In regard to vowels, based on the source-filter theory of speech production, the vocal tract as an air tube has infinite number of resonance peaks, i.e., formants. This would mean that the frequency range of vowels may be infinite. There are two critical limits, however. The first is that the vocal source has a downward spectral tilt of -6 dB (after radiation). This means that the energy of very high formants is too low to be audible. The second limit, which is much less well-known, is that the human articulatory system is not able to effectively manipulate formants higher than F3, because the distance between adjacent nodes and anti-nodes becomes increasingly short as the formant order becomes higher. The human articulators are simply not flexible enough to simultaneously constrict the vocal tract at all the nodes or anti-nodes of a formant to effectively lower or raise it in frequency beyond F3. The combination of the two limits is why formants beyond F5 is usually not even considered, and those beyond F3 are rarely measured.

In regard to consonants, the highest spectral peak of [s] is around 5-6 kHz. The effect energy of [f] can go much higher, but the perceptual effect of its high-energy components is questionable.

This is why acoustic analysis of vowels and consonants rarely goes over 8 kHz. The energy beyond 8 kHz has more to do with overall sound quality than with distinctive properties of speech sounds.

A good source of reference would be: Stevens, K. N. (1998). Acoustic Phonetics. Cambridge, MA: The MIT Press.

Hello, You can search Peter Roach's "English Phonetics and Phonology" and Allard Jongman's " Phonetics and Psycho linguistics Laboratory" for the supporting evidence you are looking for. Best regards, R. Biria

In regard to vowels, based on the source-filter theory of speech production, the vocal tract as an air tube has infinite number of resonance peaks, i.e., formants. This would mean that the frequency range of vowels may be infinite. There are two critical limits, however. The first is that the vocal source has a downward spectral tilt of -6 dB (after radiation). This means that the energy of very high formants is too low to be audible. The second limit, which is much less well-known, is that the human articulatory system is not able to effectively manipulate formants higher than F3, because the distance between adjacent nodes and anti-nodes becomes increasingly short as the formant order becomes higher. The human articulators are simply not flexible enough to simultaneously constrict the vocal tract at all the nodes or anti-nodes of a formant to effectively lower or raise it in frequency beyond F3. The combination of the two limits is why formants beyond F5 is usually not even considered, and those beyond F3 are rarely measured.

In regard to consonants, the highest spectral peak of [s] is around 5-6 kHz. The effect energy of [f] can go much higher, but the perceptual effect of its high-energy components is questionable.

This is why acoustic analysis of vowels and consonants rarely goes over 8 kHz. The energy beyond 8 kHz has more to do with overall sound quality than with distinctive properties of speech sounds.

A good source of reference would be: Stevens, K. N. (1998). Acoustic Phonetics. Cambridge, MA: The MIT Press.

Thank you very much for your detailed and helpful answers!