I’m starting this thread at the request of a member after a series of comments in another thread regarding this graph:
The graph, with no title and insufficient labels, is a little misleading. It does not appear to be wrong, as it shows the harmonics of the voice and orchestra with a reasonable amount of accuracy. It does not, however, show the possible fundamental pitches produced by either singers or orchestral instruments, nor does it indicate that only singers using western “art music” techniques (so-called bel canto singing technique) will be able to produce the strong formants (not harmonics, more on this later) in the 3 kHz range.
(Begin Edit): The second potential issue is we might not all be as aware as we should of what vocal ranges exist and what our speakers should be able to accurately produce. I hope to clarify what someone might expect to hear from recorded voices, including fundamental frequencies (the singer's "range"), as well as the unique way singers are able to manipulate the way the harmonic series is presented. (End Edit)
Why should you believe what I have to say? Without going into unnecessary detail, I have a doctoral degree in music and my main profession for the past two decades has been music professor. I am currently teaching choral and vocal courses at one of largest universities in the US. If you desire further reading on the subject, I would direct you to the work of Scott McCoy, James McKinney, Barbara Doscher, Richard Miller, Ingo Titze, and Johan Sundberg to name a few. The McCoy text, Your Voice: An Inside View has excellent graphs, spectral analyses, and a multimedia component that are useful. Similarly, Miller’s Training Tenor Voices contains an appendix comparing spectrographs taken from famous tenors in notable passages from the repertory.
Where orchestral instruments are concerned, the matter is generally clear and straightforward. Anyone can look up a table of the ranges of orchestral instruments (from C0 at c. 16 Hz on a large pipe organ to C8 at c. 4186 Hz possible on several instruments). The harmonics that are emphasized for each instrument determine their characteristic sound. For example, clarinets famously emphasize every other harmonic with only odd numbered partials emphasized.
I will refer to “singers,” for the purposes of this writing, as classically trained, professional level singers utilizing the common western (bel canto aka “operatic”) technique. The lowest fundamental tone commonly called for in any voice part is low B-flat (c. 58 Hz) called for in the bass part of some Russian choral music (notably both the Rachmaninoff Vespers and Liturgy of St. John Chrysostom). Most bass arias in opera or oratorio don’t delve below the E or F above that (c. 82 or 87 Hz respectively). On the other end of the spectrum, the tenor high C that made Pavarotti famous as the “King of the High Cs” is at c. 523 Hz and the high F in Mozart’s Queen of the Night aria for coloratura soprano lies at c. 1400 Hz. This is around the limit of what would be expected for the highest tenor and soprano voices in western music. Some singers (both men and women) are able to access a so-called “whistle register” to reach notes higher than that (Mariah Carey, for example, uses whistle register to sing a reported high G-sharp at c. 3322 Hz in “Emotions”), but this makes up an exceedingly small percentage of the recorded repertory, and virtually nothing from the standard western “classical” canon.
Unlike typical acoustic instruments, the resonating space of the human voice is flexible. If one compares the voice to an acoustic guitar, the vocalis muscles are the analog to the strings—they vibrate, but without a resonator they are quiet and dull sounding. The body of the guitar is analogous to the throat and mouth (pharynx) of the singer (not the chest or sinuses—I won’t cite all the research here, but despite that fact that singers frequently claim to feel vibration in these locations, there are many studies that show these regions are NOT involved in sound production—that is, there is no such thing as “chest” in the sound in any literal way).
Different from a guitar (or nearly any other acoustic instrument), however, is the fact that the size and shape of the human pharynx is constantly changing. At any given pitch the harmonic series produced by the vocal muscles vibrating is much like that of any acoustic instrument. However, “formants,” which are not the same as harmonics, are produced by changing the physical characteristics of the column of air in the pharynx. Singers generally create three main formants: the first two create the vowel sound. You can experiment with this by closing your glottis, forming a vowel with your mouth, and flicking or tapping the fleshy part of your chin or cheek. You are hearing the first vowel formant. If you whisper a vowel you are hearing the second vowel formant. If you whisper them in order from the “brightest” to the “darkest” of the cardinal vowels: ee, eh, ah, oh, oo, you will hear the pitch descend with each vowel. All spoken or sung vowels must contain both formants. The pitches of these two formants fall within a very narrow range and do not change from person to person or vary by the pitch level of the fundamental tone. This is why when women sing very high notes it does not sound like a vowel anymore—they are singing pitches higher than the first (or sometimes the first and second) vowel formant. This is what makes vowel modification necessary, but I digress.
