| Conclusions |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| • | Dialects
can have very different tune shapes, but they make
similar |
|||||||||||||
| question/statement distinctions. |
||||||||||||||
| • | Old reliable f0 … |
|||||||||||||
| • | … is
indeed the best way to make the 2-way D/DQ
decision. |
|||||||||||||
| • | …doesn’t tell you much about the 4-way
distinction. |
|||||||||||||
| • | Other
acoustic properties… |
|||||||||||||
| • | …carry a substantial amount of
information. |
|||||||||||||
| • | The 2-way
Declarative/DQ distinction is… |
|||||||||||||
| • | …
clearest at the end, but |
|||||||||||||
| • | …spread across the utterance, including the very
beginning (e.g. |
|||||||||||||
| spectral
slope) |
||||||||||||||