sound lab 2————————————————————————————–install Praat on a sound-enabled computer of your own, or one that you have regular
access to. It will be most convenient to install Praat on a laptop that you can bring to
class, but you can certainly do all of these tasks on a desktop computer at home. If you do
not have access to a computer, please let me know so we can make alternate
arrangements.
1. Setup
Start by downloading the necessary files:
•
•
Install Praat:
Sound files: Download the following TWELVE files from the Canvas course website
under Files > Sound Labs. These files are available in a single zipped folder named: SL2
Sound Files.zip. Once you download the zip file, you must open it and transfer the
documents to your desktop or a different designated file before Praat can read them. You
will find the following 12 documents:
o
o
o
o
o
female.wav, female.TextGrid
male.wav, male.TextGrid
source_low.wav, source_high.wav
v1.FormantGrid, v2.FormantGrid, v3.FormantGrid
v1.Spectrum, v2.Spectrum, v3.Spectrum
Launch Praat. This should bring up two windows, one named Praat Objects and the other Praat
Picture. You can close the Praat Picture window; as before, we won’t use it for this lab.
2. ENGLISH VOWELS
In the Praat Objects window, click Open → Read from file…, then select the first four files:
female.TextGrid, female.wav, male.TextGrid, and male.wav. You should now see four items
listed in the window under Objects: “1. TextGrid female”, “2. Sound female”, “3. TextGrid
male” and “4. Sound male”.
2.1. Female Speaker
Select BOTH the “Sound.female” and “TextGrid female” objects. To select two objects together
at the same time, click on the first object then hold down the [Ctrl] key (Windows, Linux) or
[command] key (Mac) while clicking the second object. Now click View & Edit. A sound editor
window should open displaying: (1) a waveform; (2) a spectrogram; (3) word labels underneath
each vowel sound. Make the sound editor window full-screen for easiest viewing and
manipulation.
As we did in SL1, let’s adjust a few settings to optimize the display and make sure we all get the
same results:
p. 1 of 8
•
•
•
•
•
•
Under the Spectrum menu, make sure Show spectrogram is selected.
Under the Spectrum menu, select Spectrogram settings…. Set View range to 0-5000 Hz.
Under the Pitch menu, make sure Show pitch is selected.
Under the Pitch menu, select Pitch settings…. Set Pitch range to 75-500 Hz.
Under the Intensity menu, make sure Show intensity is NOT selected.
Under the Formant menu, make sure Show formants is selected.
•
Under the Formant menu, select Formant settings…. Set Maximum formant (Hz) to 5000
and Number of formants to 4. This is VERY important.
•
Under the Pulses menu, make sure Show pulses is NOT selected.
As a reminder from last time, you can select parts of the sound file via click-and-drag in the
waveform or spectrogram regions of the window. Once you’ve highlighted an area, you should
see a bar below the word labels with a number indicating the time duration in seconds for your
selection. Press that bar to play just your selected portion of the file.
Try selecting and listening to each word, one word at a time. Note that the text labels actually
only cover the vowel portion of each word. To listen to a full word you will have to click-anddrag to select a slightly wider area than just what’s covered by the text label.
QUESTIONS
Remember that your responses should be recorded on the answer sheets on pages 7 and 8 of this
document. You will only hand in those two pages.
Q1.
What is the mean pitch measured separately for each of the eight vowels for the female
speaker? (You’re being asked for eight separate measurements here.) To find the mean
pitch (the average fundamental frequency, or F0) for a given vowel, click on the
corresponding word label, then under the Pitch menu, select Get pitch. Note that since the
vowel duration is specified, everyone should get exactly the same pitch values here.
Record your measurements in Table 1 of the answer sheet.
Q2.
Now measure the first two formant frequencies (F1 and F2) for each vowel. Record these
in Table 1 on the answer sheet as well.
We cannot get formant frequencies by the same method we used for pitch. Move your
cursor to a single time position at which you want to measure the formants. Under the
Select menu click Move cursor to… to open a dialogue box. For the first vowel (i.e., in
“bead”), enter [0.099] for Position and click “Apply” (so that the dialogue box stays open).
