|
|
|
|
The ABI-1 corpus consists of speech from 14
locations representing different
Accents of the British Isles. In each location, 10 male and 10 female
subjects were chosen, who had lived in the location all their lives.
The corpus contains approximately 20 minutes of
read speech plus a small amount of spontaneous speech from each subject .
|
|
|
|
|
The PUMA corpus was created at the University of Birmingham as part of the
DTI/EPSRC funded PUMA project. The corpus consists of recordings
of induividuals interacting by voice
with an intelligent TV scheduler over a six month period.
|
|
|
|
|
The PF-STAR corpus is a collection of recordings of
British English children's speech which was
created by the University of Birmingham as part of the EU FP5 PF-STAR
children's speech corpus.
|
|
|
|
|
ABI-2 is the second "Accents of the British Isles" corpus and
comprises speech from 13 locations in British Isles representing accents which are not
included in the original ABI-1 corpus. The same recording procedures were used as in ABI-1.
The material recorded includes all of the ABI-1 prompts plus extra SCRIBE sentences. In
addition, ABI-2 subjects record 5 minutes of telephone conversational speech.
|
|
