Analyze Sound with a Smartphone or Computer

Learn how to measure sound levels, frequencies and periods, and how to interpret waveforms, spectra and spectrograms using a smartphone or computer. Includes 12 hands-on acoustics experiments with FizziQ.

Introduction - Can you study sound with a smartphone or a computer?
What is a sound?
How does a microphone turn sound into data?
Observing the time shape of a sound
Measuring the frequency of a sound
Understanding the sound spectrum
Understanding the spectrogram
Measuring sound level and understanding decibels
Can a smartphone or a computer replace a sound level meter?
Producing sounds for experiments
Smartphone or computer: which tool should you choose?
Combined smartphone-and-computer workflows
Twelve sound experiments with FizziQ
Common mistakes and confusions
Frequently asked questions
Conclusion
Sources and references

Introduction - Can you study sound with a smartphone or a computer?

Yes. The microphone of a smartphone, tablet or computer converts the small pressure variations of the air - sound - into an electrical signal, then into a digital signal. Once digitized, that signal can be displayed and analyzed: you can observe its shape over time, measure its period and frequency, compute its spectrum, follow the evolution of its frequency content with a spectrogram, and estimate a sound level in decibels.

Each type of device has its strengths. The smartphone or tablet is a mobile instrument: you take it into the field, move it around a source, measure in real time, and even use it as a sound generator. The computer is an analysis workstation: its large screen makes it easy to examine an oscillogram or a spectrogram in detail, and it lets you import audio files, compare recordings and write a structured report. The two uses complement each other: the smartphone easily captures and measures a real phenomenon, while the computer makes detailed analysis and interpretation of the signal easier.

One precaution from the outset: the quality of the results depends on the microphone used, on the processing applied by the device, and on calibration. A smartphone is not a certified sound level meter, and two different phones will not display exactly the same sound level. These limits, detailed throughout this guide, do not prevent excellent experiments: they simply define the frame within which the measurements are meaningful.

This guide presents, in order, the physical nature of sound, the measurement chain (from microphone to software), the three main representations of the signal (oscillogram, spectrum, spectrogram), the measurement of sound level and the decibel scale, the production of sounds for experiments, and finally twelve experiments that can be carried out with FizziQ on a smartphone (mobile app) or on a computer (FizziQ Web).

Part 1 - What is a sound?

1.1 A vibration that propagates

A sound always begins with a vibration: a guitar string oscillating, a loudspeaker membrane moving back and forth, vocal folds vibrating. This vibration sets the surrounding medium in motion. In air, the molecules close to the source are alternately compressed and rarefied: the local pressure rises slightly, then falls slightly, around atmospheric pressure.

These pressure variations are transmitted from one region to the next: each slice of air pushes the next one, which pushes the next. It is a pressure wave that propagates - the air molecules themselves do not travel with the wave: they oscillate around their average position. In air, this wave is longitudinal: the molecules oscillate in the same direction as the wave propagates.

An important consequence: sound needs a material medium to propagate. In a vacuum, there are no molecules to set in motion, hence no sound. That is why space is silent.

Key point. In air and liquids, sound propagates as a longitudinal pressure wave, transmitted step by step, and it does not propagate in a vacuum. In solids, several kinds of elastic waves can coexist, notably longitudinal and transverse waves.

1.2 The main quantities

Several physical quantities are used to describe a sound.

The period T is the duration of one complete cycle of the vibration, in seconds (s). For a periodic sound, the signal repeats itself identically every T seconds.
The frequency f is the number of cycles per second, in hertz (Hz). It is the inverse of the period:

f = 1/T

with f in hertz (Hz) and T in seconds (s). Example: a signal whose period is T = 2.27 ms has a frequency f = 1/0.00227 ≈ 440 Hz, the frequency of the standard A4 tuning fork.

The amplitude characterizes the size of the variation: for the sound wave, it is the maximum deviation of the acoustic pressure from atmospheric pressure; for the digitized signal, it is the maximum deviation of the signal value from zero.
The speed of sound c is the propagation speed of the wave in the medium, in meters per second (m/s). In air it is about 343 m/s at 20 °C (textbooks often use the rounded value 340 m/s). It increases with temperature.
The wavelength λ is the distance traveled by the wave during one period, in meters (m). It links space and time:

c = λ f

with c in m/s, λ in m and f in Hz. Example: a 440 Hz sound in air at 20 °C has a wavelength λ = 343/440 ≈ 0.78 m.

The acoustic pressure p is the small pressure variation due to the wave, measured in pascals (Pa). It is tiny compared with atmospheric pressure (about 101,325 Pa): a conversation corresponds to an effective acoustic pressure of the order of 0.02 Pa.

1.3 Physical phenomenon and auditory perception

Physical quantities describe the wave; perception describes what our auditory system makes of it. The two are related but do not coincide.

Frequency is a measurable physical quantity; pitch (low or high) is a sensation. For a pure tone, perceived pitch and frequency move together, but for a complex sound the perceived pitch generally corresponds to the fundamental frequency, even when it is not the strongest component - and sometimes even when it is physically absent from the signal (the “missing fundamental” phenomenon).
Acoustic pressure is a physical quantity (in pascals), of which the sound pressure level in decibels is a logarithmic representation; perceived loudness is the associated sensation, which also depends on frequency: at equal level, a 100 Hz tone seems quieter than a 1,000 Hz tone, because the human ear is less sensitive to low frequencies.
The spectral content (how the energy is distributed among frequencies) largely determines the timbre, which is why a piano and a flute playing the same note do not sound alike.

Hearing is therefore not a simple, direct readout of frequency and amplitude: it is a complex, non-linear, frequency-dependent process. This guide focuses on measurable physical quantities and points out, where useful, how they differ from perception.

Watch out for this confusion. “Frequency” and “loudness” are independent: you can play a very high-pitched sound very quietly, or a very low-pitched sound very loudly. Frequency is measured in hertz, sound level in decibels.

Part 2 - How does a microphone turn sound into data?

2.1 How it works in general

A microphone converts a pressure wave into an electrical signal, and the device’s electronics then convert that signal into numbers. The complete chain has several stages:

The membrane. A thin membrane is set vibrating by the pressure variations of the air.
Acoustic-to-electrical transduction. The motion of the membrane is converted into an electrical voltage. Smartphones and computers almost always use MEMS microphones (micro-electro-mechanical systems), miniaturized and based on a capacitive principle: the membrane forms one plate of a capacitor whose capacitance varies with the vibration.
Preamplification. The very weak electrical signal is amplified.
Analog-to-digital conversion. The continuous signal is turned into a sequence of numbers through two operations: sampling (the signal is measured at regular intervals, for instance 44,100 times per second) and quantization (each measurement is rounded to one of the available numerical values, for instance 65,536 values for 16-bit quantization).
Storage or processing. The resulting sequence of numbers - the digital signal - can be saved to a file or analyzed in real time by software.

The sampling rate limits the frequencies that can be observed: according to the Shannon-Nyquist theorem, a signal sampled at frequency f_s can only correctly represent frequencies below f_s/2. With 44,100 Hz sampling, sounds up to about 22,000 Hz can be analyzed, which covers the entire audible range.

2.2 Smartphone and computer microphones

Not all built-in microphones are equal, and their differences affect the measurements.

