Release 11 Enhancements (Release 11 New Features)

Release 11 New Features

DSP Blockset 3.0

Version 3.0 of the DSP Blockset is a major release, and introduces a substantial set of new features:

All blocks now transparently handle both real and complex data.

Dedicated complex blocks (such as Complex To Workspace and Mag/Angle Join) have been removed from the blockset.

Some of these complex blocks have been merged with their real counterparts. For example, the Version 2.2 FFT and Complex FFT blocks are now a single block, FFT. Other complex-data blocks, including most of those in the 2.2 Complex library, are now a part of the Simulink library (usually under different names). Examples are the 2.2 Real and Imag blocks, which are now combined as the Complex to Real-Imag block.

All blocks support frame-based processing for increased throughput rates.

Most blocks that operate on sequential time-samples now offer a Frame-based inputs checkbox in the parameter dialog box. When you check the box, the block accepts frames of buffered time-samples rather than a scalar sequence over time. Frame-based processing can return great increases in efficiency for both simulated and compiled models.

All blocks support the multirate sample time enhancements in Simulink 3.0.

The Sample time field in the parameter dialog box has been removed from almost all blocks. Blocks now automatically detect the sample times of inputs. Source blocks (such as Signal From Workspace) still retain the Sample time parameter.

Many blocks support internal buffer reuse for in-place algorithms and global sharing. Inplace algorithms reuse the same block of memory to store the intermediate results of a series of related operations. Global sharing allows a previously allocated block of memory that is no longer in use at a given time step to be recruited for a different operation. These buffer reuse modes help to reduce the memory footprint of both the simulation and generated code.
Real-time audio support for the PC.
New speech, audio, and wavelet demos have been added.
Many additional DSP block algorithms are now inlined and optimized in code generated using Real-Time Workshop Target Language Compiler™ templates.

There are also a number of new and enhanced blocks, and new libraries. The next few pages outline the new additions, and provide pointers to the complete feature descriptions in the DSP Blockset User's Guide. See Chapter 1 of the DSP Blockset User's Guide for an overview of the blockset's contents.

Also see the DSP Blockset readme file for a summary of the new additions. To view the readme file, at the MATLAB command line type

info dspblks

Note
The DSP Blockset 3.0 requires Simulink 3.0.

Running Different Blockset Versions

When you install the DSP Blockset 3.0 on your computer, Version 2.2 of the blockset is also installed.

Run Version 3.0 by typing dsplib. To run Version 2.2, type dsplib 2.

Incompatibilities Between 3.0 and 2.2

Because of the extensive changes introduced in this release to support the new Simulink complex data format, incompatibilities can arise when 3.0 blocks are used in models containing 2.2 blocks. See "Upgrading to DSP Blockset 3.0 and Communications Toolbox 1.4" in Chapter 4 for information about migrating a model to the current version.

Library Structure

The library structure has undergone further refinement for Version 3.0. The major alterations are:

The Spectrum Analysis library in Version 2.2 has been replaced by the Version 3.0 Estimation library. This library contains two additional libraries, Parametric Estimation and Power Spectrum Estimation.
A new Linear Algebra library has been added in the Math Functions library. The library primarily offers blocks for matrix factorization and linear equation solvers.
The Complex library has been removed from the Math Functions library as a result of the change in the complex data format. Blocks that were formerly in the Complex library have either been combined with their real-data counterparts, or relocated to other libraries. See Table 1-2.

Data Frames

Most blocks whose operation can benefit from block processing now accept data frames, vectors whose elements represent consecutive time samples from a single signal. Framed data is a common format in real-time systems, where the data acquisition hardware often operates most efficiently by accumulating a large number of signal samples at a high rate, and then propagating these samples to the real-time system as a block, or frame, of data. Data frames can also be constructed through the usual DSP Blockset buffering operations (using the Buffer block, for example).

See "Working with Samples and Frames" in Chapter 2 of the User's Guide for a complete discussion of the frame data format, and how to use it to improve model efficiency.

Upgrading Your Models to Use Data Frames

You can realize large improvements in the efficiency of your models by using data frames whenever possible. Although throughput gains are particularly pronounced in systems where the sampled data is introduced in a framed format (such as speech and audio), non-real-time simulations also benefit as a result of the reduction in block-to-block communication overhead.

