DSP Blockset

Filter Realization Wizard

Construct filter realizations using Sum, Gain, and Integer Delay blocks

Library

Filtering / Filter Designs

Description

Note    Use this block to implement fixed-point or floating-point digital filters built from Sum, Gain, and Integer Delay blocks. You can either design a filter by using the block's filter design and analysis parameters, or import the coefficients of a filter you have designed elsewhere. The following blocks also implement digital filters, but serve slightly different purposes: Digital Filter -- Use to efficiently implement floating-point filters that you have already designed Digital Filter Design -- Use to design, analyze, and then efficiently implement floating-point filters.

The Filter Realization Wizard is a tool for automatically implementing a digital filter. You must specify a filter, its structure, and the data types for the filter's inputs, outputs, and computations. The filter can support double-precision, single-precision, or fixed-point data types.

The Filter Realization Wizard creates a subsystem block that implements the specified filter using Sum, Gain, and Integer Delay blocks. To see the filter implementation, double-click the subsystem block. The subsystem block applies the specified filter to any sample-based input signal, or any frame-based row vector signal, and outputs the result.

The parameters of the Filter Realization Wizard are a part of a larger GUI, the Filter Design and Analysis Tool (FDATool), from the Signal Processing Toolbox. You can use all of the powerful tools in FDATool to design and analyze your filter, and then use the Filter Realization Wizard parameters to implement the filter in your models.

To learn how to use the Filter Realization Wizard, see other sections of this reference page.

Sections of This Reference Page

• Valid Inputs and Corresponding Outputs
• Steps to Implementing a Filter with This Block
• Specifying the Filter and Its Data Type Support
• Setting the Filter Structure and Number of Filter Sections
• Setting Where to Put the Filter
• Optimizing the Filter Structure
• Setting the Data Type of the Filter Implementation
• Corresponding Method for dfilt and qfilt
• Dialog Box
• References
• Supported Data Types
• See Also

Valid Inputs and Corresponding Outputs

The Filter Realization Wizard creates a new a subsystem block that implements the specified filter. The subsystem block applies the specified filter to an input signal, and outputs the result.

Valid Inputs.   The subsystem block accepts inputs that are sample-based vectors and matrices, or frame-based row vectors.

Corresponding Outputs.   The output of the subsystem block has the same dimensions and frame status as the input.

What Is Considered an Independent Channel.   The subsystem block treats each element of a vector or matrix as an independent channel.

Steps to Implementing a Filter with This Block

The parameters for this block are in a panel of the Filter Design and Analysis Tool (FDATool) GUI. In addition to using FDATool's Filter Realization Wizard parameters in the Realize Model panel, you must also use other panels to specify your filter. (To access the different panels, use the sidebar buttons in the lower-left corner of FDATool.)

To implement a filter using the Filter Realization Wizard, you must do the following:

1. Open the filter realization wizard by typing dspfwiz, or double-clicking the block in the Filter Designs library or in a model.
2. Specify a filter (including its data type support) by either designing a filter using the Design Filter panel, or by importing a filter using the Import Filter panel. To learn how, see Specifying the Filter and Its Data Type Support.
3. Optionally change the default filter structure and the default number of filter sections. To learn how, see Setting the Filter Structure and Number of Filter Sections.
4. Do the following in the Realize Model panel:
• Select the destination and the name of the filter subsystem block, and whether to overwrite a filter created previously by the Filter Realization Wizard. For details, see Setting Where to Put the Filter.
• Select the filter structure optimizations. To learn more, see Optimizing the Filter Structure.
• Set the data type of the input, output, and computations in the filter implementation to match those of the specified filter. To learn how, see Setting the Data Type of the Filter Implementation.
5. Click Realize model in the Realize Model panel to realize the filter. A new block appears in a specified model. Double-click the new block to see the filter realization, as illustrated in the following figure.
   
   

Specifying the Filter and Its Data Type Support

To specify a purely double-precision filter, you can either design a filter using the Design Filter panel, or import a filter using the Import Filter panel. (You can import dfilt filter objects as well as vectors of filter coefficients designed using functions in Signal Processing Toolbox and Filter Design Toolbox.)

You can also specify a fixed-point filter, a single-precision filter, or a filter with different input, output, and computation data types. You can specify such filters by using the Set Quantization Parameters panel, or by importing a qfilt filter object with the Import Filter panel. Both of these options require the Filter Design Toolbox.

To learn how to ensure that your filter implementation reflects the data types of the filter you specify, see Setting the Data Type of the Filter Implementation.

