Signal Processing Blockset Major Bug Fixes

Autocorrelation Block No longer Causes a Segmentation Violation While Generating C Code

Generating code from a model containing the Autocorrelation block with constant sample time and the Inline parameters check box selected in the Configuration Parameters dialog box no longer causes a segmentation violation.

Fixed-Point Optimization Error in Filter Realization Wizard Corrected

Previously, when you applied optimizations for 1, -1, and 0 gains on fixed- point realizations of lattice structures in the Filter Realization Wizard, some Data Type Conversion blocks were removed. The optimized model then produced significantly different results when compared to the results from the non-optimized model. This has been fixed.

Improved Sine Wave Block Stability for Long Run-Time

In previous releases, the Sine Wave block output, in trigonometric evaluation mode, would develop unstable behavior over long run-time due to the accumulation of small quantization errors. Also, in differential mode, the block output would, under certain circumstances, decrease in amplitude over long run-time. These behaviors were especially noticeable for single- precision floating-point outputs. In version 6.0, both of these problems have been fixed via algorithm modifications.

Initial Seed Parameter of the Random Source block is Tunable

Previously, the Initial seed parameter of the Random Source block could not be tuned. Now, the parameter can be tuned in simulation and code generation.

LMS Filter, Block LMS Filter, Fast Block LMS Filter, and RLS Filter Blocks Interpret Sample-Based Vector Signals Correctly

Previously, these blocks interpreted vector sample-based signals, input at the Input and Desired ports, as single-channel frame-based signals. Now, these blocks only support scalar sample-based signals and single-channel frame-based signals at these ports.

Resolved Complexity Issue with the LMS Filter and Block LMS Filter Blocks

Previously, the signals connected to the Input and Desired ports of the LMS Filter and Block LMS Filter blocks did not need to have the same complexity. However, if one signal was real and the other signal was complex, the adaptive filter might not converge. Now, the blocks produce an error if the signals connected to the Input and Desired ports do not have the same complexity.