COMD 5070 EVALUATION EXAM 2026 QUESTIONS AND
SOLUTIONS RATED A+
✔✔What is quantization - ✔✔the process of converting a voltage, which is analog, into
a number, which is digital (it creates a number that reflects a signal amplitude, one
number is a 'snap shot')
✔✔What is a sampling rate? - ✔✔is how often you take a 'snap shot', sample rate is in
Hz, 10 samples per second = 10 Hz
✔✔What is a Fourier transforms/analysis? - ✔✔Analyze it in such a way that it creates a
spectrum which shows what the individual components are from the time domain
waveform. A Fourier transform will give you a frequency domain display (spectrum)
✔✔What does a spectrum show? - ✔✔A complex sound has many ingredients like a
cake. The spectrum show how much of each ingredient is present
✔✔What are the filter types? - ✔✔1. high pass filter 2. low pass filter 3. band pass filter
4. band reject filter
✔✔What does a time domain display show? - ✔✔is the regular display that shows the
amplitude on the y-axis and time on the x-axis that represents a sound waveform.
✔✔What does a frequency domain display show? - ✔✔has the amplitude on the y-axis,
but instead of time on the x-axis it shows the frequency (of how much) of a frequency
there is in a signal. The frequency domain display looks similar to a bar graph.
✔✔What does a time domain display show us about sound? - ✔✔shows the amplitude
of a sound wave over time.
✔✔What does a frequency domain display show us about sound? - ✔✔shows how
much of each ingredient (different sine waves) are present in a given sound.
✔✔What are the 3 dimensions in a 3D spectrogram? - ✔✔1. x-axis is time 2. y-axis is
frequency 3. darkness indicates intensity (darker is stronger)
✔✔What are the types of spectra? - ✔✔line, FFT, LPC
✔✔fast Fourier transform (FFT) - ✔✔clearly shows harmonic energy, each peak is
harmonic, less clear at showing formants, more revealing of the source.
✔✔What is a formant? - ✔✔the characteristic resonances of the vocal tract.
, ✔✔linear prodictive codeing (LPC) - ✔✔shows spectral envelope, good at revealing
formants, does not show harmonics, more revealing of filter (vocal tract)
✔✔What does motor equivalence mean? - ✔✔the same sound can be produced
several ways, acoustics cannot reveal all details of movement
✔✔empirical means: - ✔✔that it is based on data
✔✔deterministic means: - ✔✔that it obeys physical laws (not random)
✔✔predictive means: - ✔✔if you do this ... then that will happen
✔✔parsimonious means: - ✔✔the simplest explanation possible is the best one
✔✔what does nyquist mean: - ✔✔this is half of the sample rate (the highest frequency
you can reproduce).
✔✔quantization - ✔✔is the process of converting a voltage, which is analog, into a
number, which is digital (it creates a number that reflects a signal amplitude, one
number is a 'snap shot' more decimal places for more detail.
✔✔sampling rate - ✔✔is how often you take a snapshot, it is in Hz, 10 samples per
second = 10 Hz
✔✔Why are speech acoustic measures limited in how well they reflect specific speech
movements in the vocal tract? - ✔✔acoustic patterns reflect vocal tract movements with
some ambiguity, motor equivalence: the same sound can be produced several ways,
acoustics cannot reveal all details of movement
✔✔What is high speed imaging? - ✔✔is real life slow motion version of filming
✔✔stroboscopy - ✔✔is a simulation (or optical illusion, animation) of a slow motion of
the vocal folds [the fundamental frequency must be steady for the timing to work
otherwise it will be blurry]
✔✔types of vocal perturbation - ✔✔jitter and shimmer
✔✔jitter - ✔✔when each successive cycle deviates from this 10 millisecond average by
a tiny amount (a few microseconds) these differences are what we call jitter. So jitter is
a frequency perturbation [jitter = frequency perturbation]
✔✔shimmer - ✔✔is a random cycle-to-cycle difference in amplitude [shimmer =
amplitude perturbation or amplitude instability]
SOLUTIONS RATED A+
✔✔What is quantization - ✔✔the process of converting a voltage, which is analog, into
a number, which is digital (it creates a number that reflects a signal amplitude, one
number is a 'snap shot')
✔✔What is a sampling rate? - ✔✔is how often you take a 'snap shot', sample rate is in
Hz, 10 samples per second = 10 Hz
✔✔What is a Fourier transforms/analysis? - ✔✔Analyze it in such a way that it creates a
spectrum which shows what the individual components are from the time domain
waveform. A Fourier transform will give you a frequency domain display (spectrum)
✔✔What does a spectrum show? - ✔✔A complex sound has many ingredients like a
cake. The spectrum show how much of each ingredient is present
✔✔What are the filter types? - ✔✔1. high pass filter 2. low pass filter 3. band pass filter
4. band reject filter
✔✔What does a time domain display show? - ✔✔is the regular display that shows the
amplitude on the y-axis and time on the x-axis that represents a sound waveform.
