Spectrum analyzer basics · Guide

What is a spectrum analyzer?

A spectrum analyzer is a tool that shows which frequencies are inside a signal and how strong each one is. Instead of a wiggly line over time, it draws a graph with frequency from left to right and level from bottom to top, so a hum, a whistle or a radio station appears as its own peak.

Try it

Press "Demo sound" to see a chord split into its notes, or use your microphone and whistle, hum or speak. Nothing is recorded or sent anywhere.

Press a button to start

Strongest —Note —Level —
Open the full analyzer

What a spectrum analyzer shows

Most signals are a mix of many frequencies at once. A guitar note, a voice, the noise from a fan and the signal from a Wi-Fi router are all made of several tones stacked on top of each other. On an oscilloscope, which shows the signal over time, that mix looks like one tangled wave. A spectrum analyzer pulls the mix apart and shows each part on its own.

The horizontal axis is frequency, measured in hertz (Hz): cycles per second. Low sounds such as a bass drum sit on the left, around 50 to 100 Hz; high sounds such as cymbals sit on the right, around 8,000 to 16,000 Hz. The vertical axis is level, measured in decibels (dB), where each 6 dB step roughly doubles or halves the strength of the signal.

220 Hz440 Hz660 Hz time →01 kHz Wave over timeSpectrum
The same sound twice. Over time it is one tangled wave; as a spectrum it is three tones at 220, 440 and 660 Hz, and you can read how strong each one is.

Two details make the graph easier to read. First, audio analyzers usually space the frequency axis on a log scale, so 100 to 200 Hz takes the same width as 1,000 to 2,000 Hz. That matches how we hear, because each doubling of frequency sounds like the same musical step, an octave. Second, the graph updates many times a second, so you see the spectrum change as the sound changes: a sung note slides, a drum hit flashes across the bass, and a steady hum stays put.

A tall, narrow peak is a single steady tone, such as a whistle or mains hum at 60 Hz. A row of evenly spaced peaks is one note with its harmonics. A wide, uneven carpet across the whole graph is noise. The lowest part of that carpet, where nothing seems to happen, is called the noise floor.

What it is used for

Anyone who works with signals needs to know what frequencies are present, so spectrum analyzers turn up in many places, from a home studio to a mobile phone factory.

Audio and music

Sound engineers use an analyzer to tune a PA system or a home theatre, to find a room resonance that makes one bass note boom, and to check that a mix is not too heavy or too bright. Musicians use one to see the notes and overtones of an instrument, and podcasters use one to find a 60 Hz hum or a high whine before it spoils a recording. In this case the input is a microphone or an audio file, and the range is about 20 Hz to 20 kHz, the range of human hearing.

Radio and wireless

Radio engineers and hobbyists look at frequencies millions or billions of times higher: FM radio near 100 MHz, mobile phones from 700 MHz to several GHz, and Wi-Fi at 2.4 and 5 GHz. An analyzer shows which channels are busy, how strong a transmitter is, and whether it leaks power into the channels next to it. Ham radio operators use one to check that their signal is clean before they go on air.

Electronics and EMI

Every switching power supply, motor and processor gives off some unwanted radio energy, called electromagnetic interference (EMI). Before a product can be sold, it must stay under legal limits, and engineers use a spectrum analyzer with a small probe to find the circuit that is causing a problem. The same tool finds the noise that makes a cheap USB charger buzz in your speakers.

Science and teaching

Scientists use spectrum analysis to study vibrations in bridges and machines, bird song, earthquakes and the light from stars. In schools it is one of the clearest ways to show that a sound is made of separate frequencies, and why a flute and a violin sound different while playing the same note. A free browser analyzer makes that lesson possible on every student's phone.

How it works

All spectrum analyzers answer the same question, "how much of each frequency is here?", but they get there in different ways. The three main designs trade speed, range and price.

Swept-tuned (superheterodyne)

The classic bench analyzer works like a radio that keeps turning its own tuning dial. It mixes the input with a local oscillator, passes the result through one narrow filter, and measures how much signal gets through. Then it moves to the next frequency, and the next, sweeping from the start to the end of the range many times a second. The width of that filter is the resolution bandwidth (RBW): a narrow filter shows fine detail and a low noise floor but makes the sweep slow.

Older analyzers drew the result on a glowing tube screen; modern ones store each sweep in memory, so they can average several sweeps, hold the highest level seen, and place markers on peaks to read their exact frequency and level.

