How Fast Does Sound Travel? Five Experiments to Measure the Speed of Sound
Can you measure something you cannot see? With a microphone, a tuning fork, a tube and some clever physics, the answer is yes.
Hemel Private Tuition | Practical Physics | GCSE and A-Level Physics
Sound Is All Around Us — But How Fast Is It Moving?
Imagine standing in a field and watching someone hit a metal object with a hammer.
You see the hammer strike before you hear the sound.
Or perhaps you have watched a thunderstorm. There is a brilliant flash of lightning, followed some seconds later by a rumble of thunder.
Why the delay?
The answer is that light and sound travel at enormously different speeds.
Light moves at approximately 300 million metres per second in a vacuum, whereas sound travels through ordinary air at roughly 340 metres per second.
That is an extraordinary difference.
But how do we know that sound travels at around 340 metres per second? More importantly, could we measure it ourselves?
At Hemel Private Tuition, I particularly enjoy experiments where students can discover a physical quantity rather than simply accepting a value printed in a textbook.
The speed of sound provides an excellent opportunity because there is more than one way to measure it.
We can use resonance, echoes, microphones, oscilloscopes and computer-based data logging. We can even investigate why different experimental methods sometimes give slightly different answers.
That final point is especially important for A-Level Physics.
A good experiment does not simply produce the expected answer. It produces evidence that we can analyse, question and improve.
1. What Actually Is Sound?
Before measuring the speed of sound, we need to understand what is moving.
Sound is a mechanical wave produced by vibrations.
When a loudspeaker vibrates, it causes neighbouring air molecules to move backwards and forwards. Those molecules interact with neighbouring molecules, passing the disturbance through the air.
The molecules themselves do not travel all the way from the loudspeaker to your ear. Instead, energy travels through the medium.
In air, sound waves are longitudinal waves.
This means the particles vibrate parallel to the direction in which the wave travels.
We describe the alternating regions as:
Compressions: regions where the pressure and density are slightly higher.
Rarefactions: regions where the pressure and density are slightly lower.
Sound requires a medium, such as a gas, liquid or solid. It cannot travel through a perfect vacuum.
Its speed depends upon the properties of that medium.
Sound generally travels much faster through water and steel than through air, although the precise values depend on temperature and material properties.
The fundamental wave equation
For a periodic sound wave:
Wave speed = frequency × wavelength
In symbols:
v = f × λ
Where:
v = wave speed in metres per second (m/s)
f = frequency in hertz (Hz)
λ = wavelength in metres (m)
For example, suppose we produce a sound with a frequency of 500 Hz and measure its wavelength as 0.686 m.
Then:
v = 500 × 0.686
v = 343 m/s
That is close to the expected speed of sound in air at room temperature.
But the important question is how we can measure that wavelength — or alternatively measure how long sound takes to travel a known distance.
2. Experiment One — Measuring the Speed of Sound Using Resonance
This is one of the classic experiments in sound physics.
It is elegant, relatively inexpensive and gives students a direct connection between sound waves and standing waves.
Equipment
A resonance tube, water reservoir or adjustable piston, tuning forks of known frequency, metre rule, thermometer and optional microphone or sound-level sensor.
The principle
A tube containing air can resonate when the frequency of the sound matches one of its natural frequencies.
Consider a tube that is open at one end and effectively closed at the other by water.
At the water surface, the air has approximately a displacement node.
At the open end, it has approximately a displacement antinode.
For the lowest resonant frequency, the effective length of the air column is approximately one-quarter of a wavelength.
Therefore:
L = λ / 4
Rearranging:
λ = 4L
And:
v = 4fL
This simplified equation ignores a small correction associated with the open end of the tube.
Practical procedure
Fill the resonance apparatus with water, leaving a short air column.
Strike a tuning fork gently with a suitable rubber striker.
Hold the vibrating fork just above the opening without touching the tube.
Gradually lower the water level, increasing the length of the air column.
Listen for the point at which the sound becomes noticeably louder.
