Using a model for speech recognition

You use the model parameter of a speech recognition request to indicate the model that is to be used with the request. You can specify a large speech model, previous- or next-generation model with the parameter.

For more information about the models that are available for speech recognition, see

Specify a previous-generation model example

The following example HTTP request uses the previous-generation model en-US_NarrowbandModel for speech recognition:

IBM Cloud

curl -X POST -u "apikey:{apikey}" \
--header "Content-Type: audio/flac" \
--data-binary @{path}audio-file.flac \
"{url}/v1/recognize?model=en-US_NarrowbandModel"

IBM Cloud Pak for Data

curl -X POST \
--header "Authorization: Bearer {token}" \
--header "Content-Type: audio/flac" \
--data-binary @{path}audio-file.flac \
"{url}/v1/recognize?model=en-US_NarrowbandModel"

Specify a next-generation model example

The following example HTTP request uses the next-generation en-US_Telephony model for speech recognition:

IBM Cloud

curl -X POST -u "apikey:{apikey}" \
--header "Content-Type: audio/wav" \
--data-binary @{path}audio-file.wav \
"{url}/v1/recognize?model=en-US_Telephony"

IBM Cloud Pak for Data

curl -X POST \
--header "Authorization: Bearer {token}" \
--header "Content-Type: audio/wav" \
--data-binary @{path}audio-file.wav \
"{url}/v1/recognize?model=en-US_Telephony"

Specify a large speech model example

The following example HTTP request uses the large speech model en-US for speech recognition:

IBM Cloud

curl -X POST -u "apikey:{apikey}" \
--header "Content-Type: audio/wav" \
--data-binary @{path}audio-file.wav \
"{url}/v1/recognize?model=en-US"

IBM Cloud Pak for Data

curl -X POST \
--header "Authorization: Bearer {token}" \
--header "Content-Type: audio/wav" \
--data-binary @{path}audio-file.wav \
"{url}/v1/recognize?model=en-US"

Using the default model

If you omit the model parameter from a speech recognition request, the service uses the US English en-US_BroadbandModel by default. This default applies to all speech recognition requests.

IBM Cloud Pak for Data If you do not install the en-US_BroadbandModel, it cannot serve as the default model. In this case, you must either

Use the model parameter to pass the model that is to be used with each request.
Specify a new default model for your installation of Speech to Text for IBM Cloud Pak for Data by using the defaultSTTModel property in the Speech services custom resource. For more information, see Installing Watson Speech to Text.