Position. On a smartphone, the main microphone is usually on the bottom edge; covering it with a finger or a case distorts the measurement. On a laptop it is often near the keyboard or the webcam, sometimes close to the fan - a source of spurious noise.
Directivity. A microphone does not pick up sound equally from all directions; the orientation of the device relative to the source changes the received signal.
Sensitivity and frequency response. Built-in microphones are optimized for voice. Their response is not uniform across the spectrum: very low frequencies (below roughly 100 Hz) and very high frequencies are often attenuated, to an extent that varies from one model to another.
Self-noise. Every microphone produces a slight signal even in silence; this noise floor limits the measurement of very quiet sounds.
Clipping. Beyond a certain level, the signal is clipped: the peaks of the wave are cut off and the measurement becomes meaningless.
Differences between devices. Two different smartphone models, or a smartphone and a computer, will give different measurements for the same sound. A good external microphone plugged into a computer can significantly improve acquisition fidelity.

2.3 Automatic processing

The operating systems of smartphones and computers sometimes apply processing to the microphone signal designed for telephony and video calls: automatic gain control (the amplification changes to keep the level constant), noise reduction, low-frequency filtering, compression, echo cancellation, voice-enhancement processing.

This processing is useful for a conversation, but it can distort a scientific experiment: automatic gain control makes it impossible to compare two amplitudes, and noise reduction can erase precisely the faint sound you were trying to measure. Whether this processing is present, and how strong it is, depends on the device model and operating system, and the app cannot always disable it. This is one more reason to favor relative comparisons made with the same device under the same conditions.

2.4 What the software can actually display

It is essential to distinguish what is measured from what is computed:

The device measures a sequence of digital samples, an image of the voltage produced by the microphone.
The amplitude displayed by an oscillogram is a relative quantity, expressed in the internal unit of the digitized signal - it is not directly a pressure in pascals.
The sound level in decibels is estimated from the energy of the digital signal, using an approximate calibration specific to the device.
The dominant frequency is computed by an analysis algorithm (the microphone does not “measure” a frequency).
The spectrum and the spectrogram are computed by Fourier transform on portions of the signal.

Key point. The microphone provides only one thing: a digitized signal proportional to the acoustic pressure. Frequency, spectrum, spectrogram and sound level are quantities computed by the software from that signal, each with its own assumptions and limits.

For a broader view of how the microphone compares to the phone’s other sensors, see Smartphone Sensors: How They Work, Accuracy, and Scientific Uses.

Part 3 - Observing the time shape of a sound

3.1 The waveform (oscillogram)

The waveform, also called an oscillogram, shows how the sound signal evolves over time: the horizontal axis is time, the vertical axis is the amplitude of the signal. It is the most direct representation of the digitized signal - the one historically displayed by an oscilloscope connected to a microphone.

The oscillogram makes it possible to tell several families of signals apart at a glance:

a periodic signal shows a pattern that repeats identically (a held musical note, a sung vowel, a synthesizer tone);
a sinusoidal signal - a “pure tone” - is the simplest periodic signal: a single, regular sine wave;
an aperiodic signal shows no repeating pattern;
noise is an aperiodic signal with a disordered appearance (hiss, applause, white noise);
a transient is a brief, unrepeated event (a clap, an impact, the attack of a note).

In FizziQ, the oscillogram is available on smartphone with the Amplitude instrument (Microphone category), which displays the waveform in real time with two switchable observation windows, and on computer in the Amplitude mode of the FizziQ Web audio analyzer, which displays the waveform of a recording or an imported file.

3.2 Measuring a period

Measuring a period on an oscillogram is one of the most instructive measurements in acoustics. The method:

Display a periodic signal (for instance a tuning fork or a sound from the FizziQ library) and zoom in until a few complete cycles are visible.
Locate two identical points of two successive cycles - for instance two consecutive maxima, or two zero crossings in the same direction.
Measure the time between them: that is the period T.
Compute the frequency: f = 1/T.

To improve precision, measure the duration Δ t of n consecutive periods, then divide: T = Δ t / n. The pointing uncertainty is thus spread over n periods instead of one. Example: if you measure 22.7 ms for 10 periods, then T = 2.27 ms and f ≈ 440 Hz; a pointing error of 0.2 ms now represents only 0.02 ms on the period, i.e. less than 1% error on the frequency.

3.3 Observing time-domain phenomena

The oscillogram reveals structures that the spectrum does not show:

the envelope of the signal, i.e. how its amplitude evolves over time;
the attack (rapid rise of amplitude at the start of a note) and the decay (gradual fading), characteristic of each instrument;
beats: when two sounds of close frequencies f₁ and f₂ are superimposed, the total amplitude oscillates slowly at the frequency f_{beat} = |f₁ - f₂|, visible as an undulating envelope;
echoes: attenuated, delayed repetitions of a brief sound;
reverberation: the diffuse prolongation of a sound in a room, visible as a gradual decay after the source stops;
saturation or clipping: the tops of the waveform appear “shaved off”, flattened at a maximum value - a sign that the signal has exceeded the capacity of the microphone or converter and that the measurement is no longer reliable.

Measurement limits. A clipped signal no longer allows a reliable quantitative analysis of amplitude or spectral content: clipping creates artificial harmonics and underestimates the true level. Some temporal information, such as the period, may nonetheless remain usable. Before any measurement, check on the oscillogram that the signal peaks remain well below the maximum.

3.4 Choosing the tool

Smartphone or computer?

With FizziQ mobile (Amplitude instrument): observe a sound live as it happens, move the microphone around a source, quickly compare several places or sources, run an experiment on the move.

With FizziQ Web (Amplitude mode of the audio analyzer): record a sound with the computer’s microphone or import an audio file (WAV, MP3, OGG, FLAC, M4A), navigate and zoom within the signal, measure a duration or a period precisely with the cursor, visualize the envelope with the min/max display, then add the analysis to the experiment notebook for the report.

Part 4 - Measuring the frequency of a sound

4.1 Defining frequency

The frequency of a periodic sound is the number of repetitions of its cycle per second, expressed in hertz (Hz). It is the inverse of the period: f = 1/T. The higher the frequency, the higher-pitched the sound is perceived; the lower the frequency, the lower-pitched it sounds.

The human audible range extends, as an order of magnitude, from 20 Hz to 20,000 Hz. This interval is a convenient convention, not a universal boundary: the upper limit decreases with age and varies from person to person, and the sensitivity of the ear is not uniform across the range.

4.2 Fundamental frequency and dominant frequency

These two notions are often confused, including in teaching documents. Yet they are distinct.

The fundamental frequency f₀ of a periodic sound is the inverse of its period: the repetition frequency of the complete cycle. It generally corresponds to the perceived pitch.
The harmonics are the components of the signal whose frequencies are integer multiples of the fundamental: 2f₀, 3f₀, and so on. (The components of an arbitrary sound, whether or not they are multiples of a fundamental, are called partials; bells and percussion produce inharmonic partials.)
The dominant frequency is the component with the largest amplitude in the spectrum at a given instant. It is what a simple frequency meter displays.

The dominant frequency is not necessarily the fundamental. Example: for a male voice singing a vowel around 110 Hz, the fundamental is at 110 Hz, but the harmonics between about 300 and 800 Hz - reinforced by the resonances of the vocal tract - can be stronger than the fundamental itself. An instrument that displays “the frequency” will then show a harmonic, say 330 Hz, whereas the perceived pitch corresponds to 110 Hz.

