You Are The Only One Who Hears Your Own Voice

Say something out loud. That sound reached your inner ear by two routes at once, and only one of those routes exists outside your head.

The obvious route is the air: sound leaves your mouth, curves around the outside of your head and arrives at your ear canal. That is air conduction, and it is the only one most people know about. The second route runs through you. Speaking shakes your jaw and your skull, and bone carries that vibration inward directly to the inner ear.

How the two compare was measured directly. Reinfeldt, Ostli, Hakansson and Stenfelt put 20 normal-hearing adults aged 23 to 34 in a custom 1.2 kg ear-muff, built to attenuate air-conducted sound by at least 20 dB without inducing an occlusion effect, and compared the two contributions phoneme by phoneme. Their finding, in their own words: the air-conducted and bone-conducted components are approximately the same order of magnitude — and strongly frequency dependent. Between 1 and 2 kHz the bone component led for most of the ten phonemes tested: roughly 12 dB above the air component for the nasals /m/ and /n/, and about 15 dB above it at 2 kHz for the back vowel /o/. Von Bekesy reached the same order-of-magnitude conclusion back in 1949. Porschmann put the bone-dominant band at 0.7 to 1.2 kHz.

So which route is louder depends on the sound you are making. There is no single percentage, and anyone handing you one has quietly rounded off the frequency dependence.

Now the part that is actually about you. A microphone sits out in the air. It receives the air-conducted component and nothing else, because the bone-conducted component never enters the air at all. The same is true of every other person in the room. The blend of air and bone that you have listened to your entire life is assembled inside your own skull, has never reached another human being, and cannot be captured by any device that has ever been built. The discomfort of hearing the recording even has a name in the literature — voice confrontation, Holzman and Rousey, 1966.

Which inverts the usual framing. The recording is not the weird version of your voice. It is the public version. The private one is unrecordable, and it belongs to exactly one listener.

One thing this video deliberately does not say: that bone conduction "boosts the low frequencies, which is why you sound deeper to yourself". It is the single most repeated line on this topic and the primary measurements do not support it as stated — Reinfeldt found bone dominance in the 1 to 2 kHz mid-range, and Porschmann found the air component dominating below 0.7 kHz. If you want to explain why a recording sounds unfamiliar, the honest version is that the spectral balance is different because one of two pathways has been removed. Not that the bass was subtracted.

Sources:
https://www.diva-portal.org/smash/get/diva2:345783/FULLTEXT01.pdf
https://doi.org/10.1121/1.3458855
https://pubmed.ncbi.nlm.nih.gov/20707445/
https://www.researchgate.net/publication/257128919_Influences_of_bone_conduction_and_air_conduction_on_the_sound_of_one%27s_own_voice
https://pubs.aip.org/asa/jasa/article-abstract/121/3/1576/952184/Examination-of-bone-conducted-transmission-from
https://doi.org/10.1037/h0023518
https://en.wikipedia.org/wiki/Voice_confrontation