Hands-on Hearing. Sound recording and visualization tools for the science classroom. Jesse Ellis Neurobiology & Behavior

Transcription

1 Hands-on Hearing Sound recording and visualization tools for the science classroom Jesse Ellis Neurobiology & Behavior David Rothstein Astrophysics Mya Thompson Neurobiology & Behavior Cornell Science Inquiry Partnerships Professional Development Day, Cornell University Friday, March 24, 2006, 11:15 am 1:30 pm 1. Introduction 2. How Many Times Does a Ping Pong Ball Bounce? an introduction to waveforms 3. How Do Animals Talk to Each Other? an introduction to spectrograms 4. What Do a Nuthatch and Opera Singer Have in Common? an introduction to harmonics 5. Online Sound Resources including Cornell University s Macaulay Library of animal sounds and The University of Iowa s Electronic Music Studios collection of musical instrument samples 6. Group Brainstorming and Wrap-up

2 Resources for Hands-on Hearing Sound Visualization Software Raven Lite (what we used in the workshop; for Windows and Macintosh) Audacity (a free, open-source program for Windows, Macintosh, Linux, etc.) Syrinx (another free program, for Windows only) Online Sound Resources The Macaulay Library of animal sounds at Cornell University Musical Instrument Samples collection at The University of Iowa s Electronic Music Studios There are many other online collections of musical instrument sounds, but the above is the only one we could find that has its music in a format that can be read by Raven Lite or Audacity. To get the data into these programs, choose a musical instrument from the above website, then right-click on one of the sound files and choose save link as or save target as. Once you ve saved the file, you can open it as a new sound from within the File menu of Raven Lite or Audacity. Other Resources The Cornell Science Inquiry Partnerships (CSIP) program: CSIP Curriculum Resources:

3 Tips for using Audacity (free, open-source alternative to Raven Lite) 1. Recording and Playback are done with the large buttons on the top of the screen. 2. Horizontal Zooming (along the time axis) is most easily done by choosing the zoom tool (magnifying glass icon on the top left of the screen) and then clicking and dragging across the region you want to zoom in on. You can click the right mouse button to zoom out. There are additional zoom buttons on the top right of the screen that work in a similar way as those in Raven Lite. 3. Vertical Zooming (along the pressure axis) is done in a similar way as horizontal zooming; just make sure the cursor is placed somewhere on the vertical axis labels when you click/drag it. 4. Selecting Data is done by choosing the selector tool on the top left of the screen (the one that looks like a cursor and is directly above the magnifying glass icon). Click and drag to select a region of the data. 5. Spectrograms can be made in Audacity by going to the top left of the waveform window, right next to the name of the sound file, clicking on the downward-pointing arrow, and scrolling down to choose Spectrum. Also, a plot of the spectrum (power vs. frequency) for a given period of time can be obtained by selecting a time range using the selector tool, going to the Analyze menu and choosing Plot Spectrum. Getting a spectrogram to display at the same time as the waveform is a little less convenient as compared to Raven Lite, but it can be done as follows: a. Load your waveform data. b. Choose the selector tool on the top left of the screen (the one that is directly above the magnifying glass icon). c. Hit control-a (select all), then control-c (copy), or choose those options from the edit menu on the top of the screen. d. Click the cursor somewhere outside of the waveform window (on the gray area of the screen) and hit control-v (paste). A copy of the waveform should appear. e. On the top left of this new window, click on the downward-pointing arrow and scroll down to choose Spectrum. You should now have the spectrum and waveform on the screen at the same time, and if you zoom in on one, it will automatically zoom in on the other. 6. Slow or Fast Playback can be done by selecting the portion of your data you are interested in (or control-a to select all), going to the Effect menu and choosing Change Speed. Choose a percentage, and then click OK (the preview button doesn t seem to work correctly). The way the percentages work is a little funny. If you want to play back your sound at a relative speed x, use the following equation: percent change = 100 * (x 1) For example, if you want to play it back at twice the normal speed (x=2), enter 100 for the percent change; if you want to play it back at one quarter the normal speed (x=1/4), enter -75 for the percent change.