Complex Data

All blocks in the DSP Blockset are now capable of processing both real and complex data (using Simulink's new complex data type). In cases where two separate blocks were previously provided for real and complex inputs (e.g., FFT and Complex FFT), there is now a single block (FFT) that operates on both real and complex data. This enhancement greatly simplifies the contents of most libraries, in addition to allowing the removal of the Complex library from Math Functions. Blocks in the Complex library that could not be combined with a real data counterpart (e.g., Imag) are now in the Simulink Math library (usually under a different name). Table 1-2 lists the new names and locations of all former 2.2 blocks.

If any of your models use complex data, be sure to read "Why You Need to Update Your Models to Use the New Complex Data Format" in Chapter 4 before adding any Version 3.0 blocks.

Multirate Sample Time Enhancements

As a result of the multirate sample time enhancements in Simulink 3.0, all nonsource DSP blocks now inherit and propagate their sample times. This means that you do not need to track sample times manually throughout a model; when you make a change to the sample time of a source block, all other DSP blocks in the model automatically adjust to the propagated sample time.

New and Enhanced Blocks

Table 1-1 lists the new blocks in Version 3.0. Among the most significant additions are the linear algebra blocks and real-time audio blocks.

**Table 1-1: New Blocks in the DSP Blockset 3.0**
Block Library	Block Name	Purpose
DSP Sources	Chirp	Generate a swept-frequency cosine.
	Discrete Constant	Generate a constant.
	From Wave Device	Read audio data from a standard audio device in real-time (Windows 95/98/NT only).
	From Wave File	Read audio data from a Microsoft Wave (.wav) file (Windows 95/98/NT only).
	Triggered Signal From Workspace	Acquire and output a workspace signal when triggered.
	Sine Wave	Generate one or more sine waves.
DSP Sinks	Buffered FFT Frame Scope	Compute and display the frequency content of an input sequence.
	FFT Frame Scope	Compute and display the frequency content of a framed input.
	Frequency Frame Scope	Display frame-based data.
	Matrix Viewer	Display a matrix as an image with values mapped to colors.
	Time Frame Scope	Display frame-based data.
	To Wave Device	Send audio data to a standard audio device in real-time (Windows 95/98/NT only).
	To Wave File	Write audio data to file in the Microsoft Wave (.wav) format (Windows 95/98/NT only).
	User-defined Frame Scope	Display frame-based data.
Elementary Functions	Contiguous Copy	Recreate the input in a contiguous block of memory (for code generation).
	Convert Complex DSP to Simulink	Convert complex data from the DSP Blockset v2.2 format to the Simulink v3 format.
	Convert Complex Simulink to DSP	Convert complex data from the Simulink v3 format to the DSP Blockset v2.2 format.
	Inherit Complexity	Change the complexity of the input to match that of a reference signal.
	Variable Selector	Select a subset of elements (submatrix) in a matrix.
Matrix Functions	Create Diagonal Matrix	Create a matrix from a vector diagonal.
	Extract Diagonal	Create a vector from the elements of a matrix diagonal.
	Extract Triangular Matrix	Extract the lower or upper triangle from an input matrix.
	Matrix Product	Multiply the elements on a specified matrix row or column.
	Matrix Scaling	Scale the rows or columns of a matrix by a specified vector.
	Matrix Sum	Sum the elements on a specified matrix row or column.
	Permute Matrix	Reorder the rows or columns of a matrix.
Linear Algebra	Backward Substitution	Solve the equation Ux=b for upper triangular matrix U.
	Cholesky Factorization	Factor a Hermitian positive definite matrix into triangular components.
	Cholesky Solver	Solve the equation Sx=b for Hermitian positive definite matrix S.
	Forward Substitution	Solve the equation Lx=b for lower triangular matrix U.
	LDL Factorization	Factor a Hermitian positive definite matrix into lower, upper, and diagonal components.
	LDL Solver	Solve the equation Sx=b for Hermitian positive definite matrix S.
	LU Factorization	Factor a square matrix into lower and upper triangular components.
	LU Solver	Solve the equation Ax=b for square matrix A.
	QR Factorization	Factor a rectangular matrix into unitary and upper triangular components.
	QR Solver	Find a minimum-norm-residual solution to the equation Ax=b.
	Reciprocal Condition	Compute the reciprocal condition of a square matrix in the 1-norm.
Buffers	Queue	Buffer inputs into a FIFO (first input, first output) register.
	Rebuffer	Increase or decrease the size of the input frame.
	Stack	Buffer inputs into a LIFO (last input, first output) register.
Switches and Counters	Counter	Count up or down through a specified range of numbers.
	Edge Detector	Detect transition of input from zero to non-zero value.
	Event-Count Comparator	Detect threshold crossing of accumulated non-zero events.
	Multiphase Clock	Generate multiple binary clock signals.
Parametric Estimation	Burg AR Estimator	Compute an estimate of AR model parameters using the Burg method.
	Covariance AR Estimator	Compute an estimate of AR model parameters using the covariance method.
	Modified Covariance AR Estimator	Compute an estimate of AR model parameters using the modified covariance method.
	Yule-Walker AR Estimator	Compute an estimate of AR model parameters using the Yule-Walker method.
Power Spectrum Estimation	Covariance Method	Compute a parametric spectral estimate using the covariance method.
	Magnitude FFT	Compute a nonparametric estimate of the spectrum using the periodogram method.
	Modified Covariance Method	Compute a parametric spectral estimate using the modified covariance method.
	Short-Time FFT	Compute a nonparametric estimate of the spectrum using the modified, averaged periodogram method.
Filter Realizations	Biquadratic Filter	Apply a cascade of biquadratic (second-order-section) filters to the input.
	Direct-Form II Transpose Filter	Apply an IIR filter to the input.
	Time-Varying Direct-Form II Transpose Filter	Apply a variable IIR filter to the input.
	Time-Varying Lattice Filter	Apply a variable lattice filter to the input.
Multirate Filters	Dyadic Analysis Filter Bank	Decompose a signal using a dyadic multirate filter bank.
Multirate Filters	Dyadic Synthesis Filter Bank	Reconstruct a signal using a dyadic multirate filter bank.

