07 - DSP Applications

Question 1

What is an algorithm?

Answer 1

A series of trigonometric and arithmetic operations; a series of steps used in digital processing

Question 2

Name one class of algorithm

Answer 2

Spectral analysis
Digital filtering
Coding and compressing data
Noise reduction

Question 3

Describe analog filters and digital filters

Answer 3

Analog - electronic circuits built with resistors, capacitors, and possibly transistors

Digital - algorithms that process the sampled and quantized data values

Question 4

Name 2 advantages of digital filters

Answer 4

Flexibility
Low Noise
Programmability

Question 5

What are 3 types of simple digital filters?

Answer 5

Simple gain filter: y(n)=Kx(n)
(Simple scaling algorithm independent of frequency)

Two-term difference filter: y(n)=x(n)-x(n-1)
(Acts as high pass filter because high frequencies have big differences between n and n-1 => preserved; low frequencies = low diff between n and n-1 => attenuated; any time you’re smoothing, it’s acting like a low pass filter)

Three-term average filter: y(n)=(x(n)+x(n-1)+x(n-2))/3
(Acts as low pass filter because high variation => more attenuation. E.g. 0, + 1, 0 = average of 0.33 (lost 66%), whereas lower frequency: 0, 0.5, 1 = average of 0.5 => only lost 50%)

Question 6

Name 2 filter parameters

Answer 6

Centre frequency and bandwidth

Attenuation/roll-off rate

Question 7

How many cutoff frequencies does a band pass filter have?

Answer 7

2

Question 8

The cutoff frequency is defined as where the relative amplitude has dropped by ___ dB

Answer 8

3 dB

Question 9

What is the formula to find the bandwidth?

Answer 9

Bandwidth = Fu-FL

Upper cutoff frequency - lower cutoff frequency

Question 10

What does the roll-off rate represent?

Answer 10

The slope/attenuation

Represented at the decrease in dB/octave

Question 11

What are the four types of filters?

Answer 11

Low pass, high pass, band pass, and band stop (notch is extreme band stop filter)

Question 12

What is a filter bank?

Answer 12

A bank of filters…..that’s it…

Just kidding! It’s a bank of filters that divide the input signal into separate frequency regions, each of which can be processed independently (e.g different gains or attenuation. Used with treble and bass, or to personalize a hearing aid)

Question 13

Narrower filters (e.g. 16-band filterbank compared to a 4-band filterbank) allow for _____ (finer/coarser) frequency axis resolution

Answer 13

Finer - can attenuate one small region (channel) while leaving the others intact. This can be used for identifying frequencies where there’s more noise and attenuating them)
-based on finding the appropriate set of coefficients

Question 14

True or False: In a filterbank, all of the filters must have the same bandwidth

Answer 14

False - can have differing bandwidths

Question 15

Name 2 uses for a filterbank

Answer 15

• processing (amplifying, compressing, etc) different frequency regions individually
• remove noise by attenuating channels with low signal-to-noise ratio
• use envelopes of each channel in cochlear implant speech processing
• provide smoothed spectrum (1 or 1/3 octave) measures of sound level across frequency in sound level meters

Question 16

What does the “Fast” part of “Fast Fourier Transform” refer to?

Answer 16

The fact that it can be computed in real time

Question 17

How are digital signals transformed into the frequency domain?

Answer 17

Using the Discrete Fourier Transform

Question 18

What does Fourier theorem state?

Answer 18

All complex waves can be expressed as a combination of sinusoidal components of different amplitudes, phases, and frequencies

Question 19

Using digitization, we can run the Fast Fourier Transform (FFT) and view the resulting __________

Answer 19

Spectrum

Question 20

True or False: DSP methods can detect the presence of voicing in a speech signal and extract the pitch period

Answer 20

True

Question 21

Time-domain signal averaging is especially important for what type of assessment?

Answer 21

ABR measurements

Question 22

A transfer function conveys information on how a system changes the content in the input signal. What are the 2 components of the transfer function?

Answer 22

The MAGNITUDE spectrum (how the system changes the amplitude; mainly use this spectrum with hearing aids)

The PHASE spectrum

Question 23

Two examples of transfer functions would be:

Answer 23

Loudspeaker measurements

Hearing aids

Question 24

What is one commonly used signal processing technique?

Answer 24

FFT for spectral analysis and transfer function measurements

Digital filtering and filterbanks for frequency shaping, compression, and noise reduction