4 The Physics of Bouncing Ping Pong Balls Some Ideas for Teaching 1. Have students calculate as many physical quantities as they can for each bounce of the ping pong ball (e.g., the height of each bounce, the vertical velocity of the ball before and after it hits the ground, etc.), using the equations of kinematics and the measured times at which the ball hits the ground. The above will be easier if the students enter their data into a spreadsheet program and have the spreadsheet do each calculation on all the bounces at once. This is a good way to show students the power of spreadsheets and, equally, the importance of the human who has to program the spreadsheet with the correct formula! 2. Have students explain and estimate their experimental uncertainties. How accurate were they in determining the time of each bounce, what was the limiting factor in the accuracy, and how do the uncertainties in the time of each bounce affect the accuracy with which they can determine the height and velocity of the ball? 3. Have students plot their data and look for any patterns in the way the physical quantities change from bounce to bounce. Can this information be used to predict the height or velocity of a particular bounce? Can the students find some way to estimate how many bounces the ball might make before something (irregularities in the table, etc.) causes it to stop? The attached sheets show some example graphs and calculations. With the exception of the first few bounces, the ball will lose a constant percentage of its height (and a constant percentage of its velocity) each time it hits the ground. 4. Have students pick one thing about the experiment to change and see how this affects their results (some possibilities include using a different kind of ball, bouncing the ball on a different surface, or bouncing it from a different height). An interesting result that the class might come up with if they pool all their data together is that the bounces will follow a similar pattern no matter how the experiment is changed that is, the ball will lose a constant percentage of its height (and velocity) each time it hits the table. What percentages those are may depend on the ball or surface (e.g., the ball might lose 85% of its height per bounce for a hard surface vs. 50% for a softer one), but the overall physical process is the same. 5. Have students come up with a question of their own and do whatever analysis is necessary to answer it. For example, is there a relationship between the energy the ball loses during each bounce and how loud the bounce is? What could be causing the first few bounces to differ from the pattern that the others follow? Is air resistance a possible factor? There are tons of possible questions to ask, not all of which have answers that are known!

5 Height of Each Bounce Bounce Height (m) Bounce Number

6 Skipping First Few Bounces Height = ( m) x (Bounce Number) Bounce Height (m) Bounce Number

7 How Many Bounces? If the exponential fit is true, then after around 150 bounces the height of each bounce would be about the size of an atom (~10-10 meters)!

8 Listening in to animal communication Activity by: Mya Thompson, Neurobiology and Behavior Department, March 2006 Summary: Classroom project in which students select vocal animals and analyze their sounds. Measurements made on the variety of sounds chosen can then be plotted in order to compare sounds across species. Major concepts: Animal communication, physics of sound, natural selection Skills: Graphical analysis, software use, plotting data Step by step description: 1. Introduction to animal communication. 2. Ask each student (or group of students) to choose a vocal animal to study. In addition to making or finding a recording of the animal chosen, each student researches the natural history of the species including what is known about when and why the animal makes sound. Logistics: If you are using the software Raven Lite,, a variety of great sounds come with the program. An excellent additional resource of a wide variety of animal sounds is the Diversity of Animal Sounds CD distributed by the Cornell Lab of Ornithology. It can be purchased from The sound recordings can then be transferred to your computer in for analysis. 3. Create visual representations of the vocalizations from each species. A variety of samples are attached in Appendix 1. Logistics: Raven Lite software developed in the Cornell Bioacoustics Research Program can be used on any computer to analyze natural sounds. It can be downloaded from To get multiple licenses for your school and discuss pricing, contact Tim Krein at tpk8@cornell.edu. Each student (or group of students) use this program to create spectrograms visual representations of sound useful in analysis. 4. Make measurements on sounds including duration of vocalizations and minimum and maximum frequency. A sample data sheet is attached Appendix 2. Logistics: Measurements can be made directly from printed spectrograms or using the software. I think both methods are valuable. Calculations done from a printed copy would encourage students to think more about what they are measuring. These hand-done measurements could then be compared with the ones generated by the computer. 5. Class constructs a summary of animal signals from the measurements made by plotting the duration and frequency range of each animal s vocalizations and compare across species. A sample graph is attached in Appendix Use what we know about the natural history of each species to describe why a species would make calls of a certain frequency or duration. For further information contact Mya Thompson at mt228@cornell.edu 1

