| DSP Blockset | Search Help Desk |
Using Matrices with Nonmatrix Blocks
There are a number of blocks in the blockset that are not specifically matrix-oriented, and do not contain any of the parameters discussed in the previous sections (e.g., Matrix size or Number of channels). Most of these blocks are element-oriented, and simply process matrices in the same way that they process scalars and vectors - each element independently. Additionally, there are a few blocks that are heavily vector-oriented or scalar-oriented, and are not intended to perform matrix operations. All three types are discussed below.Passing Matrices to Element-Oriented Blocks
Element-oriented blocks, such as some of those in the Elementary Functions library (e.g., Complex Exponential and dB), accept inputs of all sizes, and operate on each input element independently. The output from an element-oriented block is always the same size as the input; scalars remain scalars, vectors remain vectors, and matrices remain matrices.Passing Matrices to Vector-Oriented Blocks
Vector-oriented blocks, such as those in the Vector Functions library and many Simulink libraries, treat all inputs as vectors - including matrices. An N-by-M matrix is treated as a vector with N*M elements. That is, the vector-oriented block treats the matrix inputu as the vector input u(:).
Passing Matrices to Scalar-Oriented Blocks
Scalar-oriented blocks, such as the Short-Time FFT and adaptive filter blocks, accept only scalar inputs. Matrix inputs cause an error.