Now under Formant click Formant listing to get F1, F2, F3, and F4 values (and also a
time point). Record the values for just F1 and F2 in your table. Repeat this procedure for
each vowel. The time points at which you should get formant measurements are as
follows:
vowel
position (sec)
vowel
position (sec)
p. 2 of 8
/i/ in bead
/ɪ/ in bid
/ɛ/ in bed
/æ/ in bad
0.099
0.310
0.508
0.726
/ɑ/ in bod
/ɔ/ in bought
/ʊ/ in book
/u/ in booze
0.943
1.150
1.338
1.553
2.2. Male Speaker
Close the sound editor window you’ve had open for the female speaker recording. Now select
together both the “Sound.male” and “TextGrid male” objects and click View & Edit. As before,
a sound editor window should open displaying: (1) a waveform; (2) a spectrogram; and (3) word
labels underneath each vowel sound.
With the display settings already established as above for the female speaker, just one adjustment
is now required for the male recording:
•
Under the Formant menu, select Formant settings…. Now set Number of formants to 5.
As before, this is VERY important.
Q3.
Select and listen to each word produced by the male speaker. Using the same procedure as
in Q1 above, what is the mean pitch measured separately for each of the eight vowels?
Continue recording your measurements in Table 1 on the answer sheet.
Q4.
Measure the first two formant frequencies (F1 and F2) for each vowel as you did for
female speech. For this recording, the time points at which you should get formant
measurements are as follows:
vowel
/i/ in bead
/ɪ/ in bid
/ɛ/ in bed
/æ/ in bad
position (sec)
0.125
0.441
0.771
1.141
vowel
/ɑ/ in bod
/ɔ/ in bought
/ʊ/ in book
/u/ in booze
position (sec)
1.465
1.768
2.028
2.366
Q5-6. Calculate an overall mean fundamental frequency (F0) separately for the female and male
speaker, respectively. Find this as a mean of means, i.e., the average of the mean pitches
for the eight vowels (again, figured separately for each speaker).
2.3. Vowel Chart
Q7.
Starting with the blank chart provided on the answer sheet, make a vowel plot showing the
separate F1/F2 positions for each vowel for each speaker. Use different colors for the male
and female data. Draw corresponding colored lines to connect /i/→/ɪ/→/ɛ/→/æ/→/ɑ/→/ɔ/→
/ʊ/→/u/→/i/ and add a legend and IPA symbols next to each dot to complete the figure.
3. SOURCE-FILTER THEORY
p. 3 of 8
The basic idea of Source-Filter Theory is this: We produce speech sounds in two stages. First we
generate the sound source by vibrating the vocal folds in the larynx. This source is then filtered
as it passes through the vocal tract above the larynx (i.e. the mouth and nose). The source
produces pure tones of many different frequencies (harmonics of the fundamental frequency),
and the filtering process amplifies or dampens this set of frequencies in a specific way. By
changing the configuration of our vocal tract—for example by bunching up the back of our
tongue or opening our mouth to a different extent—we apply a different filter to the same source
in order to create a different speech sound.
Let’s try out this idea by combining sources and filters in various ways to synthesize artificial
vowels. Load the following EIGHT files into the Praat Objects window (via Open → Read from
files…):
•
•
•
source_low.wav, source_high.wav
v1.FormantGrid, v2.FormantGrid, v3.FormantGrid
v1.Spectrum, v2.Spectrum, v3.Spectrum
As you might guess from the file names, we have two sources (source_low and source_high) and
three filter functions (v1, v2, v3). The .wav files are sound files, meaning we can play and listen
to these. The .FormantGrid files provide summaries or directions for each filter function, while
the .Spectrum files each describe an intensity distribution across the different frequencies, i.e., a
spectral profile.
Q8.
Select and play (i.e. click Play instead of View & Edit) the two sound objects, source_low
and source_high. In what way do they sound different from one another?
Q9.
We can visualize a Spectrum object by selecting it in the Objects list and clicking View &
Edit on the right. Display each of the three Spectrum objects, v1, v2, and v3. Which one
looks like the following?
Q10. Keep open the Spectrum object that looks like the figure above. What is F1 for this
Spectrum object? Find this by positioning the cursor at the first spectral peak—not all the
way to the left of the display, but rather the first complete peak following the first dip.
After positioning your cursor, the frequency of the peak will be displayed in red at the top,
p. 4 of 8
just above the dotted line that now passes through the peak. Repeat this procedure to also
find F2.
