How the Deep Voice Test Works
A deep voice test measures your speaking pitch in hertz and places it on a scale of depth bands, then adds the resonance reading that explains why two voices at the same pitch can sound different. Pitch is the fundamental frequency of your voice: how many times per second your vocal folds open and close. The lower that number, the deeper the voice sounds, and this is the part of depth that can be measured directly.
Example result · sample recording, not your voice
The tool records eight seconds, cuts the audio into 40-millisecond frames, discards silence, and runs each remaining frame through the McLeod pitch method to find its fundamental frequency between 60 and 500 Hz. Frames with no clear repeating period, which is what breath, noise and most vocal fry look like to the algorithm, are gated out rather than counted. The headline figure is the median of the voiced frames, so a couple of dropped or creaky frames cannot push the reading lower than your voice actually is.
The depth band is then read straight off the median. The bands are defined by this site and published below in full, so you can see where every boundary falls before you read the label. Everything runs inside your browser; the audio is never sent anywhere.
What Counts as a Deep Voice?
Adult masculine speaking voices typically centre around 85 to 155 Hz. This tool divides that range into thirds and labels each third, then adds a band on either side. The thresholds are this site’s own, set for transparency without any claim to clinical standing, and they are the same thresholds the shipped code uses.
A reading near a boundary means what it says: the voice sits near a boundary. Ten hertz either side of 108 Hz is the same voice on a different morning, and the label should be read with that in mind.
Depth Is Pitch Plus Resonance
Two people can both speak at 115 Hz and one will be heard as deeper. The difference is resonance. Above the vocal folds sits the vocal tract, the throat and mouth, and it amplifies some frequencies more than others. The peaks it produces are called formants, and the first one, F1, sits low enough to colour the whole tone. A longer or more relaxed vocal tract tends to produce lower formants, which listeners describe as a darker, fuller, more chesty sound. A shorter or tighter tract produces higher formants and a thinner sound at the same pitch.
So the result shows your first formant next to your pitch. The tool estimates it from an LPC spectral envelope computed on a downsampled copy of the recording. It is reported as a number with no label attached, because the useful comparison is with your own earlier readings: if F1 falls while pitch stays put, your voice has become fuller without becoming lower.
Listeners also respond to steadiness. A low voice that wobbles reads as less deep than a slightly higher voice that holds still, and the stability card, built from cycle-to-cycle jitter, shows which kind of take you just gave.
How to Get a Reading You Can Trust
Do not push your voice down for the microphone. Forcing a lower pitch than your speaking voice naturally uses gives you a number for that effort, and it usually makes the tone pressed and the stability worse. The reading that matters is the one you get from talking the way you talk.
Do not mistake fry for depth. Vocal fry, the creaky rattle at the very bottom of a voice, mostly falls below the 60 Hz floor or has no stable period, so the tool discards it. A take that is largely fry produces very few voiced frames and a reading that is not representative of anything.
Record in a quiet room, a forearm from the microphone, for the full eight seconds. Distance does not change your pitch, but noise causes frames to be thrown away, and a short take leaves the percentile readings with little to work from.
Try it in the afternoon. Most voices sit lowest in the morning and settle a little higher once they have been used for a few hours. Neither reading is wrong; the afternoon one is closer to how most people hear you most of the time.
Can You Make Your Voice Deeper?
Habitual speaking pitch has a physical floor set by the length and mass of your vocal folds, and no exercise changes that. What can change is how much of your natural range you actually use and how much of the tract resonates with it. Many people speak a few semitones above the bottom of their comfortable range out of habit, tension or nerves, and settle lower once they relax the larynx, support the breath from the diaphragm and stop tightening the throat. That change lives in habit and leaves the anatomy untouched, and it shows up on this test as a lower median with the same or better stability.
Fullness responds even more readily than pitch. Opening the throat, keeping the jaw loose and speaking at a comfortable volume tends to lower the first formant, and the voice reads as darker without dropping a single hertz. This is the direction most voice coaches push first, because it costs nothing and carries no strain.
What does not work is grinding the voice down into fry or pressing it against a closed throat. Both sound deeper for a sentence and both tire the voice quickly, and neither registers well on a pitch detector, because neither produces a clean, stable vibration. If a lower or rougher voice appears suddenly and stays for more than two or three weeks without you trying for it, that belongs with a doctor rather than a website.
For the pitch reading on its own, without the depth bands, use the voice Hz test. The voice gender detector reads the same pitch and formants against the masculine-to-feminine scale, and the rate my voice page scores clarity, stability and tone from the same take.
Measured on Real Recordings: Depth Bands Across 40 Readers
We placed the median speaking pitch of the 40 LibriSpeech dev-clean readers on the same depth bands this page uses.
| Depth band | 20 men | 20 women |
|---|---|---|
| Very deep | 5 | 0 |
| Deep | 11 | 0 |
| Moderately deep | 2 | 1 |
| Not deep | 2 | 18 |
| High | 0 | 1 |
Most men in this set read Deep or Very deep, and two read Not deep. No one reached Exceptionally deep, which sits below the typical adult masculine range. Depth also depends on resonance, which pitch vs resonance measures on the same readers.
Frequently Asked Questions
Last reviewed: September 2026
How deep is my voice?
Press record and speak normally for eight seconds. The result gives your median speaking pitch in hertz, the depth band it falls in, your lowest stable and highest readings, a stability figure and your first formant. Lower hertz means deeper; the formant explains how full it sounds at that pitch.
What Hz is considered a deep voice?
On this test, anything under 132 Hz is labelled deep, with 108 to 132 Hz as deep, 85 to 108 Hz as very deep and under 85 Hz as exceptionally deep. Those bands split the typical adult masculine speaking range of 85 to 155 Hz into thirds. They are this site’s own transparent thresholds and do not come from a clinical standard.
Is 100 Hz a deep voice?
Yes. A median speaking pitch of 100 Hz sits in the bottom third of the typical adult masculine range, which this test labels very deep. It is close to the note G2. Whether it also sounds full depends on resonance, which the result reports as the first formant.
Is 120 Hz a deep voice?
A 120 Hz median sits in the middle third of the typical adult masculine range, which this test labels deep. It is close to the note B2 and is a common reading for adult male speakers.
What is the average male voice frequency?
Adult masculine speaking voices typically centre around 85 to 155 Hz averaged across a sentence, which places the middle of the range near 120 Hz. Individual voices vary widely, and pitch also moves within one person by time of day, mood and volume.
Why does my voice sound deeper to me than the test says?
You hear your own voice partly through bone conduction, which carries low frequencies to your inner ear more efficiently than air does. Everyone sounds deeper to themselves than to a microphone or another listener. The test reports what the microphone received, which is closer to what other people hear.
Can vocal fry make my voice test deeper?
Usually not. Fry mostly sits below the 60 Hz floor of the pitch detector or has no stable repeating period, so those frames are discarded. A take that is mostly fry produces very few voiced frames, so the reading comes out unreliable instead of low.
Is my recording uploaded?
No. The pitch and formant analysis runs inside your browser tab through the Web Audio API. There is no server that receives audio, and the recording is discarded when you leave the page.