Watch out for this confusion. The “Dominant frequency” instrument of FizziQ mobile displays the strongest component of the spectrum at the moment of measurement. For a pure tone or a sound whose fundamental dominates (tuning fork, flute, whistling), dominant and fundamental frequencies coincide. For a low voice or an instrument rich in harmonics, they can differ: you then need to examine the full spectrum to identify the fundamental.

4.3 Measurement methods

Four complementary methods can be used to measure a frequency, each with its limits:

Measuring the period on the oscillogram. The most instructive method: it yields the fundamental frequency (the repetition frequency of the cycle), and its precision depends on careful pointing. It requires a periodic, reasonably stable signal.
Automatic detection. The app computes and displays a frequency continuously. Fast and convenient, but the algorithm returns the dominant component, which may differ from the fundamental (see 4.2), and it becomes unstable on noisy sounds or mixtures of sources.
Spectral analysis. You read the frequencies of the various components off the spectrum. This method identifies the fundamental even when it does not dominate, but its precision is limited by the frequency resolution (see Part 5).
Comparison with a generator. Adjust the frequency of a tone produced by the synthesizer until the beats with the sound under study disappear: the generator frequency then equals the studied frequency. A very precise method for stable sounds, provided two devices are available (one produces, the other compares).

4.4 Suggested experiments

Frequency of a tuning fork: check that an A4 tuning fork vibrates at 440 Hz, by period measurement and by direct readout.
Frequency of a sung note: sing a sustained vowel and compare the displayed frequency, the measured period and the spectrum.
Comparing musical notes: measure the frequencies of two notes an octave apart and verify the ratio of 2, or explore the ratios of the equal-tempered scale.
Frequency of a siren: follow the evolving frequency of a modulated sound (the spectrogram is the ideal tool here, see Part 6).
Beats between two close frequencies: produce 440 Hz and 442 Hz with the synthesizer and verify that the beat frequency equals |f₁ - f₂| = 2 Hz.

Part 5 - Understanding the sound spectrum

5.1 What is a spectrum?

The spectrum of a sound represents how the amplitude (or energy) of the signal is distributed across frequency, computed over a given analysis duration. The horizontal axis is frequency in hertz; the vertical axis is the amplitude (or level) of each frequency component.

Beware of the wording “the spectrum shows all the frequencies present in the sound”: a spectrum is always the result of a computation performed on a portion of signal of finite duration, with a given method (discrete Fourier transform), a limited frequency resolution and a threshold below which components merge into the noise. Two different analysis settings produce two different-looking spectra for the same sound.

5.2 Fundamental, harmonics and timbre

The spectrum of a periodic sound of fundamental frequency f₀ consists of lines located at the frequencies f₀, 2f₀, 3f₀, …: the harmonic series. The relative amplitudes of these harmonics vary from one source to another, and this distribution is one of the main determinants of timbre:

a pure tone (tuning fork, synthesizer in sine mode) has a single line, at f₀;
a complex harmonic sound (voice, string, pipe) has several lines at multiples of f₀ - a clarinet favors odd harmonics, a violin shows a rich, extended series;
an inharmonic sound (bell, timpani, metal bar) has partials whose frequencies are not integer multiples of a fundamental;
noise has a continuous spectrum, spread over a wide band, with no distinct lines.

5.3 An introduction to the Fourier transform

The intuitive idea: any signal, however complicated, can be decomposed into a sum of sine waves of different frequencies, amplitudes and phases. The spectrum is the result of this decomposition: it indicates “how much” of each sine wave is needed to rebuild the signal. It is the mathematical counterpart of what the inner ear does, to a first approximation: separating the frequency components of a sound.

Going further. In practice, the software applies a discrete Fourier transform (computed with the FFT algorithm, Fast Fourier Transform) to a portion of the digitized signal - the time window - containing N samples taken at sampling rate f_s. The result is a spectrum defined at the frequencies k · f_s/N (k an integer), i.e. with a frequency step Δ f = f_s/N = 1/T_{win}, where T_{win} is the window duration. The highest analyzable frequency is f_s/2 (Shannon-Nyquist theorem). The choice of window therefore entirely determines the fineness of the analysis.

5.4 Limits of spectral analysis

Frequency resolution. The spectrum step is Δ f = 1/T_{win}: a 0.1 s window gives a resolution of 10 Hz; a 1 s window is needed to separate two components 1 Hz apart. It is therefore impossible to separate 440 Hz and 442 Hz with a 50 ms window.
Duration-resolution trade-off. Lengthening the window sharpens the frequency resolution but averages the signal over a longer time: a sound that changes during the window (attack, glissando) produces a blurred spectrum. This is a fundamental trade-off that no setting can remove.
Spectral leakage and windowing. Cutting a portion of signal abruptly creates artificial discontinuities at the edges, which spread each line’s energy onto neighboring frequencies (“spectral leakage”). Software mitigates this by multiplying the signal by a smooth weighting function (a Hann window, for instance), at the cost of slightly wider lines.
Noise and detection threshold. Weak components merge into the noise floor of the microphone and the environment: their absence from the spectrum does not prove their absence from the sound.
Clipping. A clipped signal generates artificial harmonics: the displayed spectrum no longer describes the original sound.
Aliasing. If the sound contains components above f_s/2, they reappear at false frequencies in the spectrum. Acquisition chains include anti-aliasing filters, but caution remains warranted near the limit.

5.5 Using it on smartphone and computer

Immediate analysis on the smartphone (the Spectrum analyzer instrument of FizziQ mobile, which displays the spectrum in real time up to 5,000 Hz) is enough to identify a fundamental and its first harmonics, to compare a pure tone with a complex sound, or to observe the spectrum of a vowel while singing it.

Detailed analysis on the computer (the Spectrum mode of FizziQ Web) is preferable when you need to examine the lines closely, measure frequencies with the cursor, analyze a precise portion of a recording, or export the data (frequency, magnitude) to the spreadsheet for quantitative processing. Importing an audio file also brings reproducibility: the same file can be re-analyzed by every student, or compared with a recording made in class.

Part 6 - Understanding the spectrogram

6.1 Definition

The spectrogram represents the evolution of a signal’s frequency content over time. It is built by computing the spectrum on successive windows of the signal and juxtaposing the results:

horizontal axis: time;
vertical axis: frequency;
color (or graphic intensity): the amplitude or level of each component, depending on the tool’s convention. In FizziQ Web, black corresponds to silence, blue to low intensity, yellow to medium intensity and red to maximum intensity.

6.2 Oscillogram, spectrum and spectrogram compared

Representation	Horizontal axis	Vertical axis	What it shows
Oscillogram	Time	Amplitude	Time shape of the signal
Spectrum	Frequency	Amplitude or level	Frequency content of one time interval
Spectrogram	Time	Frequency (color = intensity)	Evolution of the spectrum over time

The three representations describe the same signal from three complementary angles: the oscillogram answers “how does the signal vary?”, the spectrum “which frequencies is it made of?”, the spectrogram “how does its composition evolve?“.

