I'll check that out!

It's weird that TetGen could not recognise my deflector though, I've tried with simpler geometries (such as a sphere) and the particles seems to interact just fine with the object. I guess it's an issue with the shape of the reflector forcing some strange meshing methods that render the volume unusable... 

In the meantime, I'll read the Acousteo doc, to see if I can get the results I need. Thanks!