9 Appendix 1 Sample spectrograms of animal sounds For further information contact Mya Thompson at mt228@cornell.edu 2

10 Appendix 2 A sample data sheet For further information contact Mya Thompson at mt228@cornell.edu 3

11 Appendix 3 Sample graph For further information contact Mya Thompson at mt228@cornell.edu 4

12 What do Sounds Look Like? Understanding Natural Sounds Overview: Teacher s Guide By Jesse Ellis, CSIP Graduate Student Fellow, Cornell University This lab is intended to give students a better understanding of the complexities of sounds in the real world, and to give students the opportunity to use software designed for visualizing and measuring natural sounds. Most physics labs on sound treat sound as a well-behaved phenomenon. Those of us studying animal sounds have found that this intersection of biology and physics have produced a vast diversity of sounds, most of them pretty messy from a physics point-of-view. I hope to give students a little sample of that diversity, and a basic understanding that will make the mess a little easier to interpret. It should give students a good workout of their basic wave equations, and help them ground their understanding of sound waves. Students will to listen to a number of animal sounds, mostly birds, and try to predict what time-frequency representation of those sounds would look like (much like writing out music in musical notation). They will then examine and measure those sounds more closely, using the time-frequency representation and wave-form representation of the sounds. Students will also be challenged to draw sounds using their own soundproduction apparatus (speaking, singing, whistling, or other sound production methods), though this will require a microphone plugged into the computer. The sounds should also be available on the website to download as a.zip file, and two sound analysis program (among many) are recommended: Syrinx, for PC, free at: Raven Lite, licenses $25/license for up to 8, $10/license after that, at: Subject: Physical Sciences, Living Environment Audience: Because of the concepts dealt with here, this is probably best done with a physics class that has had an introduction to wave mechanics. However, you could easily lighten up on some of the physics concepts and use this lab to introduce them (or not) to a younger audience, perhaps middle schoolers. Objectives/Goals: The overarching goals of this lab are to give students a greater understanding of sound waves and wave mechanics by introducing them to the complexities of sounds found and

13 generated in the biological world, and to give them a visual method for interpreting and understanding sounds. Specific Objectives: Students will be able to (very) generally describe a time-frequency representation of a sound after measuring and examining the wave-form representation (i.e. rising, falling, having harmonics of any type or not, etc). Students will be able to draw a basic time-frequency representation of a sound after measuring the wave-form in multiple locations. This is basically an exercise in graphing. Students will be able to describe audible features of a sound (sharp, loud, nasal, clear, scratchy, rough, etc) in terms of time-frequency features of the sounds noisy/high bandwidth, having strong harmonic overtones, clicks, dark). Assessment Strategy Play the class an unknown sound and ask them to draw it in the frequency domain on paper, and to describe any features they hear using terms identifiable in the frequency domain. Would there be harmonics? Would the wave form be sinusoidal? Or would it be more irregular? Then have each student measure a wave-form and graph it in the frequency domain. Grading could be based upon general features identified in the unknown sounds (determining the exact frequency representation of a sound is more of an art than a science, and students should not be expected to reproduce the sound just from hearing it!), such as noisiness, general frequency changes over time, or white-noise, click type features that students should be able to identify. More specific grading rubrics could be applied to graphing sounds, since this is a much more quantitative exercise (and there ARE correct answers in this case). Teaching tips and thoughts: Teachers should attempt to familiarize themselves with the sound analysis program they have chosen ahead of time. Each program has its own quirks, and must be dealt with accordingly. It is easier to create and examine spectrographs in Raven, but Raven isn t free! The biggest challenges are creating a spectrograph and then zooming in and out in order to analyze the waveform. In Raven: Creating a spectrograph is a matter of clicking the button in the upper left with three grey horizontal bars in it. There are numerous adjustments that can be made in Raven, though Raven s default settings are usually quite good for looking at sounds. Navigating through the spectrograph is done with a series of buttons near the slider bars. Plus and Minus symbols zoom in the horizontal and vertical axes. The H symbol will zoom out to its full size on that axis, and the red square will zoom in to your selection. In Syrinx: In Syrinx you can only toggle between waveform and spectrograph. This is done most easily by right-clicking the sound and choosing the view at the top of the

14 menu. To change spectrograph details or axes, right-click on the spectrograph and choose Window Settings. To change time resolution and zoom in and out on a sound, the window must be set at Fixed Time scale. To appropriately analyze a waveform, a good time resolution is.01, while to see a whole sound file a good resolution may be 5s or 10s. Figure 1: A waveform in Syrinx of a Red-breasted Nuthatch call. At the bottom of the window are time measures of the left and right sides of the selection box. These can be used to calculate wavelengths and thus frequency. Here the measure is of the wavelength of the modulating waveform, a sinusoid AM wave. This creates the complex harmonics around the carrier frequency, around 3100 Hz. The harmonics are in bands at multiples of ~400Hz above and below the carrier.

