Voice Pitch Test: How Deep Is My Voice?

Your result names a band from Deep voice to High voice, and each band is a share of speaking voices in the 2022 study: the test works out what share of men (or women) speak higher than you, from the study’s average and standard deviation (read as a bell curve, which is how a published norm is used), and the band follows from that share. The cut-offs are the 15th, 35th, 65th and 85th percentiles, so the five bands hold about 15, 20, 30, 20 and 15% of voices.

The same five bands for women’s speaking voices, from the same study. A woman with a low voice often reads Deep against women and High against men; the test shows both.

Reading aloud, the men in the 2022 study averaged 115 Hz, close to A#2 on a piano, and the women 199 Hz, close to G3. Once age was accounted for, women spoke 76 Hz higher than men, and for every year of age speaking pitch fell by 0.37 Hz, so a 60 year old reads about 15 Hz lower than a 20 year old on average.

Older adults speaking in conversation give a second view. Among 1,154 men and 1,318 women aged 40 to 79 in Leipzig (Berg et al. 2017), men averaged 111.9 Hz and women 168.5 Hz. Men’s pitch rose with age in that group while women’s did not, and the authors note their women spoke six to seven semitones lower than earlier studies had described. A third study points the same way: women around 72 spoke lower than women around 19, while older and younger men did not differ (Goy et al. 2013). If you are over 40, read your band knowing the 2022 norms come from adults 18 to 60.

On this test, a deep voice for a man is 93 Hz or lower: lower than at least 85% of men’s speaking voices. 94 to 107 Hz is deeper than average. For a woman, deep is 170 Hz or lower. These are cut-offs on a published distribution, not a verdict on how your voice sounds: depth is one part of a voice, and resonance, loudness, pace and clarity change how deep it reads to a listener.

Is deeper better? In one listening experiment, women preferred the lower of two men’s voices until the lower one went below about 96 Hz, and from there they preferred the higher one; men preferred the higher of two women’s voices at every pitch tried (Re, O’Connor, Bennett and Feinberg 2012, PLoS ONE). That study used 19 university students and single vowel sounds, so it says something about pitch alone, not about whole voices or people.

Voicecel comes from looksmaxxing forums, where it names a man whose high or thin voice is said to cost him in looks. It is forum convention, not research: there is no agreed hertz cut-off, and the forums use it as an insult as often as a description. This page does not call anyone a voicecel.

What a measurement can say: for men, 124 to 137 Hz is higher than average (only 15 to 34% of men speak higher) and 138 Hz or higher is higher than about 85% of men. If you came from the looksmax side, note that voice is not part of a PSL score or the looksmax AI: both read the face from a photo.

Speaking pitch moves through the day and with how you speak, so take two readings on different days before you settle on a number.

The pitch tracking itself is the accurate part. We ran this page’s exact code on the passage read by 28 of the computer voices built into macOS, from 69 to 297 Hz, and compared it with Praat, the pitch software speech scientists use: the speaking pitch agreed within 0.3 Hz at the median and 6 Hz at most. Most of the gap is which moments count: Praat keeps creaky, falling ends of sentences that this test drops (any moment with an aperiodicity over 0.2), much as the 2022 study separated creak out before averaging.

The microphone matters less than people think. With room noise 15 dB under the voice, a phone-style microphone that cuts everything under 200 Hz (which removes the fundamental of most men’s voices, leaving its overtones), or a voice 30 dB quieter, the readings moved by 5.7 Hz at most. After the first 4 seconds of voice the reading was within 6.3 Hz of the whole reading, after 8 seconds within 4.2 Hz.

What we could not measure: those were synthetic voices, not people. The bigger source of error is you, from day to day and from one way of reading to another, and no tracker removes that. Take two readings, on two days, at the same volume.

This test reads your speaking voice: the pitch you talk at, which is what people mean by a deep or a high voice. Your singing range is a different number, the distance from the lowest to the highest note you can hold. To find it, and the voice type it fits from bass to soprano, use the vocal range test.