Session 3: Sound visualisation

Objectives: Humans can distinguish between the sound of a plane and the sound if a helicopter. Some people can even distinguish different types of planes only by hearing them. In this session the pupils will learn

 

  1. what a time signal is and how it looks like
  2. how the frequency analysis is used as a tool for sound and noise analysis.

Oscillations are periodic movements characterized by their amplitude and their frequency. By visualizing the time signal of a sound the pupils can “see” the movement.

With a frequency analysis of different sound sources (especially single tones) they can move on to a concept of sounds as a combination of different frequencies. This may help them to understand the nature of an “acoustic colour” and the difference between sound and noise.

The pupils should have knowledge about frequency or this session can be used to explain it.

A cooperation with a music teacher is useful to explain overtones.

Maximum duration

90 minutes

Material

Introduction/Starters

1) How can we visualize sounds?

Play music on a computer with windows media player. Turn “visualisation” to “graph”. The time-signal of the music will be displayed while playing. Questions: What is shown in the visualisation? How could we use that?

2)How can sounds be analysed?

Play music on a computer with windows media player. Turn “visualisation” to “bars”. The spectrum of the music will be displayed while playing. Questions: What is shown in the visualisation? How could we use that?

Main activities

1) Visualization of the time signal with an oscilloscope

Use a microphone with an oscilloscope to visualize the time signal. If no oscilloscope is available use a computer with soundcard and oscilloscope software (e.g. “scope”).

Explanation: Explain that the signal detected by the microphone shown in the oscilloscope is proportional to the air vibrations (or pressure / density variations) caused by the sound signal.

 2) Frequency analysis

Switch to frequency analysis. Make different sounds (whistle, hum, clap…) in front of the microphone and watch the meter. Try to whistle a tone with increasing pitch like when you call a dog. What is shown on which axis of the diagram? What does it mean when the graph is high on the left or right side of the diagram? When do we get peaks?

Explanation: In frequency analysis tools we have the sound level drawn on the vertical axis, the frequency on the horizontal axis running from low frequencies (“low tones”) on the left to the higher ones on the right. As an analogy you can use a piano, where the lower tones are played on the left and the higher ones are played on the right side of the keyboard. So we can see higher amplitudes on the left side when sound with more energy in the low frequency range is analysed. In every case we hear tones we will see peaks in the spectrum. The level is measured in Decibel (dB) with the highest possible level being the maximum level of the soundcard. Normally this level is set as “zero” so that any other levels in the range of the sound card are negative. For a soundcard recording with 16 bit dynamic range the sound level without any noise is
-96.3 dB.

3)Analyse the sounds of different sources using the software “scope”.

a) Observe the spectrum which you record when there is “no” noise in the room. What do you measure? What do you hear?

Explanation: The microphone is still recording/detecting sound, even when it is “quiet” in the room. The background noise of breathing, air fans, noise outside the room, rustling of clothes etc. won’t get the line to the ground. Even when these noises would miss, the electronics of the microphone and the soundcard would still produce some noise. When you do measurements you will have to keep that in mind to distinguish the sound you want to measure from the background noise.

 b) Measure the spectrum of a plucked string and/or of a sung (or “robot-like” spoken) vowel. List the frequencies of at least three of the lowest peaks and find a mathematical rule, which would create such frequencies.

Hint: Divide the higher frequencies by the lowest frequency.

Explanation: Most tonal sounds are produced by vibrating strings or air columns. Due to standing waves these vibrate with every possible frequency at the same time, thus producing a fundamental tone and overtones/harmonics with frequencies being multiples of the fundamental one. Our hearing will combine all these peaks in the spectrum to one single perception of a tone with a certain “colour”. The differences of the overtone-levels varying from vowel to vowel will help us distinguish a “u” from an “e” and understand speech.

 

c) Measure the spectrum of a tuning fork and draw a sketch of the measured curve

Mind the right labelling of the axis. What would the spectrum of two tuning forks with different tones sounding simultaneously be like? Measure the main frequency of the forks sound.

Hint: You can amplify the tuning forks by using a resonant body like an open box or the table. You can tune forks by adding mass, e.g. wrapping a paper clip around the vibrating parts.

Explanation: This should give a peak in the measured spectrum concerning to the main frequency of the tuning fork. Two different forks should show two peaks in the spectrum. In software like “scope” you can use a drag-and-drop-cursor to easy measure frequencies. Tuning forks have poor overtones/harmonics so a single peak for one fork should be seen in the spectrum. May be some harmonics can be detected but with an obvious smaller amplitude.

 d)Measure the spectrum of noise

(e.g. “white noise” of a noise generator or a radio/tv without being tuned to a station, a spoken “shhhh”) and draw a sketch of the measured curve. What are the differences to the spectrum of the tones?

Hint: You will need a higher volume for the noise than the tones for a proper measurement.

Explanation: The spectrum of ideal white noise is a flat line, showing that there is the same energy in every frequency (same sound levels). The measured spectra should be more “curvy” according to the quality of the speakers, the microphone or the mouth position. In random noise the sound energy is more or less evenly distributed to many frequencies as opposed to tonal sounds where the energy is concentrated in one or more peaks.

 e)  Measure the spectra of different disturbing noise

and rate them according to their disturbance (e.g. fan noise, traffic noise, aeroplane noise, wind noise, white noise). Is there a correlation between the rating and the “peakyness” (=tonality) of the spectrum?

Explanation: An important psycho-acoustical parameter is the tonality of sounds. In most cases noises with tonal components will be rated to be more disturbing than noises with more “flat” spectra.

Lessons learned

A sound spectrum shows how the sound energy is distributed to different frequencies. Tonal sounds show a spectrum with one or more peaks. Random noise shows a continuous spectrum with a flat line. Realistic sounds combine both elements.