6.3 Reading a spectrogram

A few characteristic signatures, easy to reproduce in class:

Pure tone: a single horizontal line, at the frequency of the tone.
Musical note: a stack of regularly spaced horizontal lines - the fundamental and its harmonics.
Glissando or siren: one or more oblique or undulating lines, showing the continuous variation of frequency.
Speech: a rich structure alternating harmonic zones (vowels, with their formants - reinforced frequency bands) and noisy zones (consonants such as “s” or “sh”, diffuse patches at high frequencies).
Beats: a line whose intensity pulses at the beat rate (the two components being too close to be separated by the analysis).
Doppler effect: a line whose frequency drops as the source passes the microphone.
Birdsong: fine, fast, modulated patterns, often between 2 and 8 kHz - an excellent reading exercise.

6.4 Limits of the spectrogram

The spectrogram inherits all the limits of spectral analysis, with one extra constraint: the window choice simultaneously sets the time resolution (short window = brief events well localized) and the frequency resolution (long window = frequencies well separated) - you cannot have both finely at once. Add to this the color scale (the same data looks very different depending on the chosen contrast), the noise threshold (weak components disappear or, conversely, the background fills with patches), and classic misleading interpretations: vertical stripes caused by a brief impact (which excites all frequencies), artificial harmonics caused by clipping, or ghost lines caused by aliasing.

Experiment with FizziQ. In FizziQ Web, load the “Siren” sound from the library, display the spectrogram and follow the frequency line as it rises and falls. Then load the “Risset bells”: the spectrogram reveals the mechanism of the endless-rise illusion, invisible to the ear alone. On the same theme of auditory illusions, see the ready-to-use activity The Shepard sound effect.

Part 7 - Measuring sound level and understanding decibels

7.1 Acoustic pressure

The air around us constantly exerts a pressure of about 101,325 Pa (standard atmospheric pressure). A sound wave superimposes very small variations on this pressure: the instantaneous acoustic pressure p(t), which oscillates around zero.

Since this quantity oscillates, it is characterized by its effective value p (the RMS value, root mean square): the square root of the mean of the square of p(t) over an observation period. It is this effective value that enters the definition of sound level. Orders of magnitude: about 0.00002 Pa at the threshold of hearing, 0.02 Pa for a conversation, 20 Pa near the threshold of pain - a factor of one million between the extremes.

7.2 Sound pressure level

The sound pressure level is defined by:

Lₚ = 20 log₁₀(p/p₀)

where:

Lₚ: sound pressure level, in decibels (dB);
p: effective acoustic pressure of the measured sound, in pascals (Pa);
p₀: reference acoustic pressure in air, fixed by international convention at p₀ = 20 µPa (2 × 10⁻⁵ Pa), a value close to the human hearing threshold at 1,000 Hz.

Thus, a sound of effective pressure p = p₀ has a level of 0 dB; p = 0.02 Pa (conversation) gives Lₚ = 20 log₁₀(0.02/0.00002) = 20 log₁₀(1000) = 60 dB; p = 20 Pa gives 120 dB.

7.3 Why use a logarithmic scale?

Everyday acoustic pressures span six orders of magnitude (from 20 µPa to 20 Pa). A linear scale would be unreadable: conversation, at 0.02 Pa, would be crushed against zero on an axis graduated up to 20 Pa. The logarithmic scale maps this span onto the 0-120 dB interval, closer to the way we perceive intensities.

An essential consequence: on a logarithmic scale, multiplying the physical quantity translates into adding decibels. Multiplying the pressure by 10 adds 20 dB; multiplying the power by 10 adds 10 dB. Decibels therefore never add up like ordinary quantities.

7.4 Level changes

Three benchmarks to know - and not to mix up:

Doubling the acoustic pressure adds 20 log₁₀(2) ≈ 6 dB.
Doubling the intensity or acoustic power adds 10 log₁₀(2) ≈ 3 dB.
Adding a second identical source does not systematically raise the level by exactly 3 dB at the measurement point. The “+3 dB” result assumes the two sources are incoherent (their signals are uncorrelated, like two independent noises): their intensities then add. If the sources are coherent (two loudspeakers playing the same signal), it is the instantaneous pressures that add, and the result depends on phase: depending on the measurement position, the level can rise by up to +6 dB (constructive interference) or drop sharply (destructive interference).

In real situations, the distance to the sources, their directivity, reflections from walls, and the nature of the sound field (direct or reverberant) also play a role: all reasons why a classroom measurement often departs from the theoretical value.

And above all: two 60 dB sounds never make 120 dB. At most 66 dB (coherent, in-phase sources), typically about 63 dB (independent sources).

7.5 Adding independent levels

For independent (incoherent) sources, intensities add, which translates on the levels into:

L_total = 10 log₁₀(10^(L₁/10) + 10^(L₂/10) + …)

Worked examples (worth recomputing in class - FizziQ’s calculator is well suited):

L₁ = L₂ = 60 dB: L_total = 10 log₁₀(2 × 10⁶) ≈ 63.0 dB.
L₁ = 60 dB and L₂ = 70 dB: L_total = 10 log₁₀(10⁶ + 10⁷) ≈ 70.4 dB - the louder source dominates almost completely.
Four 50 dB sources: L_total = 10 log₁₀(4 × 10⁵) = 50 + 10log₁₀(4) ≈ 56 dB.

This formula is only relevant for incoherent sources; it does not apply to two loudspeakers playing the same signal.

7.6 Weightings and averaged measurements

A few complementary notions, useful for interpreting the values found in documents:

dB SPL: sound pressure level relative to the 20 µPa reference, without frequency weighting.
A-weighting and dB(A): sound level meters often apply a filter (A-weighting) that attenuates low and very high frequencies to mimic the ear’s sensitivity; the result is expressed in dB(A). A dB(A) level and an unweighted level are not directly comparable.
Instantaneous measurement and equivalent level: a fluctuating level can be described by its instantaneous value or by an energy average over a given duration (the equivalent level, written Leq in acoustics documents). FizziQ mobile’s “Noise level” instrument provides an average of the ambient level, while the “Sound level meter” tracks the level over time.

Regulatory measurements (occupational noise, neighborhood noise) involve instruments of standardized classes, specific procedures and standards: they are beyond the scope of a smartphone measurement (see Part 8).

Part 8 - Can a smartphone or a computer replace a sound level meter?

The answer up front: a smartphone or a computer can be very useful for teaching, for relative comparisons and for many acoustics experiments. It does not, however, automatically replace a professional sound level meter that is calibrated and compliant with a standard.

8.1 Suitable uses

Relative comparison of two situations: is the corridor noisier than the classroom? By how many decibels does closing the door reduce the level?
Tracking a level over time: rising noise during a break, decay of reverberation.
Teaching experiments on the decibel scale, source addition, attenuation with distance.
Indicative noise mapping of a place (schoolyard, street, cafeteria), provided the same device is used everywhere.
Detecting a change: a threshold being crossed, a source appearing.
Comparing materials or distances: which material attenuates best? How does the level fall when the distance is doubled?