Measure the length of the air column.
Repeat using different tuning forks.
At resonance, the sound will become noticeably more intense.
This is not because the tuning fork suddenly begins vibrating faster. The air column is responding particularly strongly to the frequency being produced.
Worked example
Suppose a tuning fork has a frequency of 512 Hz.
The first resonance occurs when the measured air-column length is 0.160 m.
Ignoring the end correction:
λ = 4 × 0.160 = 0.640 m
v = 512 × 0.640
v = 327.7 m/s
The measured value is slightly lower than the expected value of approximately 343 m/s at 20°C.
Why?
One important reason is that the displacement antinode is actually located a little beyond the open end of the tube.
The effective resonating length is therefore slightly longer than the measured tube length.
An A-Level improvement — use two resonances
A particularly good method is to find two consecutive resonance positions.
For a tube closed at one end:
First resonance: L1 + e = λ / 4
Second resonance: L2 + e = 3λ / 4
Here, e represents the end correction, assuming it remains approximately constant.
Subtracting the equations gives:
L2 - L1 = λ / 2
Therefore:
λ = 2(L2 - L1)
And:
v = 2f(L2 - L1)
This method has an important advantage.
The end correction cancels out.
Suppose the first resonance occurs at 0.160 m and the second at 0.495 m.
The difference is:
0.495 - 0.160 = 0.335 m
Therefore:
λ = 2 × 0.335 = 0.670 m
v = 512 × 0.670
v = 343.0 m/s
An excellent result.
These are illustrative measurements rather than actual laboratory observations, but they demonstrate how using two resonance positions can improve the method.
What students learn
This experiment brings together frequency, wavelength, standing waves, nodes, antinodes, resonance and experimental corrections.
It is also an opportunity to discuss why taking more measurements does not necessarily remove a systematic error.
3. Experiment Two — Two Microphones and the Time Taken for Sound to Travel
This is one of my favourite ways of introducing modern experimental physics.
Instead of measuring a wavelength, we measure a time interval.
The idea is remarkably simple.
Place two microphones a known distance apart. Produce a short, sharp sound near the first microphone. Measure how long the sound takes to reach the second microphone.
Then calculate the speed.
Equipment
Two microphones, a dual-channel oscilloscope or simultaneous-channel audio recording interface, a metre rule, connecting cables, and a sharp sound source such as a finger snap or hand clap.
Setting up the experiment
Imagine two microphones arranged along a straight line.
The first microphone is positioned near the sound source.
The second microphone is placed exactly 1.00 m farther away.
When a sound pulse is produced, the first microphone detects it before the second.
The oscilloscope displays two signals with a small difference in arrival time.
The relationship is:
v = d / t
Where:
d = distance between microphones
t = difference between arrival times
v = speed of sound
Worked example
Distance between microphones = 1.00 m
Time delay = 0.00292 s
Therefore:
v = 1.00 / 0.00292
v = 342.5 m/s
This is close to the expected value at room temperature.
Why use an oscilloscope?
An oscilloscope makes extremely short time intervals visible.
A delay of just three milliseconds would be difficult to measure directly with a stopwatch.
However, on an oscilloscope it becomes the horizontal separation between two waveform features.
Students can measure the time difference using cursors.
With a dual-channel instrument, both signals must share a reliable common timebase.
This experiment is also used in university teaching demonstrations, including an example documented by the University of Maryland.
Making the experiment more accurate
I would encourage students to repeat the experiment at several different microphone separations.
For example:
Separation (m) | Illustrative delay (ms) |
|---|---|
0.50 | 1.46 |
1.00 | 2.92 |
1.50 | 4.37 |
2.00 | 5.83 |
If students plot separation (vertical axis) against time delay (horizontal axis), the gradient gives the speed of sound.
Gradient = change in distance / change in time
This is often better than relying upon one result.
It also introduces a crucial A-Level concept: a straight-line graph can identify the physical quantity we want while helping reveal systematic offsets.