The third formant is commonly referred to as the “singer’s formant” and is an emphasis on whatever harmonic(s) lies near 3 kHz (technically between 2400 and 3200Hz). This emphasis can be quite significant, leading to a situation where the third formant sounds more strongly than the fundamental pitch. Despite this, we only hear the singer’s formant as a strong “ring” or resonance. Our brains do not confuse the singer’s formant for the fundamental pitch. In the best operatic singers, this formant is so strong that it sounds more loudly in that pitch range (c. 3000 Hz) than any orchestral instrument, thereby allowing singers’ voices to “carry” over the sound of an orchestra.
It should be noted that sopranos rarely produce the singer’s formant as the natural pitch of their voices does not require it in order to be heard. Also, singers who utilize styles that are not the western bel canto style do not typically produce the necessary harmonic content required for the singer’s formant to exist. For example, a singer producing a breathy sound (such as is frequently used in jazz singing) cannot produce a singer’s formant, nor can a singer producing a tense, tight, or pressed sound. Producing such a strong third formant while singing with a microphone for amplification would undoubted prove problematic anyway.
So, returning to the graph above, it shows only the harmonic content of the orchestra and only formants of trained voices singing in a classical style. We have seen that the lowest fundamental frequencies possible from several orchestral instruments and fundamental frequencies produced by many male singers are below what is shown in the graph. The graph also clearly shows a peak at the level of the singer’s formant, which is far beyond the range of most singers, regardless of training. The graph appears to be exactly correct, but without clearer labels and explanation it is confusing.
(After writing this post and pasting it here, I discovered that the title of the image is "formant-singer.jpg" thus appearing to confirm my view of what the graph is actually showing.)
The graph, with no title and insufficient labels, is a little misleading. It does not appear to be wrong, as it shows the harmonics of the voice and orchestra with a reasonable amount of accuracy. It does not, however, show the possible fundamental pitches produced by either singers or orchestral instruments, nor does it indicate that only singers using western “art music” techniques (so-called bel canto singing technique) will be able to produce the strong formants (not harmonics, more on this later) in the 3 kHz range.
(Begin Edit): The second potential issue is we might not all be as aware as we should of what vocal ranges exist and what our speakers should be able to accurately produce. I hope to clarify what someone might expect to hear from recorded voices, including fundamental frequencies (the singer's "range"), as well as the unique way singers are able to manipulate the way the harmonic series is presented. (End Edit)
Why should you believe what I have to say? Without going into unnecessary detail, I have a doctoral degree in music and my main profession for the past two decades has been music professor. I am currently teaching choral and vocal courses at one of largest universities in the US. If you desire further reading on the subject, I would direct you to the work of Scott McCoy, James McKinney, Barbara Doscher, Richard Miller, Ingo Titze, and Johan Sundberg to name a few. The McCoy text, Your Voice: An Inside View has excellent graphs, spectral analyses, and a multimedia component that are useful. Similarly, Miller’s Training Tenor Voices contains an appendix comparing spectrographs taken from famous tenors in notable passages from the repertory.
Where orchestral instruments are concerned, the matter is generally clear and straightforward. Anyone can look up a table of the ranges of orchestral instruments (from C0 at c. 16 Hz on a large pipe organ to C8 at c. 4186 Hz possible on several instruments). The harmonics that are emphasized for each instrument determine their characteristic sound. For example, clarinets famously emphasize every other harmonic with only odd numbered partials emphasized.
