Voicing is determined by vibration of the vocal cords during the production of a single sound. If the vocal cords are vibrating, that sound is considered voiced. If they’re not vibrating, that sound is considered voiceless.
To feel this, put your hand on your throat and touch your Adam’s apple, and make the sounds [sssss] and [zzzzz]. You should feel vibration of your vocal cords during [z] but not during [s].
All sounds are either voiced or voiceless, and some pairs of sounds are identical in how they’re produced except for the voicing (such as [s] and [z] but also [f] and [v] and many other pairs). If the contrast in voicing is enough to change the meaning of the word, then these sounds are considered to be separate phonemes (see phonology.
In many Aboriginal languages, there is no voicing contrast. This means that speakers of these languages consider pairs of sounds as the same phoneme. For example, in Kunwinjku, [k] and [g] don’t change the meaning of any word, so they’re considered the same sound (which is why there’s a k but no g in the Kunwinjku alphabet), however the different sounds do tend to occur in different positions in the word.