A subtle problem
What happens if the microphone signals are not processed at exactly the same time?
Some recording systems introduce delays between channels.
If the channels are not synchronised, students could mistake an electronic delay for the travel time of sound.
For a sound travelling just one metre, even a small timing offset may create a significant error.
I would use a single synchronised two-channel system, where possible, and reverse the microphones as a useful check. The sound source should remain collinear with the microphones, and the pulses must not be so loud that they cause clipping.
The apparent simplicity of this experiment makes it ideal for discussing the difference between a straightforward formula and a genuinely reliable measurement.
4. Experiment Three — Measuring the Speed of Sound Using an Echo
An echo is sound that has travelled to a reflecting surface and returned to the listener.
It gives us another way to measure the speed of sound.
If sound travels a distance d to a wall and then returns, its total distance is 2d.
Therefore:
v = 2d / t
Where t is the total round-trip time.
The outdoor experiment
Imagine standing 50 m from a large, flat wall.
You produce a sharp sound and record the interval until its echo returns.
Suppose the time interval is 0.292 s.
The sound travels:
2 × 50 = 100 m
Therefore:
v = 100 / 0.292
v = 342.5 m/s
A pleasingly simple experiment.
However, a stopwatch is not an ideal instrument for measuring such a short interval.
Human reaction times are much too significant.
Instead, we can use a microphone and audio-recording software.
The waveform should show an initial pulse followed by a reflected pulse.
The separation between these events gives the echo delay.
An indoor laboratory version
We can reproduce the same principle using a long tube with a closed end.
Place a microphone close to the open end, pointing along the tube.
Produce a short sound pulse near the microphone.
The pulse travels down the tube, reflects at the closed end and returns.
Using a tube length of 1.50 m, the approximate return time would be:
t = 2 × 1.50 / 343
t = 0.00875 s
That is 8.75 milliseconds.
With suitable data logging, this is easy to display.
PASCO has published experiments using precisely this general technique, including configurations with a sound sensor, a closed tube and Capstone software.
The challenge of reflections
In a real laboratory, sound may reflect from walls, benches, equipment and other objects.
The microphone may detect several overlapping pulses.
Students must identify which signal is the required echo.
This is useful experimental problem-solving: the apparatus may be working perfectly, yet interpreting the signal still requires care.
For an outdoor echo experiment, choose a safe location away from roads and avoid very loud impulsive sounds. A computer-generated click through a modest loudspeaker can be a more controlled source.
5. Experiment Four — Using an Oscilloscope to Measure Wavelength
There is another approach that combines two microphones with a continuous sound of known frequency.
Instead of recording a single pulse, we use a steady tone.
A loudspeaker connected to a signal generator produces a sound at a chosen frequency.
Two microphones detect the resulting wave.
As one microphone is moved, its signal changes phase relative to the other.
When the signals return to the same phase, the microphone has moved by one whole wavelength, provided we follow successive alignments.
The procedure
Generate a continuous tone, perhaps at 1,000 Hz.
Place two microphones along the direction of propagation.
Connect them to separate oscilloscope channels.
Keep one microphone fixed.
Move the other microphone slowly away.
Identify successive positions where the waveforms are aligned in phase.
Measure the separation between these positions.
At 1,000 Hz and a sound speed of 343 m/s:
λ = v / f
λ = 343 / 1000
λ = 0.343 m
So the distance between successive in-phase positions should be approximately 34.3 cm.
The experiment demonstrates that sound behaves as a wave rather than merely as a sequence of isolated events.
A useful extension
Repeat the experiment at different frequencies.
Students should discover that, in ordinary air under the same conditions, the speed remains approximately constant while wavelength decreases as frequency increases.
A 500 Hz tone has approximately twice the wavelength of a 1,000 Hz tone.
That provides a convincing practical demonstration of:
v = f × λ
There are some practical complications. Reflected waves can disturb the phase measurements, and moving a microphone sideways rather than along the sound path will create misleading results.
