COMD 5070 STUDY GUIDE 2026 EXAM QUESTIONS AND
SOLUTIONS RATED A+
✔✔Source filter - ✔✔-in theory, glottal source and vocal tract filter are independent
-F0 can change, causing harmonics to change while tract configuration remains
constant
-tongue can move, changing filter characteristics while F0 remains the same
✔✔Vocal tract transfer function - ✔✔The change in sound between entering the vocal
tract and the one leaving it (output minus input). It is the difference that the vocal tract
makes to the sound that left the larynx.
✔✔Formants - ✔✔-resonance peaks in transfer function
-some frequencies are amplified so that they are more prominent
-there can be many for any vowel
-they do not create the sound
-they shape or alter
-the first 2 or 3 are of primary interest
-the formants on an output spectrum identify a specific vowel
✔✔"Spillover effects" of source-tract interactions - ✔✔-in conditions such as Parkinson's
disease the result of hypokinetic dysarthria causes:
-weak, monotone voice
-imprecise consonant articulation
-treated by increasing loudness which naturally improves articulation
✔✔spasmodic dysphonia - ✔✔-a neurolaryngological disorder
-spasms cause strained, strangled voice
-medical treatment is laryngeal injections of botox, does not respond to speech therapy
✔✔muscle tension dysphonia - ✔✔-creates tight, strangled sounding voice
-excessive neck muscle activity
-treated by laryngeal massage
✔✔analog signals - ✔✔-continuous in time and amplitude
-no gaps so there are infinite points in time,
*infinite amplitude values,
*cannot be fully represented by a table of numbers
*only a graph can really represent them
✔✔Analog devices - ✔✔microphone/speakers
clocks or watches with hands
mercury thermometer
tape measure
cassette recorder
, dimmer light switch
✔✔Digital signals - ✔✔-discrete in time and amplitude
-numbers represent snapshots over time
-numbers have finite precision (limited number of decimal places)
-can be fully represented by a table of numbers
-unknowns between the points
✔✔Digital devices - ✔✔digital clock
computer
CD
✔✔Sound conversion - ✔✔ADC - analog to digital conversion
DAC - digital to analog conversion
PC sound cards
✔✔Sampling rate - ✔✔-snapshots that represent amplitude values
-the more samples the better
-sample rate specified in Hz
✔✔Sampling rate pros and cons - ✔✔-higher rate gives better fidelity
-higher rate requires bigger files (i.e. more memory usage, disk space used, processing
time)
✔✔Nyquist frequency - ✔✔Represents the highest frequency that a particular recording
can reproduce. In order to get an accurate sample, you need to decide which
frequencies you're interested in and sample twice that frequency. So Nyquist frequency
is half the sample rate.
✔✔Obtaining good recording data - ✔✔-a quiet environment reduces reverberation
-use good quality microphone
-make sure the signal strength has good saturation and is not clipping
-use proper sample rate, err on the side of detail
✔✔Sound and movement - ✔✔-all sounds originate with movement
-movement characteristics determine the nature of the sound
-speech production has many degrees of freedom, i.e., there are many individual
structures/movements involved
✔✔Filtering - ✔✔-speech sounds contain many frequencies
-we can be selective about which ones we are going to keep and analyze.
✔✔Filter types - ✔✔high pass - allows high frequencies through, keeps out lower
frequencies
low pass - allows low frequencies through, keeps out higher frequencies
SOLUTIONS RATED A+
✔✔Source filter - ✔✔-in theory, glottal source and vocal tract filter are independent
-F0 can change, causing harmonics to change while tract configuration remains
constant
-tongue can move, changing filter characteristics while F0 remains the same
✔✔Vocal tract transfer function - ✔✔The change in sound between entering the vocal
tract and the one leaving it (output minus input). It is the difference that the vocal tract
makes to the sound that left the larynx.
✔✔Formants - ✔✔-resonance peaks in transfer function
-some frequencies are amplified so that they are more prominent
-there can be many for any vowel
-they do not create the sound
-they shape or alter
-the first 2 or 3 are of primary interest
-the formants on an output spectrum identify a specific vowel
✔✔"Spillover effects" of source-tract interactions - ✔✔-in conditions such as Parkinson's
disease the result of hypokinetic dysarthria causes:
-weak, monotone voice
-imprecise consonant articulation
-treated by increasing loudness which naturally improves articulation
✔✔spasmodic dysphonia - ✔✔-a neurolaryngological disorder
-spasms cause strained, strangled voice
-medical treatment is laryngeal injections of botox, does not respond to speech therapy
✔✔muscle tension dysphonia - ✔✔-creates tight, strangled sounding voice
-excessive neck muscle activity
-treated by laryngeal massage
✔✔analog signals - ✔✔-continuous in time and amplitude
-no gaps so there are infinite points in time,
*infinite amplitude values,
*cannot be fully represented by a table of numbers
*only a graph can really represent them
✔✔Analog devices - ✔✔microphone/speakers
clocks or watches with hands
mercury thermometer
tape measure
cassette recorder
, dimmer light switch
✔✔Digital signals - ✔✔-discrete in time and amplitude
-numbers represent snapshots over time
-numbers have finite precision (limited number of decimal places)
-can be fully represented by a table of numbers
-unknowns between the points
✔✔Digital devices - ✔✔digital clock
computer
CD
✔✔Sound conversion - ✔✔ADC - analog to digital conversion
DAC - digital to analog conversion
PC sound cards
✔✔Sampling rate - ✔✔-snapshots that represent amplitude values
-the more samples the better
-sample rate specified in Hz
✔✔Sampling rate pros and cons - ✔✔-higher rate gives better fidelity
-higher rate requires bigger files (i.e. more memory usage, disk space used, processing
time)
✔✔Nyquist frequency - ✔✔Represents the highest frequency that a particular recording
can reproduce. In order to get an accurate sample, you need to decide which
frequencies you're interested in and sample twice that frequency. So Nyquist frequency
is half the sample rate.
✔✔Obtaining good recording data - ✔✔-a quiet environment reduces reverberation
-use good quality microphone
-make sure the signal strength has good saturation and is not clipping
-use proper sample rate, err on the side of detail
✔✔Sound and movement - ✔✔-all sounds originate with movement
-movement characteristics determine the nature of the sound
-speech production has many degrees of freedom, i.e., there are many individual
structures/movements involved
✔✔Filtering - ✔✔-speech sounds contain many frequencies
-we can be selective about which ones we are going to keep and analyze.
✔✔Filter types - ✔✔high pass - allows high frequencies through, keeps out lower
frequencies
low pass - allows low frequencies through, keeps out higher frequencies
