Dataset Viewer
Auto-converted to Parquet
Search is not available for this dataset
audio
audioduration (s)
0.1
10.8
label
class label
13 classes
7present01
3object01
2no
0down
7present01
6on
4object10
8present10
5off
1left
1left
2no
3object01
1left
5off
10stop
5off
2no
3object01
2no
3object01
5off
9right
10stop
11up
9right
3object01
6on
0down
10stop
6on
9right
8present10
12yes
9right
4object10
6on
4object10
1left
7present01
8present10
6on
11up
4object10
4object10
1left
5off
4object10
7present01
5off
3object01
10stop
12yes
6on
3object01
10stop
5off
10stop
11up
3object01
11up
10stop
4object10
0down
4object10
1left
2no
4object10
3object01
6on
4object10
5off
12yes
6on
11up
5off
6on
1left
11up
9right
1left
12yes
1left
9right
1left
11up
2no
4object10
10stop
12yes
2no
1left
11up
3object01
9right
1left
3object01
5off
10stop
3object01
End of preview. Expand in Data Studio

This is a draft dataset of prosodic minimal pairs. It includes two stress doublets, the noun and verb uses of object and of present.
Remaining words are from superb ks. object10 and present10 tokens lexically have left stress. object01 and present01 tokens lexically have right stress.

There are separate validation splits for each of object10, present10, object01 and present01. In each of these splits, all the tokens have the same segmental content and lexical prosody. Test splits are not here yet.

Downloads last month
54