How to Use the SDK to Enable Automatic Language Identification in AWS Transcribe Streaming ?

Question

As the title suggests, I am trying to use the Python SDK to utilize the Transcribe service. I found that the console allows the use of automatic language identification during streaming, but I can't find any parameter settings to use this feature in the SDK. Is there a way to enable automatic language identification in Transcribe streaming using the SDK?

Below are the parameters I have found:

```
language_code: str,
media_sample_rate_hz: int,
media_encoding: str,
vocabulary_name: Optional[str] = None,
session_id: Optional[str] = None,
vocab_filter_method: Optional[str] = None,
vocab_filter_name: Optional[str] = None,
show_speaker_label: Optional[bool] = None,
enable_channel_identification: Optional[bool] = None,
number_of_channels: Optional[int] = None,
enable_partial_results_stabilization: Optional[bool] = None,
partial_results_stability: Optional[str] = None,
language_model_name: Optional[str] = None
```

ehsanonyx · Answer

Here's a sample Python script that demonstrates how to enable automatic language identification:

`import boto3
import json
import os

# Define the required parameters
language_options = ['en-US', 'es-US']  # Define the possible languages you want to identify
media_sample_rate_hz = 16000  # Common sample rate for streaming audio
media_encoding = 'pcm'  # Encoding type for the audio stream

# Create a client for the Transcribe service
transcribe_client = boto3.client('transcribe')

# Start the transcription stream with automatic language identification enabled
response = transcribe_client.start_stream_transcription(
    LanguageCode='auto',
    MediaSampleRateHertz=media_sample_rate_hz,
    MediaEncoding=media_encoding,
    IdentifyLanguage=True,  # Enable automatic language identification
    LanguageOptions=language_options,  # Provide a list of potential languages
    VocabularyName=None,
    SessionId='my-session-id',
    EnablePartialResultsStabilization=True,
    PartialResultsStability='medium',
    ShowSpeakerLabel=False,
    EnableChannelIdentification=False,
    NumberOfChannels=1,
    VocabularyFilterMethod='mask',
    VocabularyFilterName=None,
    LanguageModelName=None
)

# Print the response to check the status of the transcription stream
print(json.dumps(response, indent=4))
`

How to Use the SDK to Enable Automatic Language Identification in AWS Transcribe Streaming ?

Define the required parameters

Create a client for the Transcribe service

Start the transcription stream with automatic language identification enabled

Print the response to check the status of the transcription stream

Relevant content