Week 2 - SEGMENTATION Flashcards
First Perspective in PHONETICS:
Articulatory Phonetics
the study of the ORGANS OF SPEECH and their use in producing speech sounds
Second Perspective in PHONETICS:
Acoustic Phonetics
the study of the PHYSICAL PROPERTIES of the SOUNDS produced in speaking
Third Perspective in PHONETICS:
Auditory/Perceptional Phonetics
the study of the PERCEPTION of /speech sounds/ by the EAR, NERVOUS SYSTEM, and BRAIN
The Phonetic Stream of Speech - How does it look on a spectogram?
Shading at a point on a spectrogram indicates the intensity at that frequency at that point in time.
we can see: NO CLEAR CONSISTENT BOUNDARIES BETWEEN PHONES IN ACOUSTIC TERMS
Segmentation of the phonetic stream
Sequence of Discrete Phones Abstraction
Speech = Continuous stream of sound
CONTAINING: freq. LESS THAN CLEAT divisions into SEGMENTS
Speech as a SEQUENCE OF DISCRETE PHONES is a useful abstraction - eg. k^h - æ - t
Segmentation into phones = NOT explicict in forms of speech. they result from analysis of how speech is produced
USEFUL ANALYSES??
- YES, in description of language
- NO, in Electronic Speech Recog
Segmentation in terms of; PHONETIC TARGETS
We can analyse the phonetic stream as the result of moving the articulators from PHONETIC TARGET to Phonetic target
- individual sounds can be concieved of as having a sepcific ‘INTENDED’ articulation = ARTICULATORY TARGET
((and/or specific acoustic target))
BUT: articulators req. time n effort to move between targets, (sometimes articulatory targets aint reached)
Articulatory Transitions
- BRIEF, but auditorilary important
The transition of a VOWEL –> FOLLOWING CONSONANT provides auditory CUES in advance as to what that consonant is.
Even when the final consonant is digitally removed, or rendered indistinct due to eg sudden noise.
BAG –> BA– ( we still know its bag due to way a transitions)
Articulatory Undershoot
(normal speech) Articulators freq DO NOT MEET (presumed) ARTICULATORY TARGETS
UNDERSHOOT varies according to:
- STYLE N SPEED OF SPEECH
EG (g in “recognition”. airflow stops when careful pronounce. airflow keeps going in casual)
- PHYSICAL N EMOTIONAL STATE
- PHONETIC TARGETS OF NEIGHBOURING SOUNDS
What can an articulatory target be summarised in terms of?
Remember: Orchestra of Speech
a single phone results from ORCHESTRATED ACTION of many parts of vocal system
articulator target can b summarised in terms of:
Rascal Vermin Sneak Back To Back, Leaving Laura’sJAMBOREE
- Respiration
- Vocal Folds
- Soft Palate
- Body of Tongue
- Tip/Blade of Tongue
- Lips
- Lower Jaw
Alignment of articulatory actions
Individual articulatory actions are not perfectly
aligned - might be why its so hard to SEGMENT speech specifically (my thought)
Co-articulation
Transition between targets can result in an INCIDENTAL SIMULTANEOUS ARTICULATION which is NOT a specific PHONETIC TARGET: co-articulation.
eg /kt/ sequence in cocktail
Complex Consonants
artic. of SOME consonants cant be desc in terms of SINGLE PRIMARY CONSTRICTION
eg single primary constriction in velar [k]
THIS is distinct from incidental co-articulation, COMPLEX CONSONANTS = PHONETIC TARGET WITH 2 CONSTRICTIONS
- some 2 primary const
- some primary and secondary const.
Timing in Complex consonants
- (Near-) Simultaneous: both targets at same time
- Sequential: One after other
- Simultaneous onset, Sequential Release (EG: Clicks, Ejectives, Implosives)
TWO PRIMARY CONSTRICTIONS
NEAR-SIMULTANEOUS
- LABIAL-VELAR
> Labial-Velar Approx: voiced [w]> Voicless labial-velar fricative: [ʍ] in Scottish Eng [ʍɪtʃ] ‘which’> Labial-Velar Stops and Nasal: [k^p] [g^b] [m͡ŋ] - NB IPA bridge symbol
eg Yoruba [ak^pa] ‘arm’
Bridge symbol should look like [m͡ŋ]
PRIMARY AND SECONDARY CONSTRICTIONS
SEQUENTIAL
AFFRICATES: stop + fricative
- voiceless [tʃ] as in ‘church’
> compare English ‘catch it’ [tʃ] vs ‘catshit’ [t.ʃ]
> stop phase of affricate is typically not an actual apico-alveolar, but more retracted
> voiced [dʒ] as in ‘judge’
GERMAN:
- pfeffer “pepper” [pf}
- zeit “time” [ts]
