09-07-2012, 09:59 AM
Filter Design Toolbox
Filter Design Toolbox.pdf (Size: 5.44 MB / Downloads: 74)
What Is Filter Design Toolbox?
Filter Design Toolbox is a collection of tools built on top of the MATLAB®
computing environment and the Signal Processing Toolbox. The toolbox
includes a number of advanced filter design techniques that support designing,
simulating, and analyzing fixed-point and custom floating-point filters for a wide range of precisions.
Filter Design Functions in the Toolbox
In a system that has unlimited power and size, any filter structure that met
your performance specifications would do. You would design a floating-point
filter whose frequency response achieved your aims and implement that filter
in your system.
When you need a fixed-point filter to meet your requirements, the filter
structure you choose can depend very much on how quantization affects the
performance of the filter. Filter Design Toolbox offers both FIR and IIR filter
design tools and structures that let you experiment with multiple filter designs
to see how each responds to quantization effects.
Quantization Functions in the Toolbox
Designing floating-point filters solves only part of the filter design problem. In
most cases, floating-point filter realizations are not appropriate for digital
signal processing applications. Many real-world DSP systems require that
their filters use minimum power, generate minimum heat, and do not induce
computational overload in their processors. Meeting these constraints often
means using fixed-point filters. Unfortunately, converting a floating-point
filter to fixed-point realization (called quantizing) can result in lost filter
performance and accuracy. To simulate and determine the effects of
quantization, and allow you to investigate how switching from floating-point to
fixed-point arithmetic affects the performance of your filter, the toolbox
includes quantization functions. You use the toolbox quantization functions for
constructing, applying, and analyzing quantizers, quantized filters, and
quantized fast Fourier transforms (FFT).
Data Quantizers
To determine how quantization affects a signal, you construct quantizers that
you use to quantize a signal or data set in MATLAB. By adjusting the
quantization parameters of your quantizer, you can investigate the output
from various quantization schemes when you apply them to a data set or
signal. In addition to experimenting with data quantization, quantizers
determine how quantized filters and quantized FFTs quantize data to which
they are applied.
Quantized Fast Fourier Transforms
In developing digital signal processing (DSP) algorithms, the fast Fourier
transform (FFT) is one of the essential building blocks. It may be the most
common transform for handling data and signals. To implement an FFT on a
fixed-point DSP, you must consider the effects of word length on the output of
the transform, in much the same way that you must consider the quantization
effects in a digital filter.