Note    Running a model containing implementations of non-double-precision filters requires the Fixed-Point Blockset, but you can still edit models containing such filter implementations without the Fixed-Point Blockset. For more information, see the topic on licensing information in the Fixed-Point Blockset documentation.

See the following topics to learn how to use the panels to specify your filter:

• Design Filter -- Topic on the Filter Design and Analysis Tool (FDATool) in the Signal Processing Toolbox documentation.
• Import Filter -- Topic on importing a filter design in the Signal Processing Toolbox documentation.
• Set Quantization Parameters -- Topic on quantizing filters in the Filter Design and Analysis Tool (FDATool) in the Filter Design Toolbox documentation.

To open a panel, click the appropriate button in the lower-left corner of FDATool.

Setting the Filter Structure and Number of Filter Sections

The Current Filter Information region of FDATool shows the structure and the number of second-order sections in your filter.

Change the filter structure and number of filter sections of your filter as follows:

• Select Convert Structure from the Edit menu to open the Convert Structure dialog box. For details, see the topic on converting to new filter structures in the Signal Processing Toolbox documentation.
• Select Convert to Second-order Sections from the Edit menu to open the Convert to SOS dialog box. For details, see the topic on converting to second-order sections in the Signal Processing Toolbox documentation.

The Filter Realization Wizard supports the following structures:

Direct form I Direct form II Direct form I transposed Direct form II transposed Second order sections for direct form I and II, and their transposes Direct form FIR Direct form FIR transposed Direct form antisymmetric FIR Direct form symmetric FIR

Lattice ARMA Lattice AR Lattice MA (same as lattice minimum phase) Lattice all-pass Lattice maximum phase Cascade Parallel

Note    You may not be able to directly access some of the supported structures through the Convert Structure dialog of FDATool. However, you can access all of the structures by creating a qfilt or dfilt filter object with the desired structure, and then importing the filter into FDATool. (To learn more about the Import Filter panel, see the topic on importing a filter design in the Signal Processing Toolbox documentation.)

Setting Where to Put the Filter

The Filter Realization Wizard creates a subsystem block that implements the specified filter. Set where the Filter Realization Wizard puts the subsystem block by doing the following:

1. Open the Realize Model panel in FDATool by clicking the Realize Model button in the lower-left corner of FDATool.
2. Set the Destination parameter to one of the following:
• New model -- Places the subsystem block in a new model.
• Current model -- Places the subsystem block in the most recently selected model.
3. In the Block Name parameter, type a name for the subsystem block.
4. If you set the Destination parameter to Current model, the Overwrite generated `Filter' block parameter becomes enabled, which has the following behavior:
• Select the Overwrite generated `Filter' block check box -- If there is a previously implemented filter subsystem in the current model that has the name specified in the Overwrite block parameter, that subsystem is overwritten by the currently specified filter. If there is no block to overwrite in the current model, the Filter Realization Wizard creates a new filter subsystem with the specified name.
• Clear the Overwrite generated `Filter' block check box -- The Filter Realization Wizard creates a new filter subsystem in the current model without overwriting any previously implemented filters. If there already exists a block in the current model with the name specified in the Overwrite block parameter, the Filter Realization Wizard appends the next available number to the new subsystem's name. For example, if you specified the name myFilter, and there are blocks in the model named myFilter, myFilter1, and myFilter2, your new filter subsystem is named myFilter3.

  

Optimizing the Filter Structure

The Filter Realization Wizard creates a subsystem block that implements the specified filter using Sum, Gain, and Integer Delay blocks. To optimize the filter implementation, (for instance, by removing unity gains), do the following:

1. Open the Realize Model panel in FDATool by clicking the Realize Model button in the lower-left corner of FDATool .
2. Select the desired optimizations in the Optimization region of the Realize Model pane. See the following descriptions and illustrations of each optimization option.
   
   

• Optimize for zero gains -- Remove zero-gain paths.
• Optimize for unity gains -- Substitute gains equal to one with a wire (short circuit).
• Optimize for -1 gains -- Substitute gains equal to -1 with a wire (short circuit), and change the corresponding sums to subtractions.
• Optimize delay chains -- Substitute any delay chain made up of n unit delays with a single delay by n.

  

Setting the Data Type of the Filter Implementation

The filter you specify can have various data types for its inputs, outputs, and computations, as described in Specifying the Filter and Its Data Type Support.

To ensure that the Filter Realization Wizard's filter implementation accurately reflects the data type(s) you specified for your filter, do the following:

1. Open the Realize Model pane by clicking the Realize Model button in the lower-left corner of FDATool .
2. Appropriately set the parameter in the Block Type region of the Realize Model pane (see the following table and figures). The default setting of the parameter depends on the filter, and is always the most appropriate setting for the specified filter.