Q11. Compare the F1 and F2 formants that you just found in Q10 to the formant table that you
created in the first part of the lab (Table 1 on the answer sheet). Which vowel might you
guess that this spectral filter will produce? (HINT: Choose only among /i/, /ɑ/, and /u/.)
Next, we’ll apply filters to our “pure” sources. To apply a filter function to a source sound in
Praat: (1) In the Objects list, select a FormantGrid and a sound object together (as before, using
[Ctrl] or [command] + click); (2) click the Filter bar on the right. For example, to apply filter v3
to source_low, you would select FormantGrid v3 and Sound source_low together then click
Filter. You should now see a new Sound object in the list (e.g., Sound source_low_filt). The new
Sound object is the output of the filter function applied to the source sound. Select the new object
and click Play to listen to it.
Q12-14. Apply each of the three filter functions (v1, v2, v3) separately to source_low and listen
to each output. Each output should sound like a different vowel in English. Which
filter function produces output that sounds like:
(12) /i/: the vowel in ‘key’?
(13) /u/: the vowel in ‘cool’?
(14) /ɑ/: the vowel in ‘calm’?
Q15.
Was your guess from Q11 accurate?
Q16.
Finally, apply each of the three filter functions (v1, v2, v3) to source_high. In what
ways do these outputs sound similar to and/or different from the corresponding sounds
produced with the same filters applied to source_low in the previous question?
[END OF QUESTIONS] — Proceed to the answer sheet on following page.
p. 5 of 8
Answer Sheet
Name: _________________________________
Part 1: English Vowels
Table 1: F0, F1, and F2 for English vowels produced by a female and a male speaker.
/i/
bead
/ɪ/
bid
/ɛ/
bed
/æ/
bad
/ɑ/
bod
/ɔ/
bought
/ʊ/
book
/u/
booze
Female
F0 (Hz)
F1 (Hz)
F2 (Hz)
Male
F0 (Hz)
F1 (Hz)
F2 (Hz)
Female speaker’s mean F0 (based on the eight measurements above) = ___________.
Male speaker’s mean F0 (based on the eight measurements above) = ___________.
Part 2: Vowel chart
Plot F1/F2 positions for each vowel for each speaker. Use different colors for male and female.
Draw corresponding colored lines to connect /i/→/ɪ/→/ɛ/→/æ/→/ɑ/→/ɔ/→/ʊ/→/u/→/i/, add a legend
explaining which color represents which gender, and write the appropriate IPA symbol next to
each dot to complete the figure.
F2 (Hz)
3000
2500
2000
1500
1000
500
200
300
400
500
600
F1 (Hz)
700
800
900
1000
p. 6 of 8
Part 3: Source-Filter Theory
Q8.
In what way do source_low and source_high sound different from each another?
_____________________________________________________________________
Q9.
Which Spectrum object looks like the given figure? ________
Q10. What are F1 and F2 for the Spectrum object identified in Q9? F1: _______
F2: ______
Q11. Which vowel might you guess the filter identified in Q9 will produce? ________
Q12. Which filter produces the vowel in ‘key’? ________
Q13. Which filter produces the vowel in ‘cool’? ________
Q14. Which filter produces the vowel in ‘calm’? ________
Q15. Was your guess in Q11 accurate? ________
Q16. In what ways do the outputs created by applying the set of filters to source_high vs.
source_low sound similar to and/or different from each other?
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
[END]
p. 7 of 8
Purchase answer to see full
attachment
Why Choose Us
- 100% non-plagiarized Papers
- 24/7 /365 Service Available
- Affordable Prices
- Any Paper, Urgency, and Subject
- Will complete your papers in 6 hours
- On-time Delivery
- Money-back and Privacy guarantees
- Unlimited Amendments upon request
- Satisfaction guarantee
How it Works
- Click on the “Place Order” tab at the top menu or “Order Now” icon at the bottom and a new page will appear with an order form to be filled.
- Fill in your paper’s requirements in the "PAPER DETAILS" section.
- Fill in your paper’s academic level, deadline, and the required number of pages from the drop-down menus.
- Click “CREATE ACCOUNT & SIGN IN” to enter your registration details and get an account with us for record-keeping and then, click on “PROCEED TO CHECKOUT” at the bottom of the page.
- From there, the payment sections will show, follow the guided payment process and your order will be available for our writing team to work on it.