What does the discrete Fourier transform (DFT) do?
Show answer & solution
Answer: (B) Converts a finite sequence of samples into frequency-domain components
The FFT is a fast algorithm for the DFT.
Electrical Engineering · Communication & Signals
30 questions, each with the answer and a worked solution.
Practice interactively and track your score →What does the discrete Fourier transform (DFT) do?
Answer: (B) Converts a finite sequence of samples into frequency-domain components
The FFT is a fast algorithm for the DFT.
What is the main advantage of the fast Fourier transform (FFT)?
Answer: (D) It reduces DFT computations from about N² to about N log₂N
For N = 1024, operations fall from about a million to about ten thousand.
Which filter has an impulse response that becomes exactly zero after a finite number of samples?
Answer: (C) FIR filter
FIR stands for finite impulse response.
What is a key advantage of FIR filters?
Answer: (D) They are always stable and can have exactly linear phase
FIR filters have no feedback paths.
What is an IIR filter?
Answer: (A) A recursive digital filter whose impulse response continues indefinitely because of feedback
IIR filters achieve sharp responses with fewer coefficients.
What is a disadvantage of IIR filters?
Answer: (B) They can become unstable and generally have non-linear phase
Careful design keeps poles inside the unit circle.
What is the z-transform used for?
Answer: (B) Analysing discrete-time signals and systems
It is the discrete counterpart of the Laplace transform.
What is the purpose of window functions such as the Hamming window in DSP?
Answer: (C) To reduce spectral leakage when analysing finite-length signals
Windows taper the edges of the data block.
What is quantisation in digital signal processing?
Answer: (A) Approximating sample values to a finite set of levels
Quantisation introduces a small error called quantisation noise.
Which arithmetic operation do DSP processors perform very efficiently in a single cycle?
Answer: (D) Multiply–accumulate (MAC)
Filters and FFTs rely heavily on MAC operations.
Which method is commonly used to convert an analog filter design into a digital IIR filter?
Answer: (C) Bilinear transformation
It maps the s-plane to the z-plane without aliasing.
What is decimation in DSP?
Answer: (D) Reducing the sampling rate by keeping every Mth sample after filtering
A low-pass filter prevents aliasing before decimation.
What is interpolation in DSP?
Answer: (A) Increasing the sampling rate by inserting and estimating new samples
Upsampling is followed by filtering.
What is spectral leakage?
Answer: (B) Spreading of signal energy into adjacent frequency bins when a non-integer number of cycles is analysed
Windowing reduces leakage.
What is the frequency resolution of an N-point DFT with sampling rate fs?
Answer: (B) fs/N
More points give finer resolution.
What is the frequency resolution of a 1,000-point DFT sampled at 10 kHz?
Answer: (B) 10 Hz
Δf = 10,000/1,000.
What is a moving-average filter?
Answer: (A) An FIR filter that outputs the average of the latest N samples
It smooths noise but blurs sharp changes.
Where must the poles of a stable digital filter lie?
Answer: (C) Inside the unit circle of the z-plane
Poles on or outside the circle cause instability.
What is the main purpose of an anti-aliasing filter before an ADC?
Answer: (D) To remove signal components above half the sampling frequency
It is an analog low-pass filter.
What is digital convolution used for in filters?
Answer: (D) Computing the filter output from input samples and filter coefficients
y[n] = Σ h[k] x[n − k].
What is zero padding in FFT analysis?
Answer: (A) Adding zeros to a data block to obtain a finer-looking spectrum interpolation
It does not add new information about the signal.
What is the main advantage of digital filters over analog filters?
Answer: (D) Precise, stable characteristics that do not drift with temperature or ageing
Digital filters are easily reprogrammed.
What is the number of complex multiplications needed for a direct 1,024-point DFT approximately?
Answer: (B) About one million
N² = 1,048,576; the FFT needs only about 5,000.
What is the transfer function of a discrete system in the z-domain?
Answer: (B) The ratio of the z-transform of output to that of input
Poles and zeros are plotted in the z-plane.
What is the unit delay operator in the z-domain?
Answer: (C) z⁻¹
Each z⁻¹ represents a delay of one sample.
What does a digital low-pass filter do?
Answer: (D) Passes low-frequency components and attenuates high-frequency components of a sampled signal
It is used for smoothing and anti-alias decimation.
What is the difference equation of an FIR filter?
Answer: (A) y[n] = Σ bₖ x[n − k], using only present and past inputs
IIR filters also use past outputs.
What is an adaptive filter?
Answer: (B) A filter whose coefficients adjust automatically to minimise an error signal
The LMS algorithm is commonly used for echo and noise cancellation.
What is the power spectral density of a signal?
Answer: (D) The distribution of signal power over frequency
It is the Fourier transform of the autocorrelation.
Why is the FFT size usually chosen as a power of two?
Answer: (C) Radix-2 algorithms are simplest and most efficient for such lengths
Mixed-radix FFTs handle other sizes.