8.2 Limits

Microphones differ from one model to the next (sensitivity, frequency response): two devices give different values for the same sound.
Absolute calibration is uncertain: without a reference, the displayed level can be several decibels off the true value. FizziQ mobile actually offers a sound-level-meter calibration setting (a dB offset) to align the display with a known reference.
The system’s automatic processing (automatic gain, noise reduction) can distort measurements (see 2.3).
Clipping caps the measurement of loud sounds; the range displayed by FizziQ mobile (0 to 110 dB) does not mean every microphone faithfully covers that whole range.
The limited frequency response underestimates very low or very high sounds.
The electronics’ noise floor masks very quiet sounds: very low displayed levels are of little significance.
The case, the hand, the position of the microphone alter the measurement.
The environment (reflections, wind on the mic, position in the room) strongly influences the result.
Operating systems process the audio stream differently: the same model under two OS versions can give different values.

8.3 Precautions

For usable measurements: keep the same device throughout a series of measurements; keep the same microphone orientation toward the source; keep the same distance; check for the absence of clipping; repeat the measurements (for close values, summarize them by their mean or median; for a rigorous energy-average level, first convert the decibels to linear quantities or use an equivalent level Leq); record the device model in the report; explicitly distinguish relative comparison from absolute value; calibrate the app against a reference when necessary and possible.

8.4 Usage limits

A measurement from an uncertified smartphone or computer must not be turned into a regulatory measurement, a medical diagnosis, professional expertise, a conclusion about hearing safety, or legal evidence. For those uses, only sound level meters compliant with the applicable standards, calibrated and operated according to the prescribed procedures, are authoritative.

Part 9 - Producing sounds for experiments

Studying sound also means producing it: having a controlled, reproducible source changes the nature of the experiments you can run.

9.1 Sine tones

A sine-tone generator (the FizziQ synthesizer, available on mobile and on FizziQ Web, covers 20 Hz to 20,000 Hz) gives control over: the frequency of the sine wave, its amplitude (the volume of the generated signal), its phase (the time offset between several channels, adjustable on mobile) and its duration. You thus have “reference tones”: a signal whose frequency is known before you measure it.

9.2 Superimposing sounds

The FizziQ mobile synthesizer can superimpose up to three simultaneous channels. Superimposing signals illustrates several phenomena:

signal addition: at every instant, pressures (and signals) add algebraically;
beats: two close frequencies (440 and 442 Hz) produce an amplitude pulsing at 2 Hz;
chords: frequencies in simple ratios (C-E-G) produce a sound perceived as consonant;
interference: two loudspeakers playing the same signal create, under certain conditions (coherent sources, controlled geometry), zones of reinforcement and attenuation in the room.

One must distinguish the instantaneous addition of pressures (always true, and phase-dependent) from the energy addition of intensities (valid only for incoherent sources) - the same distinction already met in 7.4 for adding levels.

9.3 White noise and other signals

White noise (ideally) is a random signal whose power spectral density is constant per unit frequency; white noise produced digitally or by a real loudspeaker is only white over a limited frequency band. It is a convenient source for studying the addition of independent sources or a system’s response. (The FizziQ sound library provides it.)
A frequency sweep runs continuously through a range of frequencies: useful for studying a loudspeaker’s response, locating a system’s resonances or observing the limits of the measurement chain.
An impulse (a brief clap) excites all frequencies at once: it is the source used to measure an echo or reverberation.
The complex sounds in the FizziQ library (instruments, sirens, Doppler sounds, Risset bells) provide reproducible digital reference signals for analysis.

9.4 Loudspeaker limits

The digital waveform generated by the synthesizer is precisely defined in software, but the sound radiated by the loudspeaker may be filtered or distorted. A smartphone’s tiny loudspeaker has a limited frequency response: it reproduces low frequencies very poorly (typically below 200 to 300 Hz, depending on the model, the emitted level drops sharply). It introduces distortion (harmonics absent from the original signal), all the more so as the volume is high, and it can saturate. Finally, two different smartphones or computers, or the same device at two volume settings, produce different sounds for the same digital signal.

Watch out for this confusion. Three distinct objects: the digital signal produced by the synthesizer (perfectly defined), the sound actually emitted by the loudspeaker (filtered and distorted by it), and the sound measured by the microphone (filtered again by the acquisition chain). When you generate 100 Hz and measure something else, the discrepancy usually comes from the loudspeaker, not from the generator.

9.5 FizziQ tools

Verified sound-production features in the FizziQ ecosystem: the frequency synthesizer (20 Hz to 20,000 Hz; up to three simultaneous channels with frequency, volume and phase settings on mobile; several waveforms on FizziQ Web: sine, square, triangle, sawtooth) and the sound library (tuning forks and notes, instrument sounds, white noise, sirens, sounds for the Doppler effect and the Shepard tone illusion, beats). The library can be played on mobile, and used in FizziQ Web as a direct analysis source. For experiment ideas built on the synthesizer, see Six science experiments with the tone generator.

Part 10 - Smartphone or computer: which tool should you choose?

Neither platform is better in all circumstances: the choice depends on the experiment, the mobility required, the precision sought, the equipment available and how the class is organized. The table below compares the verified features of the two tools.

Need	FizziQ mobile (smartphone/tablet)	FizziQ Web (computer)
Mobile, in-the-field measurement	Yes - its core purpose	Possible, but less suited to measurements on the move
Acquisition and live analysis from the mic	Yes	Yes (computer microphone)
Importing an audio file	No	Yes (WAV, MP3, OGG, FLAC, M4A)
Sound library	Yes (playback)	Yes (playback and direct analysis)
Oscillogram	Yes, real time (Amplitude instrument)	Yes, with zoom, cursor, min/max
Sound level (dB)	Yes (Sound level meter, Noise level)	Yes (Sound level mode)
Frequency	Yes (Dominant frequency, 50-10,000 Hz)	Yes (Frequency mode)
Spectrum	Yes, real time (Spectrum analyzer, 0-5,000 Hz)	Yes, with cursor and export
Spectrogram	No	Yes
Sound production (synthesizer)	Yes (3 channels, phase control)	Yes (several waveforms)
Doppler effect in a real situation	Yes (moving source or mic)	Analysis of a recording or a library sound
Simultaneous measurement in several places	Yes (several devices)	Limited
Large-screen work	More limited on a phone, comfortable on a tablet	Yes
Report writing	Experiment notebook, PDF export	Notebook + final report (PDF, Word), spreadsheet, LaTeX
Chromebook / smartphone bans	-	Yes, runs in the browser
Comparing several files	No	Yes (successive imports, notebook)
Acoustic stopwatch (speed of sound)	Yes	No
Data transfer and export	QR code, CSV, Python (export)	CSV, Python, image export; QR code to send to mobile

In short: does the experiment happen where the phenomenon is (schoolyard, street, music room)? The smartphone is the obvious choice. Does the experiment consist of finely analyzing a signal, comparing files or producing a document? The computer is the obvious choice. Many lessons benefit from combining both - the subject of the next part.

Part 11 - Combined smartphone-and-computer workflows

The FizziQ tools share the same data formats (QR codes, CSV, notebook files), which makes it possible to build lessons where each device does what it does best.

Workflow 1 - Measure on smartphone, analyze on computer

Carry out the experiment with the smartphone (level measurement, frequency, acoustic stopwatch…) and save the observations in the FizziQ notebook.
If the experiment produces a sound to analyze, also record it as an audio file with the phone’s recording app.
Transfer the measurement data by QR code or CSV export to FizziQ Web, and import the audio file into the FizziQ Web audio analyzer.
Examine the signal closely on a large screen: zoom into the oscillogram, spectrum, spectrogram.
Insert tables, graphs and screenshots into the final report and export it as PDF or Word.

