| DSP Blockset Release Notes | |
New Versions of Previous Blocks
Elementary Math Blocks
Many of the Elementary Math blocks appearing in DSP Blockset Version 1.0a have migrated into the Simulink/Nonlinear block library, yielding higher performance for both simulation and code generation.
sin, cos, tan, asin, acos, atan, atan2, sinh, cosh, tanh
exp, log, 10^u, log10, square, sqrt, pow, reciprocal, hypot, rem, mod
floor, ceiling, round, fix
These blocks were previously implemented as C-MEX S-functions, and were implemented as one function per block. The three new Simulink blocks are implemented with pop-up menu function selection for better usability. Copies of these Simulink blocks may be found in the DSP Blockset Math Functions/Elementary Math block library for easy access.
Note that you may continue to use the earlier versions of these blocks, but the older CMEX S-function implementations will continue to be employed. Such usage is now considered obsolete, and a warning will be generated. You may suppress the warnings by typing warning off from the MATLAB command line, or you may replace instances of the obsolete blocks with copies of the new blocks.
Length Block
The Length block from Version 1.0a has been migrated into the Simulink Connections library where it is now named the Width block. A copy of the Width block now appears in the Math Functions/Vector Math library, along with a Complex Width counterpart, for convenience.
Note that you may continue to use the earlier version of this blocks, but the older CMEX S-function implementation will continue to be employed. Such usage is now considered obsolete, and a warning will be generated. You may suppress the warnings by typing warning off from the MATLAB command line, or you may replace instances of the obsolete block with copies of the new block.
Buffer Block
Changes which will affect the behavior of simulations involving buffering and/or unbuffering operations have taken place in Version 2.0. Essentially, these blocks have been modified such that their behaviors are simpler and more consistent with real-time code practices.
Assume a buffer block configured with buffer size N and overlap OV. The buffer block now outputs an initial buffer of length N at time 0 (i.e., the start of simulation), while acquiring the next (N-OV) number of new samples. After (N-OV) samples have been acquired, at time (N-OV+1), the latest buffer will be output, and the acquisition of new samples resumed.
Contrast this with the earlier buffer operation, in which the buffer did not output an initial buffer of data; the first output was made at time (N-OV).
Unbuffer Block
The unbuffer block now supports only the "pipeline" mode of execution in order to guarantee consistency between simulation and real-time code execution.
There are now two unbuffer blocks. One supports "complete" unbuffering only, in which the block unbuffers the entirety of the input vector. The second supports "partial" unbuffering, in which just a portion of the input vector may be specified on which to perform the unbuffering operation.
| | Block Libraries | Enhancements to Old Blocks | |