A directional sound source, careful alignment and a suitable distance from reflective surfaces all help.
6. Experiment Five — Using PASCO Sensors and Data Logging
A major benefit of computer-based data logging is that it allows the student to inspect what actually happened during an experiment.
Instead of simply writing down one number, we can examine the waveform itself.
PASCO Capstone can display sound waveforms in a scope-style view. PASCO also provides published methods for measuring sound speed from reflected pulses, including a procedure sampling at 20 kHz.
A suggested PASCO investigation
I would use a sound sensor and an adjustable closed-end tube to perform an echo experiment.
The procedure would be:
Measure the tube length.
Position the microphone close to the open end.
Set an appropriately high sampling rate.
Record a short impulse and the returning reflection.
Use the software cursors to measure the time interval.
Calculate the speed of sound.
Change the tube length and repeat.
Why sampling rate matters
Suppose the equipment records 20,000 measurements per second.
The interval between samples is:
1 / 20,000 = 0.00005 s
This is 50 microseconds.
During that time, sound travelling at 343 m/s moves approximately:
343 × 0.00005 = 0.01715 m
That is around 1.7 cm.
It means choosing a waveform peak only to the nearest sample can introduce appreciable timing uncertainty, particularly over short distances.
Students can investigate whether increasing the tube length reduces the percentage uncertainty.
They can also examine whether the waveform provides a sharply defined event or a less obvious series of peaks.
One caution: not every PASCO sound sensor or interface supports simultaneous high-speed recording from multiple microphones. The actual hardware specifications and recording mode must be checked before attempting a two-microphone time-of-flight experiment.
For this reason, a single-sensor echo method is often the more straightforward starting point.
7. Does Sound Always Travel at the Same Speed?
Here is a fascinating question to ask a student:
Would sound travel faster through the air in our laboratory on a hot summer afternoon or on a cold winter morning?
The answer is that sound travels faster through warmer air.
For ordinary temperatures, we can use the approximate relationship:
v = 331 + 0.6T
Where:
v = speed of sound in m/s
T = air temperature in degrees Celsius
This is a useful approximation for air under normal conditions.
Approximate theoretical values for air, calculated using v = 331 + 0.6T.
At 0°C:
v = 331 + (0.6 × 0) = 331 m/s
At 20°C:
v = 331 + (0.6 × 20) = 343 m/s
At 30°C:
v = 331 + (0.6 × 30) = 349 m/s
That difference is measurable.
Why does temperature matter?
Increasing temperature increases the average kinetic energy of the air molecules.
The relationship between temperature and the speed at which pressure disturbances propagate can be described using the properties of a gas.
At A-Level, we can go further:
v = sqrt(γRT / M)
Where:
γ = ratio of specific heat capacities
R = molar gas constant
T = absolute temperature in kelvin
M = molar mass of the gas
For a given ideal gas composition, the speed increases in proportion to the square root of absolute temperature.
This also explains why we should measure the laboratory temperature whenever we undertake a speed-of-sound experiment.
Comparing an experimental value directly with 343 m/s without checking the temperature is not always appropriate.
A further investigation
Students could repeat one of the experiments on different days and compare the results against the predicted temperature dependence.
For an indoor experiment, this is preferable to deliberately heating air in an open laboratory, where temperature gradients and convection might introduce further complications.
Humidity also affects sound speed, although temperature is generally the dominant factor in ordinary school experiments.
8. Which Experiment Gives the Best Answer?
This is where the experiments become particularly valuable for A-Level students.
Different methods may produce different results even when measuring the same physical quantity.
Consider the following fictional set of results, illustrating what students might obtain at a laboratory temperature of 20°C.
Method | Measured speed | Difference from 343 m/s |
|---|---|---|
Resonance, one position | 328 m/s | -4.4% |
Resonance, two positions | 343 m/s | 0.0% |
Two microphones | 342 m/s | -0.3% |
Echo method | 346 m/s | +0.9% |
Phase and wavelength | 339 m/s | -1.2% |
These are illustrative results, not measurements from my laboratory.