Workflow 2 - Prepare on computer, experiment on smartphone

Analyze a reference sound in FizziQ Web (for instance the 440 Hz tuning fork from the library): oscillogram, spectrum.
Identify the relevant quantities (period, frequency, harmonics).
Formulate a hypothesis (for instance: “the tuned guitar string will produce the same fundamental”).
Carry out the experiment in the field or in class with FizziQ mobile.
Compare the experimental result with the reference and conclude.

Workflow 3 - Working without smartphones

Use a computer or a Chromebook (FizziQ Web runs in the browser, with no installation).
Record a sound with the computer’s microphone, or load an audio file, or pick a sound from the library.
Study the oscillogram, the spectrum or the spectrogram.
Export the data to the built-in spreadsheet for quantitative analysis.
Produce the report with the report editor (PDF or Word export).

This workflow directly addresses situations where smartphone use is restricted in the school.

Workflow 4 - Using several smartphones

Use a first device as the source (synthesizer or sound library).
Optionally use a second device as a second source (beats, source addition) or as a trigger.
Use a third device as the measuring instrument (sound level meter, frequency meter, acoustic stopwatch).
Control the distances and orientations of the devices, which directly condition the results.
Compare the results across groups; QR-code sharing makes it easy to gather all measurements on one device.

This is the standard setup for measuring the speed of sound (two devices in acoustic-stopwatch mode a known distance apart) and for studying the addition of two sound sources.

Part 12 - Twelve sound experiments with FizziQ

The experiments are ordered by increasing difficulty. Each sheet links, where available, to a detailed resource on the English site. General precaution: never play loud sounds through headphones, and keep synthesizer volumes moderate.

Experiment 1 - Observing the shape of a pure tone

Scientific question: what does the signal of a pure tone look like? - Level: middle school. - Duration: 20 min. - Equipment: a computer (or a smartphone). - Recommended tool: FizziQ Web. - Instrument: audio analyzer, Amplitude mode, “440 Hz tuning fork” sound from the library. - Protocol summary: load the sound, zoom in until 10 to 20 oscillations are visible, describe the observed shape; compare with a voice recording. - Observed quantity: amplitude versus time. - Representation: oscillogram. - Expected result: a regular sine wave for the tuning fork, a complex pattern for the voice. - Precaution: moderate volume. - Main source of uncertainty: none - no quantitative measurement here; the point is observation. - FizziQ activity: Tuning forks: pitch of A.

Experiment 2 - Measuring the period and frequency of a signal

Question: how do you determine a sound’s frequency from its period? - Level: middle school / lower high school. - Duration: 30 min. - Equipment: computer. - Tool: FizziQ Web. - Instrument: Amplitude mode, measurement cursor. - Protocol: display a periodic sound, measure the duration of 10 periods with the cursor, compute T then f = 1/T; compare with the value shown by the Frequency mode. - Measured quantities: duration, period, frequency. - Representation: oscillogram. - Expected result: for the library tuning fork, f ≈ 440 Hz. - Precaution: zoom in enough to point precisely. - Uncertainty: pointing the ends of the interval (hence the value of measuring several periods). - FizziQ activity: Chromatic scale and frequencies.

Experiment 3 - Measuring the frequency of a tuning fork

Question: does the tuning fork really vibrate at 440 Hz? - Level: middle school / lower high school. - Duration: 20 min. - Equipment: a tuning fork with its resonance box, a smartphone. - Tool: FizziQ mobile. - Instrument: Dominant frequency, then Amplitude. - Protocol: strike the fork, bring the microphone close, note the displayed frequency; take several captures and average them in the notebook. - Quantity: frequency (Hz). - Representation: instantaneous value and measurement table. - Expected result: about 440 Hz. - Precautions: strike the fork on a soft support; avoid background noise. - Main uncertainty: detection stability as the amplitude decays. - Related resource: the ready-to-use acoustics activities in the FizziQ activity catalog. - FizziQ activity: Tuning forks: pitch of A.

Experiment 4 - Comparing the timbres of two instruments

Question: why don’t two instruments playing the same note sound alike? - Level: middle school / lower high school. - Duration: 45 min. - Equipment: two instruments (or the instrument sounds from the FizziQ library), one analysis device. - Tool: FizziQ Web (comfortable spectrum) or FizziQ mobile (Spectrum analyzer). - Protocol: play or load the same note on two instruments; for each, note the frequencies and relative amplitudes of the first spectral components; compare. - Quantities: harmonic frequencies and amplitudes. - Representation: spectrum. - Expected result: same fundamental, different harmonic distributions. - Precaution: play the notes at comparable levels, without clipping. - Uncertainty: frequency resolution and variability of instrumental playing. - FizziQ activity: Timbre and harmonics and Harmonic and inharmonic sounds.

Experiment 5 - Observing the harmonics of a voice

Question: what is the spectrum of a sung vowel made of? - Level: lower high school. - Duration: 30 min. - Equipment: one device. - Tool: FizziQ mobile (real time) or FizziQ Web (record then analyze). - Instrument: Spectrum analyzer / Spectrum mode. - Protocol: sing a sustained vowel (“ah”, then “ee”) at constant pitch; identify the fundamental and the harmonics; observe how the spectrum changes when you change vowel without changing note. - Quantity: component frequencies. - Representation: spectrum (and spectrogram in FizziQ Web to see the formants). - Expected result: lines at multiples of the fundamental; different reinforced zones depending on the vowel. - Precaution: sing 20-30 cm from the microphone. - Uncertainty: instability of the sung pitch. - FizziQ activity: Spectrum of sung vowels and Spectrogram of bird song.

Experiment 6 - Producing and measuring beats

Question: what happens when two close frequencies are superimposed? - Level: upper high school. - Duration: 45 min. - Equipment: one smartphone (2-channel synthesizer) or two devices, plus one measuring device. - Tool: FizziQ synthesizer (mobile or Web) + analysis. - Protocol: generate 440 Hz and 442 Hz simultaneously; listen to the pulsing; record and observe the envelope on the oscillogram; measure the beat period and verify f_{beat} = |f₁ - f₂| = 2 Hz; repeat with 445 Hz. - Quantity: envelope period. - Representation: oscillogram (FizziQ Web’s min/max display makes the envelope very readable). - Expected result: beats at 2 Hz then 5 Hz. - Precaution: equal volumes on both channels. - Uncertainty: accuracy of the frequencies actually emitted by the loudspeaker. - Related resources: the Acoustic beats activity and the article Six science experiments with the tone generator.

Experiment 7 - Comparing a pure tone and white noise

Question: what distinguishes a pure tone from noise? - Level: lower high school. - Duration: 30 min. - Equipment: a computer. - Tool: FizziQ Web. - Protocol: load “440 Hz tuning fork” and then “White noise” from the library; compare the three representations: oscillogram, spectrum, spectrogram. - Quantity: appearance of the signal and the spectrum. - Representations: all three. - Expected result: sine wave / single line / horizontal line for the pure tone; disordered signal / spread-out spectrum / colored sheet for the noise. - Precaution: same display settings for both sounds. - Uncertainty: not applicable (qualitative observation). - FizziQ activity: White noise: spectral content.

