| DSP Blockset Release Notes | |
Matrix Support
Matrix support is now provided natively within Simulink, which allows the DSP Blockset to generate and process matrix signals in a much more natural manner than was previously possible.
Generating Matrix Signals
Matrix signals can now be generated in the same way as vector or scalar signals, using appropriate source blocks such as Signal From Workspace or DSP Constant. Where you would have previously specified a vector or scalar, you can now use standard MATLAB notation to specify a matrix. For example, enter [1 2 3;4 5 6] in the Constant value parameter of the DSP Constant block to generate the 2-by-3 output matrix.
When Signal dimensions is selected from the model window Format menu, Simulink displays the dimensions of an M-by-N matrix signal as a pair of bracketed integers, [MxN]. Simulink does not display any size information for a scalar signal. Dimension information for a signal can also be displayed in a model by attaching a Probe block with Probe signal dimensions selected.
Row vectors and column vectors are matrices with M=1 or N=1, respectively. These are generated by the DSP Constant block when a vector is specified in the Constant value parameter and the Interpret vector parameters as 1-D check box is not selected.
In addition, Simulink recognizes one-dimensional (1-D) vectors, which are ordered lists of values. These are generated by the DSP Constant block when a vector is specified in the Constant value parameter and the Interpret vector parameters as 1-D check box is selected. When Signal dimensions is selected from the model window Format menu, Simulink displays the size of a 1-D vector signal as an unbracketed integer. For convenience, DSP Blockset blocks treat 1-D vectors as column vectors for most operations.
Processing Matrix Signals
The most noticeable benefit of the new matrix enhancements is that you no longer need to manually specify the size of matrix inputs to a block. All DSP Blockset blocks now automatically detect input dimensions, and will generate an error if those dimensions are not appropriate for the specific operation.
Since input dimensions are detected automatically by new (Version 4.0) blocks, a matrix output of an older block may need to be altered before you use is as an input to a new block, and vice versa. See Adding Release 12 Blocks to Release 11 Models for more information.
The interpretation of sample-based and frame-based matrices is the same as in previous releases. An M-by-N sample-based matrix represents M*N independent channels, each containing one sample. An M-by-N frame-based matrix represents M consecutive samples from each of N independent channels. That is, each column of a frame-based matrix represents M consecutive samples from a single channel.
| | DSP Blockset 4.0 Release Notes | Frame Support | |