Swept analyzers can cover a huge range, from a few kilohertz to tens of gigahertz, with very low noise. Their weak point is time: while the filter looks at one frequency, it is blind to all the others, so a short burst somewhere else can be missed.

FFT analyzers

An FFT analyzer does not sweep. It records a short piece of the signal as numbers (samples) and runs a maths method called the Fast Fourier Transform, which works out the level of every frequency in that piece in one go. With 8,192 samples at 48,000 samples per second, for example, it measures 4,096 frequencies about 5.9 Hz apart from just 0.17 seconds of sound.

This is how almost every audio analyzer works, including the one on this site: your browser takes the samples from the microphone and runs the FFT dozens of times a second. You can read more about the FFT in its own guide. FFT analyzers are fast and cheap, but the sampling rate limits how high they can go, so on their own they suit audio and lower radio frequencies.

Real-time analyzers

A real-time spectrum analyzer is an FFT analyzer fast enough to process every sample with no gaps between the pieces. Because nothing is skipped, it catches events that last only microseconds, such as a Bluetooth hop or a glitch in a radar pulse. Many show the result as a colour "density" picture, where common signals glow brightly and rare ones show as faint traces.

In audio, "RTA" usually means something slightly different: a real-time analyzer that groups the spectrum into fixed bands, often 31 bands of 1/3 octave, for tuning rooms and PA systems.

Audio vs RF spectrum analyzers

The idea is the same, but the hardware is not. An audio analyzer covers about 20 Hz to 20 kHz (some go to 100 kHz), takes its input from a microphone or a line cable, and shows levels in dBFS or dB SPL. Your phone, your laptop and this website can all do it, because a sound card already turns sound into samples.

An RF analyzer covers radio frequencies, from kilohertz up to gigahertz, and takes its input from an antenna or a cable with a coaxial plug. Levels are shown in dBm, which is power compared with one milliwatt. A microphone cannot hear radio, so for RF you need hardware: a handheld analyzer, a bench instrument, or a USB radio receiver such as an RTL-SDR that turns a computer into a basic RF analyzer.

Key settings in one minute

Three settings matter most on any analyzer. Span is the width of the frequency range on screen, for example 20 Hz to 1 kHz to study hum, or 88 to 108 MHz to see the FM band. RBW (resolution bandwidth, or the FFT size on an audio analyzer) sets how close two tones can be and still show as two peaks. Reference level is the level at the top of the screen; set it just above the strongest signal so nothing is cut off. Each one is covered in detail in spectrum analyzer settings, and the next step after this page is how to read one.

Do you need to buy one?

For sound, probably not. A free online spectrum analyzer running on a phone or laptop is enough to find hum, check a speaker, see the notes in a recording or learn the basics. A USB measurement microphone makes room and speaker readings much more accurate, and costs far less than any instrument.

The limits of a phone are worth knowing. Its microphone is tuned for voices, so it reads low below about 50 Hz and may add a small lift around 2 to 5 kHz, and it cannot give a true sound pressure level without a calibrated mic. For finding a problem, comparing before and after, or learning how sound works, none of that matters much.

For radio, you need hardware, but it can be cheap. A USB receiver dongle is fine for learning and for looking at broadcast signals. A handheld analyzer is the next step for checking antennas, filters and transmitters, and a bench analyzer is for professional design and compliance work. Our guide to which one to buy compares them, and if you are deciding between tools, spectrum analyzer vs oscilloscope explains when each one is the right choice.

This page is part of spectrum analyzer basics.

FAQ

What is a spectrum analyzer used for?

To see which frequencies are in a signal and how strong they are: tuning sound systems, finding interference, testing radios and checking electronics.

How does a spectrum analyzer work?

Swept models tune a narrow filter across the range. FFT models sample the signal and work out all frequencies at once with maths.

What is a real-time spectrum analyzer?

One that analyses every sample without gaps, so it catches short bursts that a swept analyzer can miss.

What are the main types?

Swept-tuned, FFT or vector signal analyzers, and real-time analyzers. There are also audio analyzers, and optical spectrum analyzers for light.

Can a phone be a spectrum analyzer?

For sound, yes, through the microphone. For radio, it needs extra hardware such as an RTL-SDR dongle.

Written and tested by the SpectrumAnalyzer.app team · Last checked

How we tested: The live demo uses the same engine as the main analyzer, which reads the site's 1 kHz test tone at 1000.0 Hz and −6.0 dBFS; the demo chord shows its four notes A3, C#4, E4 and A4.