Experiment 8 - Studying the resonance of an air column

Question: at which frequencies does an air column resonate? - Level: upper high school. - Duration: 1 h. - Equipment: a tube (test tube, PVC pipe), possibly partly filled with water; one source smartphone; one measuring smartphone. - Tool: FizziQ synthesizer + Sound level meter. - Protocol: place the source near the opening, sweep the frequency slowly, note with the sound level meter the frequencies at which the level peaks; compare with the expected resonance frequencies of a tube closed at one end (fₙ = (2n-1)c/4L, to a first approximation, neglecting the end correction). - Quantities: frequency and sound level. - Representation: level/frequency table, graph in the notebook. - Expected result: maxima close to the computed frequencies. - Precautions: moderate volume; stable geometry. - Uncertainties: effective tube length (end correction neglected), device positions. - FizziQ activity: Air columns and pitch and Helmholtz resonance: a bottle.

Experiment 9 - Measuring the speed of sound

Question: how fast does sound travel in air? - Level: high school. - Duration: 1 h. - Equipment: two smartphones, a tape measure, a brief sound source (clapper). - Tool: FizziQ mobile, acoustic stopwatch on both devices. - Protocol: place the two devices a known distance d apart (at least 10 to 20 m); an operator stands near each device; arm both acoustic stopwatches; a first clap, made near the first device, starts both stopwatches; a second clap, made near the second device, stops them; the difference between the two measured durations then equals 2d/c, so c = 2d/Δt; repeat and average. - Quantities: distance, duration. - Representation: measurement table. - Expected result: about 340 to 345 m/s depending on temperature. - Precautions: quiet environment; adjust the stopwatches’ sensitivity and dead time. - Main uncertainty: the time measurement (a few milliseconds out of a few tens of milliseconds) - hence the value of a large distance. - Related resource: Speed of sound - seven activities with a smartphone. - FizziQ activity: Speed of sound: two stopwatches and Speed of sound: a test tube.

Experiment 10 - Studying the Doppler effect

Question: why does the perceived frequency change when the source moves? - Level: upper high school. - Duration: 1 h. - Equipment: a source smartphone (synthesizer, stable tone around 1,000 Hz), a sturdy string or a moving carrier, an analysis device. - Tool: FizziQ mobile (source) + FizziQ Web (spectrogram analysis of the recording), or the Doppler sounds from the library. - Protocol: move the source past the microphone while recording; display the spectrogram; measure the frequency before and after the pass; deduce the source speed from the Doppler formulas. - Quantities: perceived frequencies on approach and on recession. - Representation: spectrogram. - Expected result: a clear frequency drop at the pass; a speed consistent with the actual motion. - Precautions: attach the source smartphone securely; clear the area. - Uncertainty: reading the frequencies on the spectrogram (resolution). - Related resources: Five Doppler effect experiments with a smartphone, the Doppler effect activity and the sound pendulum activity.

Experiment 11 - Studying the addition of two sound sources

Question: do two identical sources add 3 dB? - Level: upper high school. - Duration: 1 h. - Equipment: two source smartphones (white noise), one measuring device. - Tool: FizziQ sound library (white noise) + Sound level meter. - Protocol: measure the level L₁ with the first source alone, L₂ with the second alone (adjust so that L₁ = L₂), then L_total with both; compare with the theoretical value L₁ + 3 dB; repeat with two sources playing the same sine tone and observe that the result depends on position. - Quantity: sound level (dB). - Representation: comparative table. - Expected result: about +3 dB for independent white noises; a position-dependent result for coherent sine tones. - Precautions: rigorously constant positions and distances; same measuring device; make sure the two white noises are independent (two separate, unsynchronized playbacks, otherwise the incoherence condition is not met). - Uncertainty: level fluctuations, reverberant environment. - Related resource: the sound-addition and interference activity in the FizziQ activity catalog. - FizziQ activity: Adding sound levels in decibels and Destructive sound interference.

Experiment 12 - Mapping the sound level of a place

Question: how is noise distributed across the school? - Level: middle to high school (project). - Duration: 2 h (measurements + analysis). - Equipment: one or more smartphones, a map of the place. - Tool: FizziQ mobile, Noise level instrument (average). - Protocol: define a grid of measurement points; at each point, measure the average level for a fixed duration (for example 30 s), with the same device, at the same height and orientation; plot the values on the map; comment. - Quantity: average sound level (dB). - Representation: table and annotated map. - Expected result: marked contrasts between quiet and noisy zones. - Precautions: measure at the same time of day; present the results as indicative and relative, not as regulatory values. - Uncertainty: temporal variability of the noise (repeat the measurements). - FizziQ activity: Sound level and distance.

Further ideas extend this list: speech analysis with the spectrogram, echo and reverberation with an impulse, comparing sound absorption of materials, comparing a real recording with a reference file from the library, the Shepard illusion. The corresponding full activities are available in the FizziQ activity catalog and in the articles on the site.

Part 13 - Common mistakes and confusions

Each entry follows the same pattern: the erroneous statement, why it is wrong or incomplete, the correct formulation.

“Sound and signal are the same thing.” Wrong: the sound is a pressure wave in the air; the signal is its electrical, then digital translation, filtered by the microphone and the acquisition chain. Correct formulation: the digitized signal is a representation - an imperfect one - of the sound wave.

“The displayed amplitude is the sound level.” Incomplete: the oscillogram’s amplitude is a relative quantity with no direct physical unit; the sound level in dB is computed from the effective value of the signal and a calibration. Correct: the amplitude describes the digital signal; the sound level in decibels is a calibrated estimate of the effective acoustic pressure.

“Sound level and sound intensity are the same.” Wrong: sound intensity is a power per unit area in W/m²; sound level is its logarithmic expression in decibels. Correct: the level (in dB) is the logarithm of the ratio between the intensity (or the square of the pressure) and a reference value.

“The higher-pitched a sound, the louder it is.” Wrong: frequency and amplitude are independent. Correct: frequency determines pitch (low/high), amplitude determines level (quiet/loud).

“The frequency displayed by the app is the fundamental frequency.” Not always: the app displays the dominant frequency, the strongest component, which may be a harmonic. Correct: to identify the fundamental of a complex sound, examine the spectrum and look for the regular spacing of the harmonics.

“Frequency is the perceived pitch.” An approximation: pitch is a sensation, generally linked to the fundamental, but phenomena such as the missing fundamental show that perception is not a mere frequency readout. Correct: frequency is a physical quantity; pitch is the sensation most often associated with it.

“Spectrum and spectrogram are the same thing.” Wrong: the spectrum describes a single time interval (frequency/amplitude axes); the spectrogram shows how the spectrum evolves over time (time/frequency axes, intensity as color). Correct: the spectrogram is a succession of spectra.

“The spectrum shows the shape of the wave.” Wrong: the oscillogram shows the time shape; the spectrum shows the frequency composition. Correct: oscillogram = time, spectrum = frequency.

“Any frequency present in a sound is a harmonic.” Wrong: a harmonic is an integer multiple of the fundamental; other components are partials (possibly inharmonic) or noise. Correct: harmonic = component at n × f₀, with n an integer.

“White noise is any random noise.” Wrong: white noise has a precise definition - energy uniformly distributed over frequency. Other random noises (pink noise, traffic noise) have different distributions. Correct: white noise is random noise with a uniform spectral density.

