Voice frequency in Hz: understanding and measuring your voice

When you're told that your voice is around 120 Hz, what does that actually mean?

This number mainly refers to one acoustic characteristic of your voice: its fundamental frequency, or F0. It provides a way to measure the pitch of your voice and track how it changes over time.

But a number in hertz does not tell the whole story. It changes while you speak, depends on how the measurement is taken, and cannot by itself determine whether a voice is “normal,” masculine, feminine, or healthy.

Here's how to interpret it properly.

Fundamental frequency, voice pitch, and Hz: what do they mean?

One hertz (Hz) equals one cycle per second.

A voiced sound is produced when the vocal folds vibrate. When you produce a voiced sound, such as a vowel, your vocal folds vibrate repeatedly. The rate of this vibration is called the fundamental frequency, or F0.

An F0 of 120 Hz therefore means that this vibration cycle repeats about 120 times per second.

As a general rule: the lower the F0, the lower the voice is perceived to be; the higher the F0, the higher the voice is perceived to be.

This is what connects fundamental frequency to voice pitch.

The two concepts are not exactly the same, however. F0 is an acoustic quantity that can be estimated from a sound signal. It is what a machine measures or estimates. Pitch is a perception produced by our auditory system. It is what we hear.

In ordinary speech, the two are closely related, but two voices with nearly the same F0 can still sound very different.

How is voice frequency measured?

A pitch detector does not analyze your entire voice all at once.

It breaks the sound into very small segments, looks for a repeating pattern, and estimates its rate of repetition, or frequency. This process is repeated several times per second to build a curve showing how your voice changes while you speak.

Not everything can be measured reliably. Some consonants, breathy sounds, and background noise do not have enough periodicity to produce a reliable F0 estimate. The detector therefore has to distinguish usable measurements from moments when the analysis is less reliable.

It also has to deal with octave errors: a detector may sometimes interpret a frequency as being twice as high or twice as low as it actually is. It therefore needs to try to identify and reject these errors.

How Basaltone performs this measurement

In Basaltone, the signal is analyzed about 30 times per second using windows of 2,048 samples, or about 43 milliseconds with a microphone operating at 48 kHz.

The app rejects segments with insufficient periodicity, as well as values outside the 60 to 600 Hz range. A sudden change greater than 5 semitones is also treated as suspicious until the following measurements confirm it.

The retained values are then smoothed so that a single unstable reading does not cause the curve to jump suddenly.

These parameters are Basaltone design choices. They are not clinical thresholds or universal standards for voice measurement.

Basaltone also asks the browser, when supported, not to apply certain processing commonly used for audio calls, such as echo cancellation, noise reduction, or automatic gain control, so that the signal is modified as little as possible.

What is the average voice frequency in Hz?

According to a large population study published in the Journal of Voice, the average F0 in conversational speech is about 112 Hz in men and 169 Hz in women. Berg and colleagues measured 2,472 adults aged 40 to 79 in Germany, including 1,154 men and 1,318 women.

These numbers are useful, but they are not targets.

They describe the averages of two specific groups, in a specific population and under specific measurement conditions. The study itself shows that frequency varies with speaking intensity, among other factors, and that some results also vary with age.

The authors also found that the average observed in women was about 6 to 7 semitones lower than some values previously used in the literature.

In other words, there is no single number that defines the “correct” frequency for a voice. Being above or below a population average is also not enough to conclude that there is something wrong with a voice.

This is where many pages go wrong: they quote a range without a source and present it as a target to reach, the supposedly “normal” voice.

Your first measurement is therefore a particularly useful reference for tracking your own progress, because it allows you to compare later measurements with your own starting point.

How can you get truly comparable voice measurements?

A single measurement does not tell you very much: it shows where your voice is at a given moment. To track changes over time, it is much more useful to compare measurements taken under similar conditions.

This is also why the lowest note you can produce for a fraction of a second does not necessarily represent your usual speaking frequency.

Basaltone only counts a session toward your progress after at least 60 seconds have been measured. This threshold is a design choice intended to prevent a very short segment or a randomly chosen sentence from having too much influence on your progress tracking.

Why not look only at the minimum and maximum?

An extreme value can appear for a few milliseconds because of a real variation in the voice or an imperfect measurement.

That is why Basaltone does not define the low and high ends of a session using the absolute minimum and maximum. Instead, the app uses the 10th and 90th percentiles: in simple terms, this reduces the influence of the most extreme values and provides a range that is more representative of the session.

Likewise, progress is not determined by a single session. Basaltone uses the median of your last 5 measurements of at least 60 seconds, and moves by no more than one step per session.

Why Basaltone also uses semitones

Comparing two values in hertz can sometimes be misleading. A difference of 10 Hz does not have the same significance at every starting frequency.

For example, moving from 100 to 110 Hz represents a proportionally larger change than moving from 200 to 210 Hz, even though the difference is 10 Hz in both cases.

This is where semitones become useful. Instead of looking only at the number of hertz between two frequencies, semitones make it possible to compare their ratio on a logarithmic scale. One octave, for example, always corresponds to 12 semitones and a doubling of frequency.

Basaltone therefore uses semitones internally to compare differences between frequencies more meaningfully, and divides your progress into steps of about 0.75 semitone.

In practice, the same step does not always correspond to the same number of hertz:

A difference of 6 Hz therefore represents a larger change around 100 Hz than around 165 or 200 Hz.

This is also why a small difference in hertz between two sessions is not enough on its own to show lasting progress. To track changes, it is more useful to look at the trend across several comparable measurements.

Voice training duration and progress explains in more detail how to interpret these changes over several weeks.

Hz does not describe your whole voice

Fundamental frequency is only one dimension of the voice. It does not describe resonance, formants, timbre, articulation, or intonation on its own. Two people with nearly the same F0 can therefore sound very different. The guide pitch, timbre, resonance, and formants explains these differences in more detail.

A number in hertz also cannot determine your identity, your health, or how another person will perceive your voice. Think of it simply as a reference point for tracking your progress. If a number pushes you to force your voice, it has stopped being useful: see vocal fatigue, tension and safe practice.

To understand how this measurement can fit into a broader practice, see training for a lower voice.

You can also measure your voice frequency with Basaltone's free test. The sound is analyzed locally in your browser, and no audio recording is stored.