Managing audio resources

Acoustic model customization is available only for previous-generation models. It is not available for next-generation and large speech models.

The customization interface includes the POST /v1/acoustic_customizations/{customization_id}/audio/{audio_name} method, which is used to add an audio resource to a custom acoustic model. For more information, see Add audio to the custom acoustic model. The interface also includes the following methods for listing and deleting audio resources for a custom acoustic model.

Listing audio resources for a custom acoustic model

The customization interface provides two methods for listing information about the audio resources of a custom acoustic model:

The GET /v1/acoustic_customizations/{customization_id}/audio method lists information about all audio resources that have been added to a custom model.
The GET /v1/acoustic_customizations/{customization_id}/audio/{audio_name} method lists information about a specified audio resource for a custom model.

Both methods return the name of the audio resource plus the following additional information:

duration is the length in seconds of the audio. For an archive file, the figure represents the accumulated duration of all of the audio files that are contained in the archive.
details identifies the type of the resource: audio or archive. (The type is undetermined if the service cannot validate the resource, possibly because the user mistakenly passed a file that does not contain audio.) For an audio file, the details include the codec and frequency of the audio. For an archive file, they include its compression type.

Additionally, listing all audio resources for a model returns the total_minutes_of_audio summed over all of the valid audio resources for the model. You can use this value to determine whether the custom model has enough or too much audio to begin training.

The methods also list the status of the audio data. The status is important for checking the service's analysis of audio files in response to a request to add them to a custom model:

ok indicates that the service has successfully analyzed the audio data. The data can be used to train the custom model.
being_processed indicates that the service is still analyzing the audio data. The service cannot accept requests to add new audio or to train the custom model until its analysis is complete.
invalid indicates that the audio data is not valid for training the model (possibly because it has the wrong format or sampling rate, or because it is corrupted).

List all audio resources example

The following example lists all audio resources for the custom acoustic model with the specified customization ID:

IBM Cloud

curl -X GET -u "apikey:{apikey}" \
"{url}/v1/acoustic_customizations/{customization_id}/audio"

IBM Cloud Pak for Data

curl -X GET \
--header "Authorization: Bearer {token}" \
"{url}/v1/acoustic_customizations/{customization_id}/audio"

The acoustic model has three audio resources. The service has successfully analyzed audio1 and audio2; it is still analyzing audio3.

{
  "total_minutes_of_audio": 11.45,
  "audio": [
    {
      "duration": 131,
      "name": "audio1",
      "details": {
        "codec": "pcm_s16le",
        "type": "audio",
        "frequency": 22050
      }
      "status": "ok"
    },
    {
      "duration": 556,
      "name": "audio2",
      "details": {
        "type": "archive",
        "compression": "zip"
      },
      "status": "ok"
    },
    {
      "duration": 0,
      "name": "audio3",
      "details": {},
      "status": "being_processed"
    }
  ]
}

List an audio-type resource example

The following example returns information about the audio-type resource named audio1:

IBM Cloud

curl -X GET -u "apikey:{apikey}" \
"{url}/v1/acoustic_customizations/{customization_id}/audio/audio1"

IBM Cloud Pak for Data

curl -X GET \
--header "Authorization: Bearer {token}" \
"{url}/v1/acoustic_customizations/{customization_id}/audio/audio1"

The resource is 131 seconds long and is encoded with the pcm_s16le codec. It was successfully added to the model.

{
  "duration": 131,
  "name": "audio1",
  "details": {
    "codec": "pcm_s16le",
    "type": "audio",
    "frequency": 22050
  }
  "status": "ok"
}

List an archive-type resource example

The following example returns information about the archive-type resource named audio2:

IBM Cloud

curl -X GET -u "apikey:{apikey}" \
"{url}/v1/acoustic_customizations/{customization_id}/audio/audio2"

IBM Cloud Pak for Data

curl -X GET \
--header "Authorization: Bearer {token}" \
"{url}/v1/acoustic_customizations/{customization_id}/audio/audio2"

The resource is a .zip file that contains more than nine minutes of audio. It too was successfully added to the model. As the example shows, querying information about an archive-type resource also provides information about the files that it contains.

{
  "container": {
    "duration": 556,
    "name": "audio2",
    "details": {
      "type": "archive",
      "compression": "zip"
    },
    "status": "ok"
  },
  "audio": [
    {
      "duration": 121,
      "name": "audio-file1.wav",
      "details": {
        "codec": "pcm_s16le",
        "type": "audio",
        "frequency": 16000
      },
      "status": "ok"
    },
    {
      "duration": 133,
      "name": "audio-file2.wav",
      "details": {
        "codec": "pcm_s16le",
        "type": "audio",
        "frequency": 16000
      },
      "status": "ok"
    },
    . . .
  ]
}

Deleting an audio resource from a custom acoustic model

Use the DELETE /v1/acoustic_customizations/{customization_id}/audio/{audio_name} method to remove an existing audio resource from a custom acoustic model. When you delete an archive-type audio resource, the service removes the entire archive of files. The service does not allow deletion of individual files from an archive resource.

Removing an audio resource does not affect the custom model until you train the model on its updated data by using the POST /v1/acoustic_customizations/{customization_id}/train method. If you successfully trained the model on the resource, until you retrain the model, the existing audio data continues to be used for speech recognition.

Delete an audio resource example

The following method deletes the audio resource that is named audio3 from the custom model with the specified customization ID:

IBM Cloud

curl -X DELETE -u "apikey:{apikey}" \
"{url}/v1/acoustic_customizations/{customization_id}/audio/audio3"

IBM Cloud Pak for Data

curl -X DELETE \
--header "Authorization: Bearer {token}" \
"{url}/v1/acoustic_customizations/{customization_id}/audio/audio3"