✔✔What does a frequency domain display show? - ✔✔has the amplitude on the y-axis,
but instead of time on the x-axis it shows the frequency (of how much) of a frequency
there is in a signal. The frequency domain display looks similar to a bar graph.
✔✔What does a time domain display show us about sound? - ✔✔shows the amplitude
of a sound wave over time.
✔✔What does a frequency domain display show us about sound? - ✔✔shows how
much of each ingredient (different sine waves) are present in a given sound.
✔✔What are the 3 dimensions in a 3D spectrogram? - ✔✔1. x-axis is time 2. y-axis is
frequency 3. darkness indicates intensity (darker is stronger)
✔✔What are the types of spectra? - ✔✔line, FFT, LPC
✔✔fast Fourier transform (FFT) - ✔✔clearly shows harmonic energy, each peak is
harmonic, less clear at showing formants, more revealing of the source.
✔✔What is a formant? - ✔✔the characteristic resonances of the vocal tract.
, ✔✔linear prodictive codeing (LPC) - ✔✔shows spectral envelope, good at revealing
formants, does not show harmonics, more revealing of filter (vocal tract)
✔✔What does motor equivalence mean? - ✔✔the same sound can be produced
several ways, acoustics cannot reveal all details of movement
✔✔empirical means: - ✔✔that it is based on data
✔✔deterministic means: - ✔✔that it obeys physical laws (not random)
✔✔predictive means: - ✔✔if you do this ... then that will happen
✔✔parsimonious means: - ✔✔the simplest explanation possible is the best one
✔✔what does nyquist mean: - ✔✔this is half of the sample rate (the highest frequency
you can reproduce).
✔✔quantization - ✔✔is the process of converting a voltage, which is analog, into a
number, which is digital (it creates a number that reflects a signal amplitude, one
number is a 'snap shot' more decimal places for more detail.
✔✔sampling rate - ✔✔is how often you take a snapshot, it is in Hz, 10 samples per
second = 10 Hz
✔✔Why are speech acoustic measures limited in how well they reflect specific speech
movements in the vocal tract? - ✔✔acoustic patterns reflect vocal tract movements with
some ambiguity, motor equivalence: the same sound can be produced several ways,
acoustics cannot reveal all details of movement
✔✔What is high speed imaging? - ✔✔is real life slow motion version of filming
✔✔stroboscopy - ✔✔is a simulation (or optical illusion, animation) of a slow motion of
the vocal folds [the fundamental frequency must be steady for the timing to work
otherwise it will be blurry]
✔✔types of vocal perturbation - ✔✔jitter and shimmer
✔✔jitter - ✔✔when each successive cycle deviates from this 10 millisecond average by
a tiny amount (a few microseconds) these differences are what we call jitter. So jitter is
a frequency perturbation [jitter = frequency perturbation]
✔✔shimmer - ✔✔is a random cycle-to-cycle difference in amplitude [shimmer =
amplitude perturbation or amplitude instability]
