CSD 303 Final Exam || 100% correct.
Convert between seconds and milliseconds correct answers Divide seconds by 1000
Wave's frequency and period correct answers Inverse of each other
Periodic waves correct answers repeating same waveform
Simple periodic (sine) wave correct answers pure tone: only one frequency
Match time domain and frequency domain signals correct answers Find period of a wave from time domain, then take the inverse and match to frequency domain
Add 2 sine waves correct answers Share period of the slower wave, so pick an interval to repeatedly find the average amplitude between the two waves, plot that amplitude at every interval (over 4 seconds, measure the average amplitude of the waves every .5 second, draw line connecting dots)
Frequency domain of 2 sine waves correct answers Find fundamental frequency (greatest common denominator), and every multiple in the wave
3 types of filter correct answers High-pass, low-pass, and band-pass
Draw a filter given its cutoff frequencies correct answers only the region inside the bounds passes through, all other frequencies are "shaded out" (can't pass through)
How does a filter affect a complex periodic sound correct answers Only the frequencies inside the filter's bounds can pass through; a sound with 200Hz and 600 Hz passing through a
400Hz low-pass filter will sound ONLY LIKE 200Hz (600Hz filtered out)
Assess relative bandwidths of different filters correct answers Bandwidth= 1/2 the area of the filter; area of bandwidth=area of skirts
A smaller bandwidth will have steeper skirts than a larger bandwidth
2 real-world examples of speech sources in source-filter model correct answers Larynx/vocal folds, and turbulence (in fricatives)
Harmonics in a power spectrum correct answers Multiples of the fundamental frequency
Identify quantal regions in a graph correct answers Quantal regions= horizontal line of glottal
width (articulation) vs. acoustic output graph Glottal width is changing, but the acoustic output (sound produced) doesn't change
Consonant produced based on tube model configuration of vocal tract correct answers Fricative: narrow constriction formed by an obstacle; closed tube model in back cavity, front cavity is closed-open tube at place of constriction
Stop: complete closure of lips, followed by rapid opening; back tube closed on both ends, and
front tube is closed at place of articulation, but open at lips
Nasal: VP port at back of pharynx is open (branching tubes) Identifying fricatives or stops in a spectrogram correct answers Fricatives: spread out white noise, sibilants have higher energy in higher frequencies than other fricatives; voicing bar near bottom
Stops: absence of energy (gap), followed by burst of energy
Real world examples of speech filters in source-filter model correct answers Vocal tract, front
cavity
Identify whether an individual cavity in a tube model is open or closed on each end correct answers Closed-open tube model: glottis to pharynx (back cavity), then from pharynx to mouth (front cavity)
Entire vocal tract one long tube for vowels
Closed tube model: when constriction is involved in a tube
Calculate resonant frequency given closed or closed-open tube correct answers Closed tube: nc/2L
Closed-open tube: (2n-1)c/4L
Calculate anti-resonance for a nasal vowel given the nasal cavity length correct answers ONE
closed-open tube (closure at glottis and opening at nostrils); LONG vocal tract because vocal tract is just one long tube (F= (2n-1)c/4L)
Mouth cavity is also closed-open tube; these resonant frequencies cancel out some of the frequencies in acoustic output of sound
Describe how anti-formants affect a speech sound spectrum correct answers Anti-formants from mouth cavity cancel out some frequencies in acoustic output; rebound in oral cavity and
flip polarity, then when they interfere with speech signal they "cancel" out those waves at the same resonant frequency
In a spectrum, these appear as valleys, attenuating the overall sound and reducing energy of the sound
Resolve where individual formants come from in a multi-tube model correct answers L=(2n-
1)c/4F
Recognize a Helmholtz resonator correct answers low fundamental frequency set by back cavity; constriction is in place between 2 tubes (like a bottle)
Assign lyrics in an opera to different voice parts and justify your choice: what is it about the pitch of a sung vowel that determines intelligibility? correct answers Higher frequency sounds are less intelligible: less-frequent harmonics because of a higher fundamental frequency, so less easily distinguishable speech
Difference between spectra and spectrograms correct answers Spectra: instantaneous moment
in time, and which frequencies are present at that moment in time (amplitude/height correlated to energy at that moment); is one "slice" out of a spectrogram
Spectrogram: shows a range of time, and how the frequencies present in the sound change over time (amount of shading/darkness indicates how much energy is in that frequency)
Identify harmonics in a spectrum or spectrogram correct answers Multiples of the fundamental frequency (a peak on a spectrum)
