Phoneticians define phonation as "use of the laryngeal system to generate an audible source of acoustic energy, i.e., sound, which can then be modified by the articulatory actions of the rest of the vocal apparatus."
A voiced sound is produced when the vocal folds vibrate, by an overpressure of air in the lungs. This produces a pure tone. It is further modified by movements in the vocal tract, by the volume of the airflow and by the degree of constriction of the folds. The vowels are usually voiced, but there are also voiced consonants.
If the vocal folds do not vibrate, then the sound (usually a consonant) is called voiceless.
Other sounds may be produced by completely blocking off the airstream and then releasing the folds. The sound of this type produced at the glottis is called a glottal stop.
The voice source is used to change intonation and the tone of words by varying the subglottal pressure as well as the tension of the vocal folds. This leads to changes in the frequency of vibration, which are in turn perceived by the listener as modifications in pitch and/or in loudness. During speech the flow of air is relatively small because of constrictions of the folds.
Subglottal pressure is regulated by a number of factors, namely: the respiratory muscles, gravity and elasticity.
In English, every voiced fricatives corresponds to a voiceless one and vice versa. The main distinction between the pairs of English stop consonants, however, is not voiced vs. voiceless since English /b d g/ may be voiceless in certain dialects. English /p t k/ are aspirated.
Certain English morphemes have both voiced and voiceless allomorphs, e.g. the ending -s (voiced in kids but voiceless in kits /kɪts/) or the past tense ending -ed (voiced in buzzed /bʌzd/ but voiceless in fished /fɪʃt/.
In languages such as French, all obstruents occur in pairs of a voiced one and an unvoiced one.
Mandarin Chinese has pairs of aspirated and unaspirated voiceless plosives and aspiration is the main distinction rather than phonation.
A few languages, such as Finnish or Alemannic, have no voiced obstruents but pairs of long and short consonants instead.