Skip to main content

Best practices when simulating in Treble

Wave solver

To accurately predict the impulse response of a space using Treble’s wave solver, the amount of detail in the imported 3d model is critical. The higher the transition frequency between the wave and geometrical solvers, the higher the level of detail necessary to model. A good rule of thumb is to model detail down to approximately a quarter wavelength of the transition frequency; or as detailed in the following table:

Transition frequencyOptimal level of detail
177Hz [125 Hz octave band]0.4m
355Hz [250 Hz octave band]0.2m
710Hz [500 Hz octave band]0.1m
1420Hz [1000 Hz octave band]0.05m
2840Hz [2000 Hz octave band]0.025m
  • When modelling furniture, windows, doors and other similar interior fittings it’s recommended to model each individual element. It is not recommended to make broad geometrical simplifications such as modelling a whole audience area as a single box.

  • Windows and doors should be modelled as indentation into the surface they’re placed on. The depth of the indentation should correspond approximately with the depth of the windowsill/doorframe.

  • Traditional chairs with a backrest should be modelled as two 2d planes.

  • A simple table on four feet should be modelled as a single 2d plane. In the case of a desk with a more complex structure the overall shape should be modelled.

  • Computer monitors should be modelled as vertical 2d planes.

  • When modelling cupboards and shelves the transition frequency must be given careful consideration. At lower transition frequencies modelling a shelfing unit as a single box can be a good approximation. When the gap between the shelves becomes less than a quarter of the wavelength of the transition frequency it’s recommended to model each individual shelf.

  • The height of a room with a suspended acoustic ceiling should be modelled up to the acoustic ceiling. In other words, the airgap between the acoustic ceiling and structural ceiling should not be modelled. The Acoustic ceiling should be assigned a material from the material database that corresponds with the plenum depth.

  • Curtains should be modelled as an extrusion from a wall surface and not as a separate surface in front of it. They should be assigned a material from the material database that takes into account the mounting distance from the backing wall. Creases in the curtain should only be modelled if their depth are within the optimal level of detail of the chosen transition frequency.

Below are reference pictures of a model that was created based on an office space, where the target transition frequency was 710 Hz:

Office space

Office space

Model of office space

Model of office space

Geometrical acoustics solver

  • For the geometrical acoustics solver it’s also recommended to model each individual structural/furnishing element individually in the same manner as previously detailed for the wave solver when setting up a 3d model.

  • For a good simulation result it’s imperative to consider the scattering values of all surfaces. Generally, no surface should have a scattering coefficient lower than 0.1.

  • The following table contains a range of scattering coefficients we recommend using as a starting point:

MaterialsScattering coefficient
Audience area0.6 – 0.8
Furniture0.4 – 0.7
Exposed bricks0.3 – 0.4
Rough unpainted concrete0.15 – 0.25
Smooth surfaces0.1 - 0.15
  • Additionally, the following factors should be considered:
    • It’s generally more detrimental to underestimate scattering rather than overestimating it, when comparing with measured values.
    • Small objects and edges are generally more diffusive than larger surfaces and should ideally be assigned a scattering coefficient in the higher range of the respective field in the table.
    • The more geometrical detail omitted in the 3d model, the more important it becomes to input a sufficiently high scattering coefficient. E.g. when omitting modelling small objects on a table it’s important to increase the scattering coefficient of said table when compared with an empty one.