10/7/2018
1: Introduction to Sound Signals
Introduction
Most of the labs will use various tools to visualize sounds. For example, we will record sound with computers that store the recording as a list of numbers (similar to the way a CD represents sound).
Today, you will talk into a microphone, which converts the sound pressure into an electrical signal, and the computer will measure the strength of that electrical signal about 22,000 times each second. The result is that each second of recorded sound is represented by a time-series of 22,000 numbers that we will call a "sound signal". Once the sound data is stored on the computer, there are many things that the computer allows us to do to analyze or modify the sound recording. One thing the computer can do is to make a graph of this data. Here is how the computer graphs a part of the recording of the word "hello"
The vertical axis is the strength of the electrical signal that is proportional to sound pressure, and the horizontal axis is time. The computer connects the points on the plot to help us to see the patterns.
Recording sound
A.Let's try it out.
1.Turn on the computer if it is not already on. After a short time, the computer monitor will light up and you will see the "desktop." Using the mouse, point the cursor arrow at the symbol in the lower left that is labeled "Praat." Then, quickly click the mouse button twice [i.e., "Double-click"]. This automatically starts up the program called Praat. Praatis a computer application that phoneticians and other researchers use to analyze, synthesize, and manipulate speech and other sounds as well a providing high quality figures to document their work. Wait a minute for the program to load itself. Two windows should appear: a Praat objects window and a Praat picture window.
B. To record from the microphone, perform the following steps:
- Choose Record mono Sound… from the New menu in the Praat objects window. A SoundRecorder window will appear on your screen.
- Press the button on the computer’s microphone to activate the microphone. Use the Record and Stop buttons to record a few seconds of your speech. Experiment with how loud to speak and how close to the microphone. The level meter will indicate the appropriate combination, i.e., you want to maintain the level to show green with the rare “spike” into the yellow or red levels.
- Use the Play button to hear what you have recorded.
- Repeat the two steps above until you are satisfied with what you have recorded. To make sure that the sound file does not get too large, don’t record more than a few seconds of your voice for now.
- Type in a name for your sound file in the text box below the Save to list: button. Hit the Save to list: button and the text string “Sound name” should appear in the Praat objects window that indicates file where your sound is recorded.
- The right part of the Praat objects window shows what you can do with the sound. Try the Play, Edit, and Draw buttons.
C. Understanding the Displays
We will be examining waveforms that consist of sound pressure on the vertical axis and time on the horizontal axis.
1. Edit window
Again, you get the Edit window by clicking the Edit button in the Praat objects window when your sound file is selected. It will show a waveform of your sound displayed in the top half of a window that has your sound file as its name. The bottom half of the display will be a spectrogram of your sound that we will learn about later in the course. An example is shown below:
Top: waveform of "test, test, test" spoken in to the microphone, with sound pressure along the vertical axis, and time along the horizontal axis. Bottom: corresponding spectrogram that will be discussed later in the course.
With this window up, hit the long horizontal Window button below the displays. A vertical line will sweep across the display, indicating what portion of the sound that is being played at a given time, and your sound will be played back.
Click and drag (i.e., point the cursor to some part of the sound signal, click and hold the mouse button down, and drag to another part of the sound signal) in the display to change the size of a red rectangle in the display. This selects a portion of the waveform that can subsequently be played back by clicking on the button that is lined up directly below the red selection rectangle.
Play back different parts of your recording. Try selecting less than an entire word and playing it back. Can you still understand what is being said?
Go through the whole chain several times, i.e., saying (or singing) different things into the microphone, recording the results, storing it in to a sound file, and then using the Edit button to examine the different sound waveforms. We will develop the tools to analyze the differences systematically during the semester.
2. Praat picture window
You can also draw the waveform of your sound by clicking and dragging a red window in the Praat picture window, and then hitting the Draw button in the Praat objects window when your sound file is selected. The picture will always fill the drawn red rectangle. Drawing always overlays a new picture on top of an existing picture, so you will have to choose Erase all from the Edit menu of the Praat picture window to get rid of the previous drawing.
When you hit the Draw button, a choice of range of times (from time tstart to tend) to display is given. If the second, tend, time entered is 0.0, the entire waveform will be displayed by default. Try hitting the Draw button a few times with different time ranges (remembering to Erase all from the Edit menu each time) to “zoom in” on interesting portions of the waveform.
3. Printing your recording
Once you hae a good recording in the Praat picture window, try printing the display. To do so, choose Print from the File menu of the Praat picture window. On your printout, lable the loudest and softest parts of your recording, and if possible, the words or letters corresponding to each section of the waveform.
