Lecture 6 - Consonants

Consonants: Produced with substantial constriction - sometimes no air flow 

What are the consonant choices in English? 

1. Minimal pair = a pair of words that differ in only one phoneme 

   1. pin/bin/din/tin/chin

      1. Different words buy they have the same /in/, what differs is the consonant that comes first which allows us to generate different words 

2. We have 24 different choices in English (approximately because of different accents/dialects) 

3. Pulmonic = air is flowing from the lungs 

4. Bilabial - produced by the closure of the two lips - (|constriction)

5. Labio-dental - lower lip and upper teeth closure 

6. Inter-dental - between the teeth 

7. Alveolar - the constriction is the tip of the tongue moving alongthe alveolar ridge 

8. Palatal alveolar - between hard ridge and alveolar ridge 

9. Palatal - top of palatal - /j/

10. Velar - back of tongue is constricted against velum 

11. Glotal - /h/ no air…?

How are they articulated? 

1. Stops are made by stopping airflow along the larynx-pharynx-oral cavity path

2. The velum is raised closing off the nasal cavity from the pharynx and oral cavity 

3. There will be a “burst” of air flow when the constriction is stopped 

4. /p/ /b/ /k/ /d/ /t/

1. Fricatives are made by restricting airflow enough to cause TURBULANCE along the larynx-pharynx-oral cavity path

2. The degree of constriction is slightly less than that of a stop - not completely cutting off airflow

1. Nasals (nasal stops: the velum is open, there is no airflow in the oral cavity so it goes through the nasal cavity 

2. When the velum lowers we see resonance of the nasal cavity 

1. Affricates are ‘stop’ like followed by a fricative like part 

2. Stopping airflow along the larynx-pharynx-oral cavity path but no burst the release of the closure is slower so the burst is much longer and more fricative like 

Approximants: vowel-like constants - accoustically similar to vowels but phonologically (show up in the same place as consonants) 

/w/ + /y/ → very similar to vowels /u/ and /I/ → semi-vowels or semi-consonants 

/l/ + /r(our one)/ - liquid approximants →it is very difficult for second language learners and children 

We apply source filter theory to understand their acoustics 

1. They have clear formant structure 

2. Lots of resonance coming through 

What is the difference between [aia] : where you clearly hear 3 different vowels, and [aja] 

The only difference is the duration of the sound! The vowel holds it for a long time / This makes us perceive a difference between vowels and semi-vowels. Given they are so short in duration, we can use them in places of a syllable where you can only find consonants e.g. “yellow” “mellow” “bellow” 

Liquids have both rotics /r/ and laterals /l/ (airflow around the sides of the constriction)

Nasals - they are considered nasal stops because there is complete closure of the oral cavity however instead of lowering the velum it stays lifted, allowing air to flow through the nasal cavity instead. 

Both liquids and nasals allow for air to flow from somewhere  → so areodynamically, they are very similar. It is often that you will confuse a nasal for a liquid or viceversa 

However, nasals have anti-resonance that bounces back 

Nasals have a similar acoustic as that of vowels so we apply the same theory  → a version of source filter theory (a bit more complicated). The length of the tube affects what sound we make! If the tube is shorter we end up with /ŋ/ since it is close to wear the air goes but if it is longer, we end up with /m/ as it resonates in the tube for longer. 

Fricatives 

Layers of fluid flow

1. Laminar flow - predictable 

2. Turbulent - unpredictable - the rate of the flow of fluid is so fast that the walls of the vocal tract have some sort of friction as it flows against the pipe. Because it is so fast, it starts to tear off the wall and breaks off other parts of the fluid, and this makes random fluids bounce. When it exits, we get a random noise because they are broken apart

Narrow constriction + same volume of fluid = increased pressure/velocity - if we have the same amount of fluid we need a faster flow and if we have fast flow the layers interact, break off and go in random directions  

AIR IS A FLUID THE SAME PRINCIPLE APPLIES - IF WE HAVE THE SAME AMOUNT OF AIR COMING OUT OF A SMALLER VOCAL TRACT, THE FLUID WILL HAVE TO MOVE FAST AND IT MOVES SO FAST THAT YOU GET TURBULENT FLOW = FRICATIVES!! WE CANT SEE BUT WE CAN HEAR AND IT CAUSES A STATIC SOUND THEY COME IN BOTH VOICED AND VOICELESS 

As the sound propegates from mouth to ear is also random, we get a random waveform 

The sound source for a fricative is essentially flat because of the nature of turbulent flow. You can still filter this since it is a flat amount of energy 

MANY DIFFERENT REPETITIONS OF SAYING THE SAME SOUND OVERLAYED - we can see “peaks” these will be the resonance of the distance in the cavity. As the constriction moves back we get earlier peaks because the cavity is getting further and further. 

ON A SPECTROGRAM it looks random and static depression of energy across different frequencies that has more energy in some areas because of how it is filtered. The resonance frequency is lower when the tube is longer 

Some fricatives are particular: sibilance - have higher amplitude - we naturally use sibilance to get attention or call the cat - this cmes from the fact that teh constriction is right behind the front teeth 

Voiced fricative: air flow from vibration and air flow from constriction 

Plosives - time is so important 

1. Closing phase 

   1. Moving into the position 

2. A hold phase 

   1. The complete closure is help for air pressure to build 

3. A release phase 

   1. The air is released 

4. OPTIONAl aspiration 

   1. A small burst / a very short period of turbulent flow - is then filtered and comes our 

FOR THE SOUND [B] = VOICED AND UNASPIRATED 

1. You have to consider the sound that comes before it , because during phase A you have a transition of the vocal tract shape and you get the resonance pointing in the direction of what they would be if there was air resonating. It cant resonate because there is no air as it is constricted but the formants point to what COULD have been 

The bursts are short intervals of direction occurring as the closure is released. 

The formants point to where there would be sound if there was sound 

VOICE ONSET TIMING (VOT) - vocal cord are not adducted 

Plosives are complete closure 

In order for voicing to occur, you need a pressure difference in the glotal - these two things are at odd with eachother since you can have flow if the vocal tract is closed. 

VOT the delay between the release of a plosive and then start of phonation of the following sound - so you have to take in the phases - if voicing occurs before the release this is a NEGATIVE VOT- if they happen at the moment of release =  0 VOT - voicing occurs after release = POSITIVE VOT 

The difference between these is how we chracterise difference 

VOICING IS ABOUT WHETHER THE VOCAL FOLDS ARE READY FOR VIBRATION WHEN THE BURST IS RELEASED - 

1. NEGATIVE -  yes because theyve already been vibrating 

2. POSITIVE - no they are not because there is a gap between 

Are the vocal folds vibrating while there is turbulence? 

[p] voiceless, aspirated 

The vocal folds are not ready to vibrate, air flows , turbulent air, = aspiration 

Voicing for stops is more difficult because it becomes more theoretical “would it occur” 

The way that the VOT continuum is used differs from language to language.

The moment the lips are released voicing does not happen straight away 

English uses 0VOT for voiced consonant but spanish uses it for voiceless = misinterpretations 

coursework WEEK 6: production and perception - it is VOT that differentiates voicing for STOP consonants not whether voicing is occurring. For english it is near 0VOT  

PERCEPTIALLY SENEISTIVE TO SMALL PHONETIC CATEGORIES. BASED ON AERODYNAMIC DIFFERENCES. EVEN IF THEY ARE NOT IN THE PHONOLOGY OF THE LANAGUGAE