The important question is not simply which experiment produced a value closest to 343 m/s.
We should ask why the answers differ.
Random errors and systematic errors
Random errors cause repeated measurements to vary.
For example, students may find it difficult to identify the exact resonance position because the sound becomes louder gradually rather than suddenly.
Systematic errors can shift every measurement in one direction.
Examples include:
Ignoring the end correction in a resonance tube.
Using microphones with a fixed electronic timing offset.
Measuring the wrong acoustic path.
Selecting a reflected waveform instead of the direct sound.
Using an assumed temperature rather than the measured air temperature.
Repeating measurements can help reduce the influence of random errors, but it will not automatically eliminate systematic errors.
This distinction is frequently important in GCSE and A-Level experimental questions.
Calculating percentage difference
We can compare an experimental result against a predicted value using:
Percentage difference = |measured value - predicted value| / predicted value × 100
If the measured speed is 339 m/s and the predicted value is 343 m/s:
Percentage difference = |339 - 343| / 343 × 100
Percentage difference = 1.17%
However, a small percentage difference does not automatically mean a measurement is reliable.
We must also consider its uncertainty.
For example, a result of 343 ± 15 m/s is less precise than one of 341 ± 2 m/s, even though the first happens to agree exactly with our expected value.
That is a distinction I particularly want students to understand.
9. Measuring Uncertainty — An A-Level Challenge
Suppose students use two microphones separated by:
d = 1.000 ± 0.005 m
They measure a time interval of:
t = 2.92 ± 0.05 ms
Their calculated speed is:
v = 1.000 / 0.00292 = 342.5 m/s
For independent uncertainties in distance and time, a first-order estimate of the fractional uncertainty is:
Δv / v = sqrt[(Δd / d)^2 + (Δt / t)^2]
Substituting:
Δv / v = sqrt[(0.005 / 1.000)^2 + (0.05 / 2.92)^2]
Δv / v = 0.0178
This corresponds to approximately 1.8%.
The absolute uncertainty is:
Δv = 342.5 × 0.0178
Δv = 6.1 m/s
So we might report:
v = 343 ± 6 m/s
This uncertainty includes only the quoted distance and timing contributions. It does not account for any unrecognised systematic errors.
For a conservative school-level approach, adding the percentage uncertainties gives a larger estimate of about 2.2%.
Either way, students begin to see that reporting a measurement properly is more than writing down the result of a calculation.
10. Can We Combine the Experiments into One Investigation?
This is the experiment I would be particularly interested in running at Hemel Private Tuition.
Rather than carrying out each method in isolation, I would organise a comparative investigation.
Students would measure the speed of sound using three approaches:
Resonance in an air column.
Time of flight between two microphones.
The delay between an initial sound pulse and its echo.
All three methods would be used under approximately the same laboratory conditions.
We would measure the air temperature, repeat each experiment and record the uncertainty.
Then we would compare the findings.
Interactive experiment comparison
Select a method to see what students measure and the main source of uncertainty.
What to measure
Frequency and the distance between successive resonance positions.
Equation
v = 2f(L2 - L1)
Main uncertainty
Identifying resonance positions, frequency uncertainty and acoustic end effects.
An investigation suitable for GCSE and A-Level
At GCSE level, students could focus on calculating speed, drawing graphs and explaining differences between results.
At A-Level, I would extend the exercise to include:
Uncertainty propagation.
Distinguishing precision from accuracy.
Identifying systematic errors.
Using best-fit lines and gradients.
Evaluating assumptions about wave behaviour.
Deciding whether different results agree within their uncertainties.
It is an excellent example of how one physical concept can support several levels of teaching.
11. From the Laboratory to the Real World
The speed of sound is much more than a textbook number.
Thunderstorms
When we see lightning, the light arrives almost instantaneously compared with the sound.
At approximately 343 m/s, sound travels about one kilometre in three seconds.
