Sound Generation Methods

This section describes the generation method for each type of sound source supported by Sound Composer.

  • Harmonics - 1 parameter uses the same method as that used for Generating a Signal from Waterfall, following the RPM profile indicated by the user. The order numbers are used to get the frequency F of each sinusoid (F(order N) = N x RPM / 60).
  • Harmonics - 2 parameters uses the same sound generation method as Harmonics - 1 parameter. The main difference is that in this case, the level is controlled by two parameters (frequency is still controlled by RPM, which is one of the two control parameters). To get the level values at the unknown points of the dataset, the IDW (Inverse-Distance Weighting) method is used for each order in the dataset (see below for more information).
  • Audio sources require no transformation of the input. The input sound is used as is.
  • Spectrum uses the Inverse FFT synthesis method, as described in Methods for Sound Creation.
  • Broadband noise - 1 parameter uses the inverse FFT synthesis method (see Methods for Sound Creation), but the spectrum levels evolve according to the control parameter profile specified by the user. For each frequency, the spectrum level value at a given control point is therefore created by linear interpolation of the dB levels taken from the two surrounding known spectra (the two closest spectra in the dataset whose control values are higher and lower than the control parameter value where the sound has to be generated).
  • Broadband noise - 2 parameters uses the same sound generation method as Broadband noise - 1 parameter. The main difference is that in this case, the level is controlled by two parameters. To get the level values at the unknown points of the dataset, the IDW (Inverse-Distance Weighting) method is used at each frequency of the spectra (see below for more information).
Note: The more frequency points the spectra have, the longer the generation process.

Details about the IDW interpolation method

To get the level at a given point (control1 C1, control2 C2), a 2-dimensional interpolation has to be done using the data from the input dataset (the known points). Ansys Sound: Analysis and Specification uses an interpolation based on the IDW method (Inverse-Distance Weighting). The dB level at an unknown point U (C1, C2) is based on the value of each known point K, weighted by its distance to the unknown point.

The dB level at point U is then:

If dist(K,U) ≠ 0, ∀ K ∈ {input dataset} .

L(U) = L(K) if U is a known point.

where

  • K are the known points
  • U is the unknown point
  • the weight WK(U) is defined as follows:
  • dist(K,U) is the Euclidean distance between K and U:
Important: In order to be independent of the order of magnitude of parameters C1 and C2 (imagine one is a speed in km/h for instance between 0 and 50 km/h, the other is a revolution speed in rpm, between 1000 and 8000 rpm), the dataset control parameter values are pre-processed to generate control parameter values between 0 and 1 for each, before calculating the distances and weights.
Note: The bigger the known dataset, the longer the calculation time.