“dB and dB(A) are equivalent.” Wrong: dB(A) applies a frequency weighting that attenuates lows and extreme highs; the two values differ, especially for bass-rich sounds. Correct: only compare levels expressed with the same weighting.

“Doubling the pressure or doubling the power - same thing: +3 dB.” Wrong: doubling the pressure adds about 6 dB (the level depends on the square of the pressure); doubling the power adds about 3 dB. Correct: +6 dB per pressure doubling, +3 dB per power doubling.

“To add two sounds, you add their decibels.” Wrong: levels are logarithms; you add intensities (incoherent sources) or instantaneous pressures (coherent sources), never dB. Correct: 60 dB + 60 dB ≈ 63 dB for independent sources.

“My smartphone is a sound level meter.” A misuse of language: an app displays an estimate of the level, without the metrological guarantees (accuracy class, calibration, standardized weightings) of a standards-compliant sound level meter. Correct: a smartphone allows indicative, comparative measurements; a certified sound level meter allows standardized measurements.

“My 62 dB reading is the true value.” Incomplete: without calibration, the absolute value is uncertain by several dB. Correct: prefer differences between measurements made under the same conditions with the same device.

“The 100 Hz tone I generated is the sound emitted in the room.” Wrong: the loudspeaker filters and distorts the signal, especially in the bass. Correct: the generated digital signal and the sound actually radiated differ; only the microphone measurement tells you about the sound actually produced.

Part 14 - Frequently asked questions

How do you measure sound with a smartphone? The smartphone’s microphone turns sound into a digital signal, which an app such as FizziQ analyzes: sound level in decibels, frequency, signal shape, spectrum. Just install the app, allow microphone access and pick the instrument you need. The measurements are reliable for comparisons; the absolute decibel value remains indicative (see Part 8).

How do you analyze a sound on a computer? With FizziQ Web, in an ordinary browser, you can record a sound with the computer’s microphone, import an audio file (WAV, MP3, OGG, FLAC, M4A) or pick a sound from the built-in library, then examine it as an oscillogram, a spectrum or a spectrogram. The cursor and zoom allow precise measurements of durations and frequencies (see Parts 3 to 6).

How do you measure decibels with a phone? Open a sound-level-meter instrument (in FizziQ: Microphone category, Sound level meter), point the microphone at the source without covering it, and read the level in dB. For usable results, keep the same device, distance and orientation from one measurement to the next (see Parts 7 and 8).

Is a sound level meter app reliable? It is reliable for comparing situations and tracking changes, less so for giving an absolute value: the microphones are not calibrated and differ between devices, with discrepancies that can reach several decibels. A regulatory measurement requires a certified sound level meter (see Part 8).

How do you measure the frequency of a sound? Three methods: read the automatically detected frequency, measure the period on the oscillogram and compute f = 1/T, or locate the lines on the spectrum. Period measurement is the most instructive; the spectrum is indispensable for complex sounds (see Part 4).

What is the difference between fundamental frequency and dominant frequency? The fundamental is the repetition frequency of the signal’s cycle, the one that sets the perceived pitch; the dominant frequency is the strongest component of the spectrum at a given moment. They coincide for a pure tone but can differ for a voice or a harmonic-rich instrument (see 4.2).

How do you see the spectrum of a sound? On a smartphone, FizziQ’s Spectrum analyzer instrument displays the spectrum in real time; on a computer, FizziQ Web’s Spectrum mode displays it for a recording or a file, with a measurement cursor and data export (see Part 5).

What is the difference between a spectrum and a spectrogram? The spectrum is a frequency snapshot of one time interval: amplitude versus frequency. The spectrogram is a film: it shows how that spectrum evolves over time, with color encoding intensity (see Part 6).

What is an oscillogram for? For observing the signal’s shape over time: recognizing a pure tone, measuring a period, seeing an attack, beats, an echo or clipping. It is the starting representation of any analysis (see Part 3).

Why do two identical sources add about 3 dB? Because for independent sources, it is the intensities that add: twice the intensity corresponds to 10log₁₀(2) ≈ 3 dB more. This result assumes incoherent sources; two loudspeakers playing the same signal can give anything from +6 dB to strong attenuation depending on position (see 7.4 and 7.5).

Why does the result change from one smartphone to another? Each model has its own microphone (sensitivity, frequency response, position) and its own software processing (automatic gain, noise reduction), and factory calibration is not guaranteed. Differences between devices are normal; they are a reminder to always compare measurements made with the same device (see 2.2, 2.3 and 8.2).

Can you use a computer’s microphone for an experiment? Yes: FizziQ Web can record directly with the computer’s microphone and then analyze the signal. Built-in microphone quality varies; an external microphone improves acquisition if you need finesse (see 2.2 and Part 3).

Can you import an audio file to analyze it? Yes, into FizziQ Web, in WAV, MP3, OGG, FLAC and M4A formats. It is the most reproducible way to work: the same file can be analyzed by the whole class and compared with a real recording (see 5.5).

How do you measure the speed of sound? The most direct method uses two smartphones in acoustic-stopwatch mode, a known distance apart: two claps a known distance apart (the first starts both stopwatches, the second stops them) give a duration difference equal to 2d/c, from which the speed follows. Expect about 340 to 345 m/s depending on temperature (see Experiment 9).

Which app should you use to study acoustics? FizziQ (free, ad-free, no account needed) covers the whole chain: microphone measurement, sound production, a library of reference sounds, analysis (oscillogram, spectrum, and spectrogram on FizziQ Web), an experiment notebook and report export.

Should you use a smartphone or a computer? The smartphone to capture and measure in the field, on the move, in real time; the computer to analyze finely, import files, compare and write. The two combine within a single lesson thanks to QR-code and CSV exchange (see Parts 10 and 11).

Is FizziQ Connect needed for sound experiments? No. The acoustics experiments in this guide rely on the microphones of smartphones, tablets and computers. FizziQ Connect is a box designed to connect other kinds of external sensors (temperature, CO₂, light, voltage…) and is mainly used in environmental or multidisciplinary experiments.

Conclusion

With a smartphone or a computer, the essential quantities of acoustics become directly accessible: period and frequency on the oscillogram, harmonic composition on the spectrum, time-frequency evolution on the spectrogram, sound level in decibels - each with limits you need to know in order to interpret the results correctly.

The smartphone and the computer are not rivals: the first captures and measures as close as possible to the phenomenon, the second analyzes, compares and documents. The educational measurement they allow - comparative, reproducible, open to discussion - is not a regulatory measurement, and that is precisely what makes it an excellent training ground for metrology: uncertainties, calibration, conditions of validity.

The best starting point is a simple experiment: display the oscillogram of a tuning fork and measure its period. From there, everything else unfolds - harmonics, beats, Doppler, the speed of sound. The FizziQ ecosystem supports every step, from capturing the sound to interpreting it and reporting on it: the mobile app for measuring and producing sounds, FizziQ Web for analyzing and writing, and the activity catalog and site articles to extend each experiment.

To go further on fizziq.org: the acoustics activities in the activity catalog, the sound and acoustics tools page, the articles on the Doppler effect, the speed of sound, the tone generator, smartphone sensors and FizziQ Web on the big screen, as well as the FizziQ and FizziQ Web user guides (English versions available).

Table of contents