In addition to the new blocks above, most Version 2.2 blocks have received enhancements for Version 3.0, and many have changed names (primarily as a result of the new complex data format). Table 1-2 below lists all of the 2.2 blocks alphabetically, and shows the corresponding 3.0 block and library location.

**Table 1-2: Enhanced Blocks in the DSP Blockset 3.0**
2.2 Block Name	3.0 Block Name	Library Location
Analog Filter Design	same	same
Analytic Signal	same	same
Angle	Complex to Magnitude-Angle	Simulink
Autocorrelation	same	same
Buffer	same	same
Buffered FFT Scope	Buffered FFT Frame Scope	same
Burg Method	same	Power Spectrum Estimation
Commutator	same	same
Complex Autocorrelation	Autocorrelation	same
Complex Buffer	Buffer	same
Complex Buffered FFT Scope	Buffered FFT Frame Scope	same
Complex Cepstrum	same	same
Complex Constant	Constant	Simulink
Complex Delay	Integer Delay	same
Complex Demux	Demux	Simulink
Complex Diagonal Matrix	Constant Diagonal Matrix	same
Complex Dot Product	Dot Product	Simulink
Complex Exponential	same	Elementary Functions
Complex FFT Scope	FFT Frame Scope	same
Complex Flip	Flip	same
Complex From Workspace	Signal From Workspace	same
Complex Gain	Gain	Simulink
Complex Kalman Adaptive Filter	Kalman Adaptive Filter	same
Complex Levinson-Durbin	Levinson Solver	same
Complex LMS Adaptive Filter	LMS Adaptive Filter	same
Complex LPC	LPC	same
Complex Matrix Constant	Matrix Constant	same
Complex Matrix From Workspace	Matrix From Workspace	same
Complex Matrix Multiplication	Matrix Multiplication	same
Complex Matrix To Workspace	Matrix To Workspace	same
Complex Multiply	Product	Simulink
Complex Mux	Mux	Simulink
Complex Normalization	Normalization	same
Complex Partial Unbuffer	Partial Unbuffer	same
Complex Reciprocal	Math Function	Simulink
Complex RLS Adaptive Filter	RLS Adaptive Filter	same
Complex Selector	Selector	Simulink
Complex Submatrix	Submatrix	same
Complex Sum	Sum	Simulink
Complex To Workspace	To Workspace	Simulink
Complex Transpose	Transpose	same
Complex Unbuffer	Unbuffer	same
Complex Unit Delay	Integer Delay	same
Complex Width	Width	Simulink
Complex Zero Pad	Zero Pad	same
Conjugate	Math Function	Simulink
Constant Exponent	Math Function	Simulink
Convolution	same	same
Convolution C-C	Convolution	same
Convolution C-R	Convolution	same
Correlation	same	same
Correlation C-C	Correlation	same
Correlation C-R	Correlation	same
Cumulative Sum	same	same
dB	same	Elementary Functions
dB Gain	same	Elementary Functions
DCT	same	same
Delay	Integer Delay	same
Detrend	same	same
Diagonal Matrix	Constant Diagonal Matrix	same
Difference	same	same
Digital FIR Filter Design	same	same
Digital IIR Filter Design	same	same
Distributor	same	same
Dot Product	same	Simulink
Downsample	same	same
FFT	same	same
FFT Scope	FFT Frame Scope	same
Filter	Discrete Filter	Simulink
Filter Realization Wizard	same	same
FIR Decimation	same	same
FIR Interpolation	same	same
FIR Rate Conversion	same	same
FIR Rate Conversion (Frame)	FIR Rate Conversion	same
