Skip to content

Commit a8826d5

Browse files
Update batch-transcription.md
1 parent fa613aa commit a8826d5

File tree

1 file changed

+54
-1
lines changed

1 file changed

+54
-1
lines changed

articles/cognitive-services/Speech-Service/batch-transcription.md

+54-1
Original file line numberDiff line numberDiff line change
@@ -67,7 +67,8 @@ Configuration parameters are provided as JSON:
6767
"properties": {
6868
"ProfanityFilterMode": "Masked",
6969
"PunctuationMode": "DictatedAndAutomatic",
70-
"AddWordLevelTimestamps" : "True"
70+
"AddWordLevelTimestamps" : "True",
71+
"AddSentiment" : "True"
7172
}
7273
}
7374
```
@@ -82,6 +83,7 @@ Configuration parameters are provided as JSON:
8283
| `ProfanityFilterMode` | Specifies how to handle profanity in recognition results. Accepted values are `none` which disables profanity filtering, `masked` which replaces profanity with asterisks, `removed` which removes all profanity from the result, or `tags` which adds "profanity" tags. The default setting is `masked`. | Optional |
8384
| `PunctuationMode` | Specifies how to handle punctuation in recognition results. Accepted values are `none` which disables punctuation, `dictated` which implies explicit punctuation, `automatic` which lets the decoder deal with punctuation, or `dictatedandautomatic` which implies dictated punctuation marks or automatic. | Optional |
8485
| `AddWordLevelTimestamps` | Specifies if word level timestamps should be added to the output. Accepted values are `true` which enables word level timestamps and `false` (the default value) to disable it. | Optional |
86+
| `AddSentiment` | Specifies sentiment should be added to the utterance. Accepted values are `true` which enables sentiment per utterance and `false` (the default value) to disable it. | Optional |
8587

8688
### Storage
8789

@@ -93,6 +95,57 @@ Polling for transcription status may not be the most performant, or provide the
9395

9496
For more details, see [Webhooks](webhooks.md).
9597

98+
## Sentiment
99+
100+
Sentiment is a new feature in Batch Transcription API and is an important feature in the call centre domain. Customers can use the `AddSentiment` parameters to their requests to
101+
102+
1. Get insights on customer satisfaction
103+
2. Get insight on the performance of the agents (team taking the calls)
104+
3. Pinpoint the exact point in time when a call took a turn in a negative direction
105+
4. Pinpoint what went well when turning negative calls to positive
106+
5. Identify what customers like and what they dislike about a product or a service
107+
108+
Sentiment is scored per audio segment where an audio segment is defined as the time lapse between the start of the utterance (offset) and the detection silence of end of byte stream. The entire text within that segment is used to calculate sentiment. We DO NOT calculate any aggregate sentiment values for the entire call or the entire speech of each channel. These are left to the domain owner to further apply.
109+
110+
Sentiment is applied on the lexical form.
111+
112+
A JSON output sample looks like below:
113+
114+
```json
115+
{
116+
"AudioFileResults": [
117+
{
118+
"AudioFileName": "Channel.0.wav",
119+
"AudioFileUrl": null,
120+
"SegmentResults": [
121+
{
122+
"RecognitionStatus": "Success",
123+
"ChannelNumber": null,
124+
"Offset": 400000,
125+
"Duration": 13300000,
126+
"NBest": [
127+
{
128+
"Confidence": 0.976174,
129+
"Lexical": "what's the weather like",
130+
"ITN": "what's the weather like",
131+
"MaskedITN": "what's the weather like",
132+
"Display": "What's the weather like?",
133+
"Words": null,
134+
"Sentiment": {
135+
"Negative": 0.206194,
136+
"Neutral": 0.793785,
137+
"Positive": 0.0
138+
}
139+
}
140+
]
141+
}
142+
]
143+
}
144+
]
145+
}
146+
```
147+
The features uses a Sentiment model which is currently in Beta.
148+
96149
## Sample code
97150

98151
The complete sample is available in the [GitHub sample repository](https://aka.ms/csspeech/samples) inside the `samples/batch` subdirectory.

0 commit comments

Comments
 (0)