15 Figure 2: A spectrograph of a Thicket Tinamou call. This call is a monotone, represented by the thin line at about 1400 Hz. It sounds like a flat whistle. However, it is a much noisier recording than the nuthatch, with a lot of background sound relative to the signal strength. This makes the background darker, since a spectrograph uses a color value to indicate amplitude or energy at a particular frequency. Zooming in on this call (using the plus symbol on the horizontal axis) would show a nearly sinusoidal wave. The small differences from a sine wave are due to the background noise. NOTE: To make measurements on a call in Raven, click and drag the small dot in the middle of the bottom of the screen (below the slider) and drag it up to reveal the measurements panel. This will show start times and end times for a selection box as well as frequency values for the selection box. Time required: 2 class periods? NSE Standards addressed: Content Standards grades 9-12: Physical Science: Interactions of energy and matter: Waves, including sound and seismic waves, waves on water, and light waves, have energy and can transfer energy when they interact with matter.

16 Conservation of Energy and the increase in Disorder: The total energy of the universe is constant. Energy can be transferred by collisions in chemical and nuclear reactions, by light waves and other radiations, and in many other ways. However, it can never be destroyed. As these transfers occur, the matter involved becomes steadily less ordered Life Science Matter, energy and organization in living systems: As matter and energy flows through different levels of organization of living systems--cells, organs, organisms, communities--and between living systems and the physical environment, chemical elements are recombined in different ways. Each recombination results in storage and dissipation of energy into the environment as heat. Matter and energy are conserved in each change. Behavior of Organisms: Like other aspects of an organism's biology, behaviors have evolved through natural selection. Behaviors often have an adaptive logic when viewed in terms of evolutionary principles. Science and Technology Understandings about science and technology: # Scientists in different disciplines ask different questions, use different methods of investigation, and accept different types of evidence to support their explanations. Many scientific investigations require the contributions of individuals from different disciplines, including engineering. New disciplines of science, such as geophysics and biochemistry often emerge at the interface of two older disciplines. Science often advances with the introduction of new technologies. Solving technological problems often results in new scientific knowledge. New technologies often extend the current levels of scientific understanding and introduce new areas of research. Creativity, imagination, and a good knowledge base are all required in the work of science and engineering. Science as Inquiry Understandings about Scientific Inquiry: Mathematics is essential in scientific inquiry. Mathematical tools and models guide and improve the posing of questions, gathering data, constructing explanations and communicating results.

17 For high school physics students: Background: What is a sound? A sound is a longitudinal wave that is propagated through a physical medium (unlike light or electromagnetic waves). Its physical manifestation is as a patterned increase and decrease in the local density of the medium that is, pressure. Basically, it is a pressure wave moving through the air or water, or even though a solid. Visualizing such a wave can often make it much easier for us to understand it but how do we turn pressure into something we can see? By completing this exercise, you will: Gain an understanding of how sound waves look and work Learn to measure sound waves using computer programs Understand that complex waves (non-sinusoid) can be broken down into individual sine waves and represented visually this way harmonics are an example of this. Procedure: 1) Open the sound Thicket Tinamou. A Thicket Tinamou is a Central American bird, similar in shape to a quail. These birds are somewhat solitary and live in open dry forests. Using Syrinx or Raven, play the sound. Describe it. Is there a change in the sound over time? Does it go up or down? How does the volume (physicists use the term amplitude) change to your ear? Write your observations here: Think back on your physics training. What will the waveform that is this sound look like? Is it frequency modulated? Is it amplitude modulated? Draw what you think the sound would look like if you graphed the frequency (musicians use the term pitch ) of the sound over time. Your independent axis should be time, and the dependent axis should be frequency.

18 Now, use Syrinx or Raven to zoom in on the waveform of the sound. Does the wave appear as you predicted? Make several measurements of the period of the sound at various points in time. Wavelength is the time it takes for one cycle of the wave to pass a particular point it s also the time difference between two peaks or troughs. What is the period of the sound? What is the frequency of the sound? Is it changing over time? Is the waveform sinusoidal? 2) Now open the sound Red-breasted Nuthatch. This species is found in coniferous forests throughout western and northern North America. It often travels with other species and in pairs or small family groups. Is this sound amplitude modulated? Is it frequency modulated? What is the frequency range of the sound? Note the banded appearance of the sound. These are harmonics. What is the spacing of the harmonics? What is the lowest frequency band? What is the highest? Now measure the period of the sound. The waveform of this signal is distinctive. Describe how it differs from the Ticket Tinamou. Measure the carrier wave. What frequency is it at the beginning of the call? At the end? Which band in the spectrogram does this match? Is it the lowest or the highest? Now measure the modulating waveform. Do this by measuring from one peak amplitude to the next of similar amplitude. What is the period of this wave? What is the frequency of this waveform? Can you see any connection to the spectrograph? Recall the spacing of the harmonics.