Viewing a sound signal on the oscilloscope
One other tool we will often use this semester is the oscilloscope. It is useful for looking at how the sound signal changes at the same time as the sound signal is being made (which is not possible with Praat). We will also use the oscilloscope to measure simple sound signals (but it is not so useful when we need to analyze complicated sounds, and for this purpose,Praatexcels).
Using the Oscilloscope
1.Make any sustained sound into thesmall black microphone that is connected to the oscilloscope. Make it louder, then softer: how does the appearance of the sound signal change?
- Then try a musical sound: make it higher in pitch, then lower: how does the appearance of the sound signal change?
What do we mean by "sound?"
When people talk about “sound,” they may mean one of several things. Usually, they mean either the sensation people have when vibrations of the air hit their ears, or they mean the actual vibration of the air, whether or not anyone is there to hear it.
The relationship between vibrations and sensual perceptions is complicated, largely because perception itself is so complicated. How loud a particular sound seems to you may change depending on your mood. Nevertheless, the three most basic qualities that are often used to describe perceived sound—loudness, pitch, and timbre—do correspond in a fairly straightforward way to physical properties of vibration that can be easily measured.
The purposes of the next part are: (1) to give you some experience with the relationship between the qualities of perceived sound and the properties of physical sound, and (2) to introduce you to another of the tools we will use for the rest of the semester: the Function Generator, a device that can be used to make several different kinds of electrical oscillations that can be perceived as sound. Most usefully, the function generator allows you to independently control loudness, pitch, and timbre. "Independently" means that you can change one of these three while leaving the other two the same, which helps you to identify how the sound changes when you change that property.
Loudness
Vibrations can be large or small. Large vibrations are said to have a large amplitude. In the world of sound, amplitude corresponds to loudness—the greater the amplitude, the louder the sound.
A.Setting the amplitude
1.Turn on your function generator (the little grey box) by pushing the red POWERON button. Choose a sine wave button, rather than a square or triangle wave. Find the knob called AMPLITUDE and make sure it is turned all the way down (counter-clockwise). Put on the headphones and then slowly turn up the amplitude. Notice that the sound gets louder.
That’s all there is to it. [not really—our perception of loudness is also influenced by pitch and timbre]
Pitch
Vibrations can happen quickly or slowly. This property of vibration is called frequency. Frequency tells how often something happens in a given unit of time. For example, the sound you just heard caused your eardrum to move back and forth about 1000 times in each second, so the frequency was about 1000 per second. The unit “per second” has an abbreviation, Hz, named after Heinrich Hertz and pronounced “hurts.” So a vibration that happens 1000 times each second is said to have a frequency of 1000 Hz.
The physical property frequency corresponds to the sensation of pitch. Try this out for yourself: put on the headphones, turn up the amplitude to a comfortable volume, and then turn the knob on the function generator that is labeled FREQUENCY. As you move the knob clockwise (toward the number .2) both the frequency and the pitch go down. Twist the knob the other way and the pitch goes up.
B.Setting the frequency
The function generator can produce any frequency from less than 1 Hz up to more than 20,000,000 Hz. This wide range is obtainable by use of the bank of grey buttons labeled FREQUENCY MULTIPLIER–Hz. The 1K button is currently pushed in. The k is an abbreviation for kilo meaning 1000 (as in kilometer or kilogram). So the generator is now set to generate frequencies near 1000 Hz. To find the frequency more accurately, you multiply this number by the number on the frequency dial, which goes from 0.2 to 2.0. So when the 1k button is pushed in and the dial is set to 2.0, the frequency is 2.0x1000 Hz = 2000 Hz.
1.What is the frequency when the 1k button is in and the dial is set to 0.2?
The dial is not very precise. Later in the semester we will learn how to measure the frequency more exactly.
To get higher or lower pitches you push a different FREQUENCY MULTIPLIER button. Push in the 10k button. This puts you in the frequency range from 2000 to 20,000 Hz, and is at the top of the human hearing range. Be sure you understand the relationship between the dial and button settings on the generator and the frequency.
2.With the multiplier set to 10k, what range of frequencies can we obtain?
a) Use the generator and counter to determine the highest frequency you can hear:
b) Without changing the amplitude knob, turn the dial from its lowest setting (0.2) to its highest (2.0). Does the loudness you perceive change?
What is the dial setting when the sound seems loudest?
3.To get low frequencies, set the multiplier to 100. What range of frequencies can you get with this setting?
4.What is the lowest frequency you can hear that sounds like a pitch?
