week 6 Flashcards
Good data is produced how
quiet environment, microphone, no clipping (bringing signal in at too high of a level) and a proper sampling rate
sound is dependent on what type of molecule
air, sound is the product of air molecule movement
acoustic analysis
non invasive, looking at the waves our sound produces, analyze what manners of articulation produced the sound, from a microphone
a disordered voice is caused by a change in ________
source, the larnyx, and vocal folds
a disordered articulation is caused by a change in _______
filters, our tounge, pallates (hard and soft), teeth etc
what it motor equivalence of speech
some vocal sounds, can be produced by different articulatory movements, and acoustic analysis cannot see that
what is a limitation of acoustic analysis?
motor equivalency of speech, it cannot capture everything. we cant see how their articulators moved from a voice recording.
what is a filter in regards to recording waves
selects specific frequencies within a complex wave for further subsequent analysis, looks at all the frequencies like a prisim, ingredients
high pass filter does what to an acoustic recording
allows high frequencies through and blocks out lower frequencies or holds them back
low pass filter does what to an acoustic recording
allows low frequencies through and blocks out higher frequencies or holds them back
band pass filter does what to an acoustic recording
allows a specific set range of frequencies through, and can hold back higher or lower frequencies
band reject filter does what to an acoustic recording
holds back or blocks out a specific set range of frequencies, and can allow both higher or lower frequencies to go through
what is a fourier theorem
all periodic sounds are made up of a combination of sine waves
how do we changing domains? apply a ______ ________
fourier transform , creates a spectrum of all the different parts of a wave from a microphone signal a time display, or time domain waveform (y is amplitude, x is time). analyzes to create a spectra of all the individual components. takes a time domain display and transforms it into a frequency domain display (y amplitude , x is frequency)
time domain display
a microphone recording
frequency domain display
a spectrum, a slice in time or, snapshot in time of the frequency of a wave
harmonics are exact multiples of the _______
fundamental
sound source comes from our____
larnyx (voice box) , vocal folds,
the human voice is nearly ________
periodic, almost repeats in cycles perfectly. Contains normally small amounts of perturbation, without perturbation (shimmer and jitter) our voices would sound artificial. has harmonics and a fundamental
spectra represents what
frequency components that are present within a sound
sine wave will have how many lines in a spectra display?
one
complex periodic signals have multiple lines in a spectra display, true or false?
true
spectral envelope does noise look like in a spectral display
all different frequencies, with different phase relationships to one another, with approximately equal amplitude. vertical lines very close set together.
spectra envelope
space not filled in under neath spectral peaks, takes away vertical lines that represent individual sine waves.
what is a (FFT) Fast Fourier transform ?
shows the range of harmonics available in a complex sound, there is a peak for each one of those harmonics. Not good at showing formants/filters, but better at showing the source of a voice/sound
what are formants?
the vocal tract filter, our teeth, tounge, hard or soft pallate
what is a (LPC) linear predictive coding spectrum?
spectral envelope, doesnt show the harmonics doesnt show the filled in bottom, good at revealing what our vocal tract/filters/formants, not good at the source. its good at revealing what our filters do with the source/our larynx gives to us.
where are the larynx and vocal tract located in relativity to their productive positions
The larynx, or the voice box is located below the vocal tract, the vocal tract extends from the point where it meets the larynx, to the jaw and mouth and lips
peaks in a display, represent
formants, vocal tract, and the energy they display
vocal tract transfer function???????
The vocal tract transfer function predicts the resonance patterns of the vocal tract for a specific articulation. represents how we moved vocal tract to produce sound
what is a spectrogram
our speech lined up over time, single slices together. shows 3 parameters of language, the darker spots being intensity, the y axis being frequency and the x axis being time. a cross breed between time domain and frequency display. shows how the strength (freqency) and intensity changes over time
turn spectral slices 90 degree to the left to create what
a frequency domain
what is a nyquist frequency?
half the sampling rate
Nyquist theory states
to produce an analog signal without aliasing error, the sample rate must be twice the signals highest frequency component.
Aliasing is
the distortion that occurs when high-frequency information is incorrectly identified as low frequency usually due to incorrect sampling rates in analyzing of sound waves. to prevent… anti-aliasing filter is used before the sampling process to remove frequencies above the Nyquist frequency
wide band spectrogram is what in regards to its parameters
temporal or time detail is clear, but frequency is poor. shows glottal pluses as vertical striations. y axis is frequency, and x axis is time. each grey line represents one kilo hrtz. Darker spots represent more immense amounts of energy displaced over time within the glottal pulses produced.
narrow band spectrogram is what in regards to its paramters
clear frequency data but very poor time resolution. horizontal bands are harmonics. no more glottal pulses, but harmonics.
why are there trade offs in regards to band width displays?
analysis is a string of numbers, and the more information we need analyzed the more information we need to input.
what do darker areas mean on a spectrogram?
larger places of energy
