Consonant sounds are divided up into many categories as can be seen on the table on the main page of this site. The first of these categories is the nasals. These sounds have sawtooth-like plots in the time domain. They are quasi-periodic and have anit-resonances that 'pull down' their spectral envelopes. Their spectrograms display the dominance of the low resonances created by long vocal tract that results from the opening of the nasal cavity.
The whisper is similar to the unvoiced fricatives (see below) in that there are no vocal fold vibrations but its source is solely the glottal opening rather than with a vocal tract constriction.
Another category of consonants is the affricates. Affricates are sometimes categorized with the transitional sounds but in this study the transitional class was limited to voiced sounds thus placing the affricates in with the consonants. The combination of plosives and fricatives (see below for both) characterize affricates. Voiced affricates are quasi periodic in their waveforms and unvoiced affricates exhibit noisy spectra.
Fricatives are the largest category of consonants and like affricates they are also further divided up into voiced and non-voiced sections. All fricatives consist of noise in their spectra that is a result of oral tract constriction narrower than that of vowels. Voiced fricatives are quasi-periodic but also have noise accompanying their temporal waveform. This noise is better illustrated in the frequency domain representations of the signals. Sometimes this voicing occurs simultaneously with the noise source and other times it occurs after. Examples of each of these are /z/ and /v/ respectively. Unvoiced fricatives are mainly just noise, see /T/ for an example.
The last category of consonants is plosives. Plosive sounds are impulsive in the time domain and also can be either voiced or unvoiced. Unvoiced plosives are quick temporal impulses with noisy spectra. Voiced plosives are also quick in the time domain but the voice bar preceding them is quasi-periodic. The quasi-periodic voice bar can be seen from .02s - .05s in /g/ and is then followed by an impulse.