Configure Requests#

Transcription configuration options specify how transcription is performed. This includes the model used, features used, and additional parameters used for specific features.

The TranscriptionConfig data structure encapsulates all available transcription configuration fields. Various TranscriptionConfig fields will be encountered in further pages, but the complete list of fields is documented in the gRPC API reference: TranscriptionConfig.

The example below shows how to set a TranscriptionConfig field when using one of the Soniox client libraries, in this case model.

In the Soniox Python library, TranscriptionConfig fields are specified as keyword arguments to any of the available transcription functions.

result = transcribe_file_short(
    "../test_data/test_audio.flac",
    client,
    model="en_v2",
)

In the Soniox Node library, TranscriptionConfig fields are specified by passing a config object to any of the available transcription functions.

const result = await speechClient.transcribeFileShort(
    "../test_data/test_audio.flac",
    {
        model: "en_v2",
    }
);

In the Soniox C# library, TranscriptionConfig fields are specified by passing a TranscriptionConfig object to any of the available transcription functions. Note that field names are adapted to the C# style.

var completeResult = await client.TranscribeFileShort(
    "../../test_data/test_audio.flac",
    new TranscriptionConfig {
        Model = "en_v2",
    });

Any parameter which is not set explicitly will an effective default value determined by its data type. This is: false for bool, 0 for numeric types, empty for string, and no elements for repeated (sequence) fields.