I will refer to “singers,” for the purposes of this writing, as classically trained, professional level singers utilizing the common western (bel canto aka “operatic”) technique. The lowest fundamental tone commonly called for in any voice part is low B-flat (c. 58 Hz) called for in the bass part of some Russian choral music (notably both the Rachmaninoff Vespers and Liturgy of St. John Chrysostom). Most bass arias in opera or oratorio don’t delve below the E or F above that (c. 82 or 87 Hz respectively). On the other end of the spectrum, the tenor high C that made Pavarotti famous as the “King of the High Cs” is at c. 523 Hz and the high F in Mozart’s Queen of the Night aria for coloratura soprano lies at c. 1400 Hz. This is around the limit of what would be expected for the highest tenor and soprano voices in western music. Some singers (both men and women) are able to access a so-called “whistle register” to reach notes higher than that (Mariah Carey, for example, uses whistle register to sing a reported high G-sharp at c. 3322 Hz in “Emotions”), but this makes up an exceedingly small percentage of the recorded repertory, and virtually nothing from the standard western “classical” canon.
Unlike typical acoustic instruments, the resonating space of the human voice is flexible. If one compares the voice to an acoustic guitar, the vocalis muscles are the analog to the strings—they vibrate, but without a resonator they are quiet and dull sounding. The body of the guitar is analogous to the throat and mouth (pharynx) of the singer (not the chest or sinuses—I won’t cite all the research here, but despite that fact that singers frequently claim to feel vibration in these locations, there are many studies that show these regions are NOT involved in sound production—that is, there is no such thing as “chest” in the sound in any literal way).
Different from a guitar (or nearly any other acoustic instrument), however, is the fact that the size and shape of the human pharynx is constantly changing. At any given pitch the harmonic series produced by the vocal muscles vibrating is much like that of any acoustic instrument. However, “formants,” which are not the same as harmonics, are produced by changing the physical characteristics of the column of air in the pharynx. Singers generally create three main formants: the first two create the vowel sound. You can experiment with this by closing your glottis, forming a vowel with your mouth, and flicking or tapping the fleshy part of your chin or cheek. You are hearing the first vowel formant. If you whisper a vowel you are hearing the second vowel formant. If you whisper them in order from the “brightest” to the “darkest” of the cardinal vowels: ee, eh, ah, oh, oo, you will hear the pitch descend with each vowel. All spoken or sung vowels must contain both formants. The pitches of these two formants fall within a very narrow range and do not change from person to person or vary by the pitch level of the fundamental tone. This is why when women sing very high notes it does not sound like a vowel anymore—they are singing pitches higher than the first (or sometimes the first and second) vowel formant. This is what makes vowel modification necessary, but I digress.
The third formant is commonly referred to as the “singer’s formant” and is an emphasis on whatever harmonic(s) lies near 3 kHz (technically between 2400 and 3200Hz). This emphasis can be quite significant, leading to a situation where the third formant sounds more strongly than the fundamental pitch. Despite this, we only hear the singer’s formant as a strong “ring” or resonance. Our brains do not confuse the singer’s formant for the fundamental pitch. In the best operatic singers, this formant is so strong that it sounds more loudly in that pitch range (c. 3000 Hz) than any orchestral instrument, thereby allowing singers’ voices to “carry” over the sound of an orchestra.
It should be noted that sopranos rarely produce the singer’s formant as the natural pitch of their voices does not require it in order to be heard. Also, singers who utilize styles that are not the western bel canto style do not typically produce the necessary harmonic content required for the singer’s formant to exist. For example, a singer producing a breathy sound (such as is frequently used in jazz singing) cannot produce a singer’s formant, nor can a singer producing a tense, tight, or pressed sound. Producing such a strong third formant while singing with a microphone for amplification would undoubted prove problematic anyway.
So, returning to the graph above, it shows only the harmonic content of the orchestra and only formants of trained voices singing in a classical style. We have seen that the lowest fundamental frequencies possible from several orchestral instruments and fundamental frequencies produced by many male singers are below what is shown in the graph. The graph also clearly shows a peak at the level of the singer’s formant, which is far beyond the range of most singers, regardless of training. The graph appears to be exactly correct, but without clearer labels and explanation it is confusing.
(After writing this post and pasting it here, I discovered that the title of the image is "formant-singer.jpg" thus appearing to confirm my view of what the graph is actually showing.)
Last edited:
