Uploaded December 2014 | Updated November 2026, 3 hours ago
Carl Gunnar Michael Fant, an electrical engineer and one of the pioneers in the area of speech synthesis, defined formants as "the spectral peaks in the sound spectrum of the voice". It can also refer to an acoustical resonance in a human voice.
Formants, like harmonics, are generally present in a series, with the lowest frequency formant f1, and the next highest frequency formant f2, then f3, etc.
Formants create the timbre of a voice, a musical instrument, and other sounds. Because of this, the analysis of formants is essential for voice prints, to positively establish the identity of a voice.
Generally, the formant relationship will be shifted whenever audio is pitch-shifted, resulting in a drastic change in timbre. For example, when a voice is pitch- shifted one octave, it sounds cartoony (like "the chipmunks").
However, some signal-processing plug-ins allow pitch shifting, while maintaining the formant relationship, so that when a voice is transposed within a fairly broad range, the result still sounds like a human voice.
Carl Gunnar Michael Fant, an electrical engineer and one of the pioneers in the area of speech synthesis, defined formants as "the spectral peaks in the sound spectrum of the voice". It can also refer to an acoustical resonance in a human voice.
Formants, like harmonics, are generally present in a series, with the lowest frequency formant f1, and the next highest frequency formant f2, then f3, etc.
Formants create the timbre of a voice, a musical instrument, and other sounds. Because of this, the analysis of formants is essential for voice prints, to positively establish the identity of a voice.
Generally, the formant relationship will be shifted whenever audio is pitch-shifted, resulting in a drastic change in timbre. For example, when a voice is pitch- shifted one octave, it sounds cartoony (like "the chipmunks").
However, some signal-processing plug-ins allow pitch shifting, while maintaining the formant relationship, so that when a voice is transposed within a fairly broad range, the result still sounds like a human voice.










