## Calculations FAQs

This FAQ section provides answers to frequently asked users questions on the following topics:

- Calculation parameters, incertainties;
- Solver, normalized calculation method;
- Data analysis;
- Topography.

When a 'new calculation' is defined, you have to select manually all the elements you want to take into account to do this calculation. This is a tedious step but this is the only way to check properly the elements that will be involved in your new calculation.

The DefaultSolver includes a simplified approach to take into account downward refracting conditions (see reference manual for more details).

A new solver called ANIME3D will be soon available and includes a transformation based method to take into account refraction by sound speed celerity gradients.

Not yet, but we will start new developments on that topic next year.

Code_TYMPAN offers the possibility to enter spectra from 16 Hz to 16kHz but the quality of the results will be:

- poor in the very low frequency domain (below 50 Hz) because the source coherence and the modal behavior of structures is not taken into account;
- poor in high frequencies (typically over 5 kHz) due to the impossibility to define very detailed geometries and propagation medium space and time variability.

So, if you think one of these effects dominates in your case study, don't use Code_TYMPAN or any other ray tracing based method.

Code_TYMPAN is able to calculate phase relation between all rays and every method based on that property could be developed but, still exist the difficulty to get realistic information of the time fluctuations of the propagation medium.

Code_TYMPAN works with third octave bands which is widely used in the engineering domain. User can enter his data in octave band, but Code_TYMPAN always compute in third octave band. However, Code_TYMPAN formulas do not make the assumption of a specific bandwith and could be tested on narrow band data if:

- one is able to define sound sources power in narrow bands;
- geometrical data can be defined with a spatial precision lower than the smallest wavelenth you want to consider.

These options are used when the sum of the different contributions from **one single point source to a receiver**. The coherent model consist in summing the contributions by taking into account their phase relation and thus can lead to a maximum value of the sum of the amplitudes (e.g. + 6dB for 2 contributions of the same amplitude) and a minimum value of zero, if the contributions are out of phase. The incoherent model sums the energy of the signal (p^{2}) and leads to a maximum increase of 3 dB for two contributions of the same amplitude.

Note: if a radiating surface has more than one point source, the contributions of the differents sources of this surface are considered has incoherent.