19 3) Next, open Eastern Screech-owl. This species can be heard calling at all times of year in the eastern United States and southeastern Canada. It has a close relative in the west. Listen to the call. Try to predict what you think the waveform will look like from your prior experiences, and try to predict what the spectrograph will look like. Is there amplitude modulation? Is there frequency modulation? Try drawing this sound make two axes, one for time (the independent axis, on the bottom) and one for frequency/pitch (the dependent axis). Now look at the waveform. Was it at all like you predicted? This is a challenge, so don t worry if it s not really what you thought it would look like. However, did you hear the amplitude modulation? The screech-owl gives a series of pulses as its call, with a break between each note. It looks as though the note may continue between each louder note, but these are probably echoes. Measure the carrier frequency of the notes. Do you think there will be harmonics when you examine the spectrograph? Why or why not? Look at the spectrograph. Do your measurements agree with the spectrograph? Are there harmonics? The faint harmonics in this spectrograph result from the fact that each individual note has some amplitude modulation within it it starts a bit quieter, quickly builds up and then quiets again. Yet the carrier wave is still close to a sine wave. In this call, the strongest harmonic is the main carrier wave, and the other harmonics are much weaker. Measure the harmonics on the spectrograph. What is their relationship to the carrier wave? _

20 List of definitions: Amplitude: The pressure of the sound at the peak of a wave. This is more or less synonymous with volume. Frequency: The number of peaks that pass by a single point in one second. 20 Hz = 20 cycles per second. This means that 20 peaks of the wave pass by a single point in one second. 4 khz (kilohertz) means 4000 peaks pass a point in a single second. Frequency is the inverse of period. Period: The amount of time it takes for two consecutive peaks in a wave to pass a single point. Another way of thinking about it when analyzing a waveform is the amount of time between two peaks. Period is the inverse of frequency. Waveform: The representation of the a sound wave using Time on the X axis and pressure on the Y axis. Pressure increases and decreases relative to the ambient pressure on the microphone as successive sound waves pass by. This is turned into an electric signal and processed by your computer to show the waveform. Spectrograph: The representation of the sound wave using Time on the X axis and Frequency on the Y axis. Any wave can be mathematically broken down into component sine waves, and thus represented as a series of frequencies (of the sine waves) over time. Carrier Wave: The main sine wave that makes up a sound. Any deviations from this sine wave are caused by other components that cause harmonics in a spectrograph. See also modulating wave. Modulating Wave: The wave or waves that affect the main carrier wave. If you trace the peaks of each wave of the carrier wave you can sometimes get a sense of the modulating wave. In the Red-breasted Nuthatch call, the modulating waveform is amplitude modulated and sinusoid. Red-breasted Nuthatch (Sitta canadensis):

21 Eastern Screech-owl (Megascops asio)

22 The Physics of Musical Instruments Some Ideas for Teaching These ideas have NOT been classroom-tested, but a simple, fun project would be to have students download sounds from many different musical instruments and try to (a) categorize them into different groups based on their spectrograms, and (b) see if they can find any relationship between physical properties of the instrument and the spectrogram that it produces. The harmonics that are present in a particular instrument depend on the standing sound waves that the instrument can support. More details are available in a physics textbook (e.g., Physics for Scientists and Engineers, Paul A. Tipler, 1991, New York: Worth Publishers), but the basic idea is that instruments with an air column that is closed on one end and open on the other (like clarinets) act a bit like a string that is fixed at one end and free to move at the other. These instruments have a fundamental frequency (first harmonic) whose wavelength is around 4 times the length of the pipe, and the odd harmonics are usually much stronger than the even ones. Other instruments, meanwhile, such as a flute (two open ends), or a guitar (strings fixed at two ends), will have a fundamental frequency whose wavelength is around 2 times the length of the pipe or string, and all harmonics, both even and odd, can be present. Therefore, the spectrograms of these different musical instruments can be used to learn something about the size of the instrument, as well as its structure! Of course, real instruments are much more complicated than the theoretical description above, and there are many subtle aspects of each instrument s construction that influence the sound it makes and the relative strength of the harmonics that are present in its spectrogram. So the real value of an activity such as this one might simply be to give students a flavor of how rich and complex real-world physics can be, and how the physical construction of various instruments can be manipulated to give rise to a wonderful variety of sounds.