Building Blocks

Building Blocks

The core functionality is implemented using a number of mathematical structures or primitives. The structures are templated to facilitate both single and double precision and alignments to support SIMD instructions - either as a result of a compiler optimizing the code or as a results of the programmers use of intrinsics or inline assembly.

Mathematical primitives

The mathematical primitives are:

• sps::point_t

  Vector used for either positions or directions

• sps::rect_t

  Rectangle made up of four positions (vertices)

• sps::euler_t

  A set of 3 Euler angles for intrinsic rotations using the y-x-y'-convention.

• sps::bbox_t

  Axis-parallel bounding box used for fast parallel computation and to optimize the cache usage. It is represented using two points for minimum and maximum range of x,y, and z.

• sps::element_t

  Element represented using two scalars for its half width, and half height, a center position, and a set of Euler angles for its orientation.

  • sps::element_t::hw
  • sps::element_t::hh
  • sps::element_t::center
  • sps::element_t::euler

  The structure further contains a number of variables used for caching, vectors used for orientation and vertices for display. These structures are read-only for the user.

  • sps::element_t::normal
  • sps::element_t::uvector
  • sps::element_t::vvector
  • sps::element_t::vertices

Transducer representation

Frequency domain responses

A transducer or transducer array can be constructed using a number of primitives. Explain how:

• Elements carry apodization
• Elements have phase and delay
• fnm::FocusingType fnm::FocusingTypeNS::Pythagorean, fnm::FocusingType::Delays, fnm::FocusingType::Rayleigh
• Sub-elements shared delay and apodization
• fnm::Aperture<T>::FocusUpdate
• fnm::FocusingTypeNS::Value
• fnm::FocusingTypeNS::Rayleigh
• Use DoxyPy to refer to FNM Calculation functions