(Nore: If you only hear clicking, you are not hearing a frequency—because the headphones are small, they do not produce low sounds very efficiently, so you will need to turn the amplitude all the way up on both the generator and on the volume control on the headphone box. You will probably want to turn it down again later, or it the sounds will be annoyingly loud.)
Timbre (pronounced “tamber” )
You have seen (and heard) that sounds can be loud or soft, depending on amplitude, and they can be high or low, depending on frequency. But that is not all there is to sound. You can distinguish sounds that have the same pitch and the same volume. For example, you can sing the vowel sound “a” and then the vowel “e”, and they sound different, even though they have the same frequency and amplitude. The sounds a and e have different tone qualities or timbres. Timbre much more complicated than loudness or pitch, and you will spend much of the semester studying it.
C.Setting the function
To begin your investigation of the timbre, listen to the effect of changing the “shape” of the vibration. The function generator can produce three different types of oscillation, called sine, triangle, and square. So far, you have been listening to the smooth sine function. You can choose one type of oscillation or another by pushing one of three MODE buttons, labeled for sine, for triangle, and for square.
1.First, set the frequency to around 250 Hz. Then push the triangle button and listen to the sound. Typically, the triangle function is said to be “brighter” or “sharper” than the sine function. You should avoid calling it “higher,” since that implies a higher pitch. Do you agree that the sound is “brighter”?
2.Now try the square function. It is much brighter, possibly even “buzzy.” It will probably also sound louder to you, so you may want to turn down the volume to make a good comparison.
3.Switch back and fourth, listening to the three timbres. Try listening to the different shapes at widely differing frequencies.
4.Is it easier to distinguish the sounds at low or at high frequency?
5.Is there any frequency at which you cannot tell the three apart?
The smooth sine function is the simplest, purest, cleanest, least complex sound there is. Every other sound is, at least to some degree, brighter, sharper, or more harsh than the sine sound.
Looking at the sounds
To see why these sounds are described by shapes, such as “square” and “triangle,” turn on the oscilloscope. The oscilloscope makes a picture of the electrical signal that the function generator produces.
D.Electronic signals on the oscilloscope
1.Set the generator to the sine function and the frequency to 1000 Hz and look at the oscilloscope. Sketch the appearance of the sound signal in the space below:
2.Increase the frequency using the FREQUENCY knob and look at the signal again.Sketch it in the space below:
3.How does the appearance of high pitches on the screen differ from low pitches?
4.How could you use the screen image to help you measure the pitch?
5.Change the frequency multiplier and observe the sound signal again. How does the picture on the screen change when you increase or decrease the multiplier?
6.Now leave the frequency at a comfortable pitch and vary the amplitude. What is the difference between the way loud and soft sounds look on the screen?
7.The volume control on the headphone box has no effect on the signal, which goes to the oscilloscope.) What happens if you push in the button on the function generator that is labeled ATT 30 dB? (Make sure you leave it in the out position when you are done.) What happens on the screen? How does the sound you hear change?
8.Now leave the frequency and amplitude fixed but change the function to the "square" wave or the "triangle" wave. Sketch the sine, square and triangle waves to show how they are different.
9.What does it mean for both the amplitude and frequency of the sound to be the same but the timbre to be different (what is different on the scope)?
Other Sounds
Turn off the function generator.
E.Looking at natural sounds
1.Strike the tuning fork with the rubber mallet and listen to it. Would you say the sound is bright, like a square function, or pure, like a sine function?
You can check by looking at the sound on the scope, but to do that you will need to use a microphone. If you need help, ask your instructor to plug in the microphone for you. Strike the tuning fork and hold the microphone very close to one of the vibrating tines.
2.What shape is the signal? Draw it. Were you right?
3.Now try whistling. What do you conclude?
4.Now try singing into the mike, holding the mike very close to your mouth. Sing all different types of vowel sounds. Which vowel (a, e, i, o, oo, ah, uh) makes the most sine-like pattern?
5.How does a very bright elook (sketch it in the space below)?
Note that even though they are more complicated, musical sounds and vowel sounds (and any other "steady" sound) are periodic just like the sounds produced by the function generator. That means that approximately the same sound signal is repeated over and over. We can use loudness, pitch, and timber to describe all periodic sounds, and we use amplitude, frequency, and shape to measure the physical properties of those sounds.
F. Looking at more complicated sounds
1.Some sounds are different: look at various noisy sounds like ssssss and shhhhhhh.. How do these sounds differ from the simpler sounds you looked at earlier?
Because of its complexity we will do a step by step introduction to the oscilloscope in a later lab.
1: Introduction to Sound - 1