Assignment: Digital Filter Design and Multirate Signal Processing
1. What are IIR Digital Filters?
Answer:
Infinite Impulse Response (IIR) digital filters are systems characterized by a recursive relat
where the output depends not only on the current and past input values but also on past output
values. These filters mimic analog filter designs and have the following properties:
- They require fewer coefficients than FIR filters to achieve the same frequency response.
- They are not inherently stable unless carefully designed.
- Examples include Butterworth, Chebyshev, and Elliptic filters.
2. What is the Impulse-Invariant Method?
Answer:
The impulse-invariant method is a technique to design IIR digital filters by mapping an analog
impulse response into the digital domain. The key steps are:
1. Begin with an analog filter's impulse response h_a(t).
2. Sample it at intervals of T to obtain h_d[n] = h_a(nT).
3. This preserves the time-domain response and ensures the sampled filter closely matches the
original.
However, this method can result in aliasing, especially at higher frequencies.
3. What is the Bilinear Transformation Method?
Answer:
The bilinear transformation is a method to convert analog filters into digital filters without
mapping the s-plane into the z-plane using the substitution:
s = (2 / T) * (1 - z^(-1)) / (1 + z^(-1))
Key features:
- Frequency warping occurs, and pre-warping is often applied to correct this.
- It ensures the stability of the analog filter is preserved in the digital domain.
4. What are the Specifications of IIR Filters?
Answer:
The typical specifications of an IIR filter include:
- Passband ripple (δp): Variation allowed in the passband.
- Stopband attenuation (δs): Minimum attenuation in the stopband.
- Cutoff frequency (fc): Frequency dividing passband and stopband.
- Transition band: Range where the filter transitions between passband and stopband.
- Order: Degree of the filter, influencing complexity and sharpness.
5. What are the Specifications of a Low-Pass (LP) Filter?
Answer:
For a low-pass filter:
- Passband Frequency (fp): Frequencies below this are in the passband.
- Stopband Frequency (fs): Frequencies above this are in the stopband.
- Passband Ripple (δp): Maximum allowable variation in the passband.
- Stopband Attenuation (δs): Minimum required attenuation in the stopband.
6. What are the Characteristics of Butterworth Filters?
Answer:
Butterworth filters are known for their flat frequency response in the passband, making them i
applications requiring minimal distortion. Key characteristics:
- Maximally flat magnitude response.
- No ripples in the passband or stopband.
- The transition band is not as steep as Chebyshev or elliptic filters.
7. What are the Structures of IIR Filters?
Answer:
IIR filters can be implemented using different structures:
1. Direct Form I and II: Simple implementation using coefficients directly.
2. Cascade Form: Decomposes the filter into a series of second-order sections for improved
stability.
3. Parallel Form: Represents the filter as a sum of parallel components.
8. What is Decimation in Multirate Signal Processing?
Answer:
Decimation reduces the sampling rate of a signal by an integer factor M. It involves:
- Low-pass filtering the signal to prevent aliasing.
- Subsampling by retaining every M-th sample.
9. What is Interpolation?
Answer:
Interpolation increases the sampling rate of a signal by an integer factor L. It involves:
- Inserting L-1 zeros between each sample.
- Applying a low-pass filter to remove spectral images.
10. How is Sampling Rate Conversion Performed by a Rational Factor?
Answer:
Sampling rate conversion by a rational factor L/M:
1. First interpolate the signal by L.
2. Then decimate it by M.
- Combined interpolation and decimation require appropriate low-pass filtering to prevent alia
11. What are the Computational Requirements of Multirate Processing?
Answer:
The computational load depends on:
- Filter Design: Low-pass filter requirements.
- Interpolation Factor L and Decimation Factor M: Higher factors increase complexity.
- Polyphase Decomposition: Reduces computational overhead by processing the signal in parallel
12. What is Polyphase Decomposition?
Answer:
Polyphase decomposition splits a filter into multiple sub-filters to improve efficiency in mul
- For interpolation, it divides the filter into L sub-filters.
- For decimation, it uses M sub-filters to process every M-th sample.
13. What is a Half-Band Filter?
Answer:
A half-band filter is a low-pass filter where:
- The transition band is symmetric around π/2.
- Approximately half the coefficients are zero, reducing computation.
14. What is a Uniform DFT Filter Bank?
Answer:
A uniform DFT filter bank divides a signal into multiple frequency bands using the Discrete Fo
- It uses polyphase implementation for efficiency.
- Each sub-band corresponds to a specific frequency range.
15. What is a Two-Channel QMF Bank?
Answer:
A two-channel Quadrature Mirror Filter (QMF) bank splits a signal into low-frequency and high-
- The low-pass and high-pass filters are mirror images in frequency.
- Useful for sub-band coding in audio and image processing.
16. How is Subband Coding Used in Speech and Audio Signals?
Answer:
Subband coding compresses signals by dividing them into multiple frequency bands and encoding
- Speech compression: Reduces data for transmission or storage.
- Audio coding: Improves quality while reducing bit rates, as used in MP3 and AAC.
End of Assignment
