PRAIRIE FIELD NOTES / MONTANA

What an FFT Spectrum Is and How It Relates to a Spectrogram

Learn what an FFT spectrum shows about sound, how amplitude versus frequency is plotted, and how a spectrogram stacks many spectrum slices over time.

7 min read5 viewsWhat an FFT Spectrum Is and How It Relates to a Spectrogram
— · 0 votes
Contents
  1. The idea behind an FFT spectrum
  2. Reading amplitude versus frequency
  3. What the window length changes
  4. How a spectrogram is built from slices
  5. Setting the spectrogram parameters
  6. Choosing between the two views
  7. Terms you will meet in the display
  8. Questions about FFT spectra and spectrograms
  9. What is an FFT spectrum?
  10. Is a spectrogram just many FFT spectra in a row?
  11. Why does a longer block give sharper frequency lines?
  12. What is the difference between a magnitude spectrum and a power spectrum?
  13. Can I read exact frequencies from a spectrogram?

An FFT spectrum is a picture of one moment of sound, split into its frequencies by a Fast Fourier Transform. It shows amplitude versus frequency for a short slice of audio. A spectrogram takes that same idea and stacks many spectrum slices side by side across time.

  • An FFT spectrum answers which frequencies are present in a short slice of sound and how strong each one is.
  • The Fast Fourier Transform is the calculation that turns a block of samples into that frequency breakdown.
  • A magnitude spectrum or power spectrum is simply the spectrum viewed as strength rather than raw signal.
  • A spectrogram is a row of spectrum slices placed along a time axis, so frequency, time and strength appear together.
  • The length of each slice sets a trade-off between frequency detail and time detail in the spectrogram.

The idea behind an FFT spectrum

An FFT spectrum describes the frequency content of a short block of sound at one instant. The Fast Fourier Transform is a calculation that takes a set of samples and reports which frequencies are inside them. Each result is a value for one frequency band, and the set of values forms the spectrum.

The transform works on a block of samples rather than on a whole recording. That block is often called a window, and its length decides how much of the signal is examined at once. A longer block separates nearby frequencies more finely, while a shorter block follows rapid changes more closely.

The output is the raw complex spectrum, which holds both strength and phase for each frequency. Phase describes where each frequency sits in its cycle at the start of the block. Many displays drop the phase and keep only strength, which is what a viewer reads as the spectrum.

Reading amplitude versus frequency

An FFT spectrum is read as a graph of amplitude versus frequency, with frequency along the horizontal axis. The height at each frequency shows how strong that component is in the slice. A tall peak marks a frequency that carries a lot of energy in that moment.

The same data can be drawn as a magnitude spectrum, which plots the absolute strength of each frequency. A power spectrum plots strength squared, which compresses tall peaks and lifts weak ones. Both describe the same slice, and the choice depends on what you want to see.

Real sound rarely produces single clean lines, so a spectrum shows peaks, slopes and a floor of low-level content. A pure tone gives a narrow peak, while noise spreads across many frequencies at once. Recognizing these shapes helps you tell a whistle from wind or a machine hum.

What the window length changes

The length of the analysis block sets the balance between frequency resolution and time resolution in an FFT spectrum. A long block resolves close frequencies but smears events that happen quickly. A short block pinpoints timing but blurs nearby frequencies together.

This limit is a property of the transform, not a flaw in the software. You cannot have unlimited detail in both frequency and time from the same slice. Choosing a block length means deciding which kind of detail matters for the sound you study.

Bird vocalizations illustrate the choice. A slow, steady whistle has stable frequencies that a longer block separates well. A quick click or trill changes too fast for a long block, so a shorter block keeps its timing visible at the cost of frequency precision.

How a spectrogram is built from slices

A spectrogram is made by computing an FFT spectrum for successive slices of a recording and placing each result in a column. Time runs along one axis and frequency along the other. Color or brightness shows the strength at each time and frequency point.

Each column of a spectrogram is therefore a spectrum slice, and the whole image is a sequence of them. The slices overlap in many tools, which smooths the picture and reduces gaps between columns. A slice taken from a spectrogram shows the spectrum at that instant.

The spectrogram adds the time dimension that a single spectrum lacks. One FFT spectrum tells you what frequencies exist in a moment, and a spectrogram tells you how those frequencies rise and fall. The two views use the same calculation and differ only in how the results are arranged.

Setting the spectrogram parameters

Spectrogram settings follow directly from the spectrum settings, because each column is an FFT spectrum. The block length and window shape chosen for the transform decide the detail in both directions. Changing them changes the picture you see.

A longer block sharpens frequency lines and lets you measure pitch more precisely. A shorter block makes short events stand out and keeps their timing readable. The trade-off is the same one described for the spectrum, applied across the whole display.

Overlap between blocks affects smoothness rather than true resolution. More overlap gives more columns and a finer time grid, but it cannot recover detail the block length has already removed. Tuning the settings means matching the analysis to the sound, not maximizing every control.

Choosing between the two views

An FFT spectrum suits questions about frequency content at a chosen moment. It gives a clean read of which frequencies are present and how strong they are. A spectrogram suits questions about how sound changes over time, such as a call that sweeps or repeats.

Use the spectrum when you want to measure a peak or compare the strength of two frequencies. Use the spectrogram when you want to see the shape of an event, its start, its end and its pattern. Field guides to bird sound lean on the time-frequency view for this reason.

The two views work well together. You can spot an event in the spectrogram and then read its spectrum slice to get exact frequency values. Moving between the image and the single slice connects the pattern you see with the numbers behind it.

Terms you will meet in the display

Several terms appear alongside an FFT spectrum and its spectrogram, and they describe the same underlying data. Frequency spectrum is the general name for the breakdown of a signal by frequency. Magnitude spectrum and power spectrum name two ways of expressing strength.

A spectrum slice is one column's worth of that data, taken at a chosen time. Amplitude versus frequency is the way the graph is laid out, and it applies to both the spectrum and each spectrogram column. Keeping the terms straight helps when you compare tools.

The spectrogram combines frequency, time and strength in one image, so it needs all three ideas at once. A reader who understands the spectrum can read any column of the spectrogram. The reverse also holds, since each column is a spectrum.

Questions about FFT spectra and spectrograms

What is an FFT spectrum?

An FFT spectrum is the frequency breakdown of a short block of sound, produced by a Fast Fourier Transform. It lists the frequencies present in that block and the strength of each one. The result is often drawn as amplitude versus frequency.

Is a spectrogram just many FFT spectra in a row?

Yes, a spectrogram is built by computing an FFT spectrum for each successive slice of a recording and arranging the results as columns. Time runs along one axis and frequency along the other, with color or brightness showing strength. Each column is a spectrum slice at that instant.

Why does a longer block give sharper frequency lines?

A longer block gives the transform more cycles of each frequency to compare, so nearby frequencies separate more clearly. The cost is that events shorter than the block get smeared in time. The same trade-off applies to every column of a spectrogram.

What is the difference between a magnitude spectrum and a power spectrum?

A magnitude spectrum plots the absolute strength of each frequency, while a power spectrum plots strength squared. The power view compresses tall peaks and lifts weak content, which can make quiet detail easier to see. Both show the same slice of sound.

Can I read exact frequencies from a spectrogram?

You can read approximate frequencies from a spectrogram and then check a spectrum slice for a sharper value. The precision depends on the block length and the display scale. For careful measurement, read the spectrum rather than the image alone.

Latest articles