| DSP Blockset Release Notes | |
New Blocks
The key new block for Release 13 is the Digital Filter block. For descriptions of this and other new blocks and links to their online reference pages, see the following sections:
Cumulative Product
The new Cumulative Product block is in the Math Operations library. This block computes the cumulative product of the row or column elements of the input matrix.
Digital Filter
The new Digital Filter block in the Filter Designs library is ideal for implementing digital filters that you have already designed. The block supports a variety of filter structures, and can implement static filters with fixed coefficients, as well as time-varying filters with coefficients that change over time. Using the filter you specify, the block individually filters each channel of the input signal and outputs the result. The block's output numerically matches the output of the filter function in the Filter Design Toolbox and the filter function in the Signal Processing Toolbox.
You can use the Digital Filter block to efficiently simulate the numerical behavior of your filter on a single- or double-precision floating-point system such as a personal computer or DSP chip. You can also use the block to generate highly optimized Real-Time Workshop ANSI C code for use in single- and double-precision floating-point embedded systems. (Note that
To implement a filter with the Digital Filter block, you must provide the following basic information about your filter:
The block supports the following filter structures:
DWT and IDWT
The new DWT block in the Transforms library computes the discrete wavelet transform (DWT) of its input. The IDWT block computes the inverse DWT of its input. These blocks are identical to the Dyadic Analysis Filter Bank and Dyadic Synthesis Filter Bank blocks in the Multirate Filters library.
The DWT block uses a filter bank with specified highpass and lowpass FIR filters to decompose the input signal into subbands that have smaller bandwidths and slower sample rates. Similarly, the IDWT block uses a filter bank to reconstruct the original signal from the subbands output by the DWT block.
The filter bank filters in both blocks can be user-defined or wavelet-based. Use of wavelet-based filters requires the Wavelet Toolbox.
Interpolation
The new Interpolation block in the Signal Operations library interpolates real-valued input signals at points between sample times by using linear or FIR interpolation. You specify the times at which the block should interpolate points, and the block outputs the interpolated values.
LPC to LSF/LSP Conversion and LSF/LSP to LPC Conversion
The new LPC to LSF/LSP Conversion block and LSF/LSP to LPC Conversion block in the Linear Prediction library are ideal for speech applications such as speech coding and speech recognition. The first block allows you to convert linear prediction coefficients (LPCs) to line spectral pairs (LSPs) or line spectral frequencies (LSFs). The second block converts LSFs or LSPs back to LPCs. When converting LPCs, you can adjust the accuracy of the LSF or LSP output and set the block to flag invalid outputs due to unstable polynomial inputs.
Overwrite Values
The new Overwrite Values block in the Matrix Operations library overwrites either a specified submatrix of the input matrix or a specified portion of the input's main diagonal.
Two-Channel Analysis Subband Filter and Two-Channel Synthesis Subband Filter
The new Two-Channel Analysis Subband Filter block in the Multirate Filters library decomposes a signal into a high-frequency subband and a low-frequency subband using the specified highpass and lowpass FIR filters. Each subband has half the bandwidth and half the sample rate of the original signal.
Similarly, the new Two-Channel Synthesis Subband Filter block uses a filter bank to reconstruct the original signal from the high-frequency and a low-frequency subbands output by the Two-Channel Analysis Subband Filter block.
| | Expanding Fixed-Point Data Type Support | Enhanced Blocks | |