This gives us a rough method of estimating the distance to a lightning discharge.
However, this is not a reliable safety test. A storm can still be dangerous even if the thunder appears distant.
Medical ultrasound
Ultrasound scanners transmit high-frequency sound waves into the body and detect reflections from boundaries between tissues.
By measuring the time taken for reflected pulses to return, an instrument can estimate depth.
The principle resembles our laboratory echo experiment, although the speed of sound within tissue is different from that in air.
Sonar
Boats and ships use sound waves to investigate what lies beneath the water.
A sonar system sends out a pulse and measures the return time.
Depth is calculated using the round-trip distance and the appropriate sound speed in water.
The speed varies with temperature, salinity and pressure.
Music and musical instruments
Wind instruments depend upon standing waves and resonance.
The length of an air column helps determine the frequencies at which it resonates.
The same physics that explains our resonance tube helps us understand organ pipes, flutes and clarinets.
As someone interested in both physics and music, I find this connection particularly satisfying.
The mathematics of waves is not confined to the science laboratory. It is fundamental to the sounds we hear in music every day.
12. Why I Believe Students Should Carry Out These Experiments
After many years of teaching, I have become increasingly convinced that there is an important difference between recognising a scientific equation and understanding what that equation means.
A student might remember:
Speed = distance / time
They might also remember:
Wave speed = frequency × wavelength
And they might be perfectly capable of using both equations in an examination.
But ask them to measure the speed of something they cannot see, and the problem becomes much more interesting.
How do we know where the sound is?
How can we measure a time interval of a few milliseconds?
Why does a tuning fork become louder at certain air-column lengths?
Why do two apparently correct methods give different answers?
What does it mean when our data do not agree with the textbook?
These are the questions that turn a practical into a genuine learning experience.
One advantage of teaching in a well-equipped home laboratory is the opportunity to move between traditional experiments and modern instruments.
A tuning fork and resonance tube can sit alongside an oscilloscope, computer-based data logger and PASCO sensors.
Students can then see how different measuring techniques reveal different aspects of the same physical phenomenon.
I would rather have a student obtain a value of 336 m/s, recognise why it might be inaccurate and suggest a sensible improvement than produce 343 m/s without understanding how the measurement was made.
The first student is beginning to think like an experimental physicist.
13. Bringing Practical Physics to GCSE and A-Level Tuition
At Hemel Private Tuition in Hemel Hempstead, I teach GCSE and A-Level Physics with a strong emphasis on practical understanding.
My laboratory allows students to investigate scientific principles using real equipment rather than relying exclusively on diagrams, simulations or videos.
We can combine experiments with digital data logging, graph analysis, worked calculations and examination-style questions.
Online students can also benefit from live demonstrations using multiple cameras, close-up views of the equipment and screen sharing of experimental data.
The intention is not simply to make Physics entertaining.
It is to help students connect scientific theory with experimental evidence and develop the confidence to tackle unfamiliar examination questions.
For A-Level students in particular, the ability to interpret data and evaluate experimental methods is just as important as remembering formulae.
Conclusion — Measuring the Invisible
Sound travels through the air around us every moment of the day.
We hear it, recognise it and use it to communicate.
Yet the disturbance itself is invisible.
With a tuning fork and a tube, we can measure its wavelength.
With two microphones, we can measure how long it takes to travel a known distance.
With an echo, we can investigate the journey of a sound pulse to a reflecting surface and back.
And with an oscilloscope or data logger, we can make the behaviour of sound visible on a screen.
The fascinating part is that these experiments use different methods but should lead us towards the same physical answer.
They also teach something more valuable than a single numerical value.
They demonstrate that science depends upon measurement, evidence, uncertainty and the willingness to question our results.
The speed of sound may be approximately 343 metres per second at room temperature. But discovering how to measure it is where the real learning begins.
For GCSE and A-Level Physics tuition, including practical laboratory work, visit Hemel Private Tuition.