Fixed Truncation	Rounding Function	Simulink
Flip	same	same
Frequency Vector Scope	Frequency Frame Scope	same
Hermitian Transpose	Transpose	same
Histogram	same	same
IDCT	same	same
IFFT	same	same
Imag	Complex to Real-Imag	Simulink
Inverse-FFT FIR Filter Design	obsolete	same
Join	Real-Imag to Complex	Simulink
Kalman Adaptive Filter	same	same
Least Squares FIR Filter Design	same	same
Levinson-Durbin	Levinson Solver	Linear Algebra
LMS Adaptive Filter	same	same
LPC	same	same
Mag/Angle Join	Magnitude-Angle to Complex	Simulink
Mag/Angle Split	Complex to Magnitude-Angle	Simulink
Magnitude	Abs	Simulink
Magnitude Squared	Math Function	Simulink
Math Function	same	Simulink
Matrix Constant	same	same
Matrix From Workspace	same	same
Matrix Multiplication	same	same
Matrix To Workspace	same	same
Maximum	same	same
Mean	same	same
Median	same	same
Minimum	same	same
Multichannel IIR Filter	Direct-Form II Transpose Filter	same
Multichannel IIR Filter (Frame)	Direct-Form II Transpose Filter	same
N-Sample Enable	same	same
N-Sample Enable w/Reset	N-Sample Enable	same
N-Sample Switch	same	same
Normalization	same	same
Overlap-Add FFT Filter	same	same
Overlap-Save FFT Filter	same	same
Partial Unbuffer	same	same
Periodogram	Short-Time FFT	Power Spectrum Estimation
Quantizer	same	Simulink
Real	Complex to Real-Imag	Simulink
Real Cepstrum	same	same
Real DCT	DCT	same
Real FFT	FFT	same
Real IDCT	IDCT	same
Real IFFT	IFFT	same
Real To Complex	Real-Imag to Complex	Simulink
Remez FIR Filter Design	same	same
Repeat	same	same
Reshape	same	same
RLS Adaptive Filter	same	same
RMS	same	same
Rounding Function	same	Simulink
Running Histogram	Histogram	same
Running Maximum	Maximum	same
Running Mean	Mean	same
Running Minimum	Minimum	same
Running RMS	RMS	same
Running Standard Deviation	Standard Deviation	same
Running Variance	Variance	same
Sample and Hold	same	same
Shift Register	same	same
Sign	same	Simulink
Signal From Workspace	same	same
Sort	same	same
Split	Complex to Real-Imag	Simulink
Standard Deviation	same	same
Submatrix	same	same
Time Varying FIR Filter	Time-Varying Direct-Form II Transpose Filter	same
Time Varying IIR Filter	Time-Varying Direct-Form II Transpose Filter	same
Time Vector Scope	Time Frame Scope	same
To Workspace	Signal To Workspace	same
Toeplitz	same	same
Transpose	same	same
Triggered Complex Matrix To Workspace	Triggered Matrix To Workspace	same
Triggered Complex To Workspace	Triggered Signal To Workspace	same
Triggered Matrix To Workspace	same	same
Triggered Shift Register	same	same
Triggered To Workspace	Triggered Signal To Workspace	same
Trigonometric Function	same	Simulink
Unbuffer	same	same
Unit Delay	Integer Delay	same
Unwrap	same	same
Upsample	same	same
Variable Fractional Delay	same	same
Variable Integer Delay	same	same
Variance	same	same
Width	same	Simulink
Window Function	same	same
Yule-Walker AR	Yule-Walker Method	Power Spectrum Estimation
Yule-Walker IIR Filter Design	same	same
Zero Pad	same	same

[ Previous | Next ]