 

Appropriate Use of Block Type Settings

Block Type	When to Use	Blocks and Data Types Used in Implementation
Floating-point blocks	Only when implementing purely double-precision filters as described in Specifying the Filter and Its Data Type Support.	Sum, Gain, and Integer Delay blocks configured in the specified filter structure, and set to make double-precision computations. See the figure called Implementations of Double-Precision and Fixed-Point Filters.
Fixed-point blocks	When implementing any of the following as described in Specifying the Filter and Its Data Type Support: Fixed-point filters Single-precision filters Filters with different input, output, and computation data types	Sum, Gain, and Integer Delay blocks configured in the specified filter structure, and set to compute using the appropriate data types Gateway In block for the input quantizer Conversion blocks for the multiplicand quantizers, and for the output quantizer See the following note and the figure called Implementations of Double-Precision and Fixed-Point Filters.

Note    The filter implementation that results from the Fixed-point blocks setting described above contains blocks from the Fixed-Point Blockset, which you must install to run the filter in simulations. You can still edit the blocks used to implement the filter without installing the Fixed-Point Blockset. For more information, see the topic on licensing information in the Fixed-Point Blockset documentation.

Implementations of Double-Precision and Fixed-Point Filters

Corresponding Method for dfilt and qfilt

The dfilt (digital filter) object in Signal Processing Toolbox and the qfilt (quantized filter) object in Filter Design Toolbox both have a method, realizemdl, that allows you to access the capabilities of the Filter Realization Wizard from the command line.

For more information about the realizemdl method, see the following topics:

• The topic on methods in the dfilt reference page in the Signal Processing Toolbox documentation
• The realizemdl reference page in the Filter Design Toolbox documentation

Dialog Box

Note    The following parameters for the Filter Realization Wizard are in the Realize Model panel of the Filter Design and Analysis Tool (FDATool) GUI. To open different panels of FDATool, click the different buttons at the lower-left corner. For more information about relevant panels, see Specifying the Filter and Its Data Type Support.

Destination

The location where the new filter block should be created: in a new model, or in the current (most recently selected) model.

Block name

The name of the new filter block.

Overwrite generated `Filter' block

When selected, the block overwrites any filter block in the current model with the name specified in the Block name parameter. Enabled when the Destination parameter is set to Current model. For more information, see Setting Where to Put the Filter.

Block Type

Determines the data type support of the filter implementation. Set to Floating-point blocks when implementing purely double-precision filters. Otherwise, set to Fixed-point blocks. For more information, see Setting the Data Type of the Filter Implementation.

Optimize for zero gains

When selected, the block removes zero-gain paths from the filter structure. For an example, see Optimizing the Filter Structure.

Optimize for unity gains

When selected, the block substitutes gains equal to one with a wire (short circuit). For an example, see Optimizing the Filter Structure.

Optimize for -1 gains

When selected, the block substitutes gains equal to -1 with a wire (short circuit), and changes the corresponding sums to subtractions. For an example, see Optimizing the Filter Structure.

Optimize delay chains

When selected, the block substitutes any delay chains made up of n unit delays with a single delay by n. For an example, see Optimizing the Filter Structure.

Realize Model

Click to create a subsystem block that implements the specified filter using Sum, Gain, and Integer Delay blocks. To see the filter implementation, double-click the subsystem block. The subsystem block applies the specified filter to any sample-based input signal or frame-based row vector signal, and outputs the result.

Note    For more information about relevant parameters in other panels of FDATool, see Specifying the Filter and Its Data Type Support.

References

Oppenheim, A. V. and R. W. Schafer. Discrete-Time Signal Processing. Englewood Cliffs, NJ: Prentice Hall, 1989.

Proakis, J. and D. Manolakis. Digital Signal Processing. 3rd ed. Englewood Cliffs, NJ: Prentice-Hall, 1996.

Supported Data Types

• Double-precision floating point
• Single-precision floating point -- Supported only when you install the Filter Design Toolbox and Fixed-Point Blockset
• Fixed point -- Supported only when you install the Filter Design Toolbox and Fixed-Point Blockset

To learn how to convert your data types to the above data types in MATLAB and Simulink, see Supported Data Types and How to Convert to Them.

See Also

Digital Filter	DSP Blockset
Digital Filter Design	DSP Blockset
qfilt	Filter Design Toolbox
filter	Filter Design Toolbox
realizemdl	Filter Design Toolbox
dfilt	Signal Processing Toolbox
filter	Signal Processing Toolbox

• Filters -- Examples of when and how to use DSP Blockset filtering blocks
• Choosing Between Filter Design Blocks

FFT

FIR Decimation