Developing Applications

The CPQD ASR can be used in a variety of applications, such as

  • Call center automation

  • Speech controlled devices

  • Message dictation

  • Mobile app speech interfaces

Before you can develop your application, you need access to an ASR server. To run a CPQD ASR in your company’s environment, contact CPQD and request the installation packages and licenses. You will need to provide the following information:

What is the desired language?

A speech recognition model trained for Brazilian Portuguese cannot be used for any other language. Check to see if there are ASR models for your desired language.

What is the audio sample rate?

If the application is for phone usage, input audio has a sample rate of 8 kHz. This information is needed to install the correct model. A model for 16 kHz audio will not work in applications involving phone calls.

Will you use free speech?

Speech recognition using grammars is included in the default APQD ASR installation. If your application needs to transcribe what is said, you need to require the free speech model installer.

Will you use MRCP integration?

The way you will integrate with the CPQD ASR depends on the context of the application. In phone applications with IVR, the integration can be restricted to standards like MRCP. If this is your case, you must require the CPQD MRCP server installer.

Hardware and software requirements and guideline for installing the CPQD ASR can be found at Installation.

If you wish to run speech recognition tests before installing the CPQD ASR, contact CPQD and request access credentials to the ASR service, available on the WEB.

The CPQD site also has an ASR demo page where you can run simple tests, using free speech recognition.

Guidelines to develop your application can be found at User Guide.