Convert between seconds and milliseconds correct answers Divide seconds by 1000
Wave's frequency and period correct answers Inverse of each other
Periodic waves correct answers repeating same waveform
Simple periodic (sine) wave correct answers pure tone: only one frequency
Match time domain and frequency domain signals correct answers Find period of a wave from time domain, then take the inverse and match to frequency domain
Add 2 sine waves correct answers Share period of the slower wave, so pick an interval to repeatedly find the average amplitude between the two waves, plot that amplitude at every interval (over 4 seconds, measure the average amplitude of the waves every .5 second, draw line connecting dots)
Frequency domain of 2 sine waves correct answers Find fundamental frequency (greatest common denominator), and every multiple in the wave
3 types of filter correct answers High-pass, low-pass, and band-pass
Draw a filter given its cutoff frequencies correct answers only the region inside the bounds passes through, all other frequencies are "shaded out" (can't pass through)
How does a filter affect a complex periodic sound correct answers Only the frequencies inside the filter's bounds can pass through; a sound with 200Hz and 600 Hz passing through a
400Hz low-pass filter will sound ONLY LIKE 200Hz (600Hz filtered out)
Assess relative bandwidths of different filters correct answers Bandwidth= 1/2 the area of the filter; area of bandwidth=area of skirts
A smaller bandwidth will have steeper skirts than a larger bandwidth
2 real-world examples of speech sources in source-filter model correct answers Larynx/vocal folds, and turbulence (in fricatives)
Harmonics in a power spectrum correct answers Multiples of the fundamental frequency
Identify quantal regions in a graph correct answers Quantal regions= horizontal line of glottal
width (articulation) vs. acoustic output graph Glottal width is changing, but the acoustic output (sound produced) doesn't change
Consonant produced based on tube model configuration of vocal tract correct answers Fricative: narrow constriction formed by an obstacle; closed tube model in back cavity, front cavity is closed-open tube at place of constriction
Stop: complete closure of lips, followed by rapid opening; back tube closed on both ends, and
front tube is closed at place of articulation, but open at lips
Nasal: VP port at back of pharynx is open (branching tubes) Identifying fricatives or stops in a spectrogram correct answers Fricatives: spread out white noise, sibilants have higher energy in higher frequencies than other fricatives; voicing bar near bottom
Stops: absence of energy (gap), followed by burst of energy
Real world examples of speech filters in source-filter model correct answers Vocal tract, front
cavity
Identify whether an individual cavity in a tube model is open or closed on each end correct answers Closed-open tube model: glottis to pharynx (back cavity), then from pharynx to mouth (front cavity)
Entire vocal tract one long tube for vowels
Closed tube model: when constriction is involved in a tube
Calculate resonant frequency given closed or closed-open tube correct answers Closed tube: nc/2L
Closed-open tube: (2n-1)c/4L
Calculate anti-resonance for a nasal vowel given the nasal cavity length correct answers ONE
closed-open tube (closure at glottis and opening at nostrils); LONG vocal tract because vocal tract is just one long tube (F= (2n-1)c/4L)
Mouth cavity is also closed-open tube; these resonant frequencies cancel out some of the frequencies in acoustic output of sound
Describe how anti-formants affect a speech sound spectrum correct answers Anti-formants from mouth cavity cancel out some frequencies in acoustic output; rebound in oral cavity and
flip polarity, then when they interfere with speech signal they "cancel" out those waves at the same resonant frequency
In a spectrum, these appear as valleys, attenuating the overall sound and reducing energy of the sound
Resolve where individual formants come from in a multi-tube model correct answers L=(2n-
1)c/4F
Recognize a Helmholtz resonator correct answers low fundamental frequency set by back cavity; constriction is in place between 2 tubes (like a bottle)
Assign lyrics in an opera to different voice parts and justify your choice: what is it about the pitch of a sung vowel that determines intelligibility? correct answers Higher frequency sounds are less intelligible: less-frequent harmonics because of a higher fundamental frequency, so less easily distinguishable speech
Difference between spectra and spectrograms correct answers Spectra: instantaneous moment
in time, and which frequencies are present at that moment in time (amplitude/height correlated to energy at that moment); is one "slice" out of a spectrogram
Spectrogram: shows a range of time, and how the frequencies present in the sound change over time (amount of shading/darkness indicates how much energy is in that frequency)
Identify harmonics in a spectrum or spectrogram correct answers Multiples of the fundamental frequency (a peak on a spectrum)
