In the 10.7 release of HPE IDOL Speech Server, you could use acoustic adaptation to adapt the Gaussian Mixture Model (GMM) acoustic models to match an audio domain. To improve speech-to-text accuracy, HPE IDOL Speech Server now includes Deep Neural Network (DNN) acoustic modeling. DNNs are not currently adaptable, but typically outperform even adapted GMM acoustic models. As a result, HPE does not generally recommend acoustic adaptation. However, in certain scenarios (for example, in cases where the language packs do not have a DNN, or where you are working with a very specific domain and believe that DNN recognition could be improved upon), acoustic adaptation can still be useful. In the latter case, you must suppress the DNN model at run time to use the newly adapted acoustic model files.
Use the following instructions to perform acoustic adaptation.
This section describes how to adapt the acoustic models provided in the HPE IDOL Speech Server language packs.
Transcription Data Requirements
Prepare the Transcription Data
Present Adaptation Data to Speech Server
Acoustic Adaptation Diagnostics
Finalize the Adapted Acoustic Model
|