Introduction

Description

The Voice Capture SDK is the client side library to be used with the Voice Liveness Server product for determining the liveness of a subject. It is used to capture audio samples and package them according to the Voice Liveness Server specifications. The Voice Liveness Server will make a determination for whether or not the subject is live. The library does not directly communicate with any servers; it provides a package for the application to deliver to the Voice Liveness Server. It is up to the application developer to implement any server communication and security.

Audio samples are captured based on the selection of a Knomi Workflow. Workflows are a set of pre-defined rules and parameters that determine the acceptable behaviors and traits of a sample in order to optimize results obtained from the Voice Liveness Server. Different Workflows will allow for different behaviors and may utilize different Liveness capabilities on the back-end.

The library is able to detect the difference between silence, noise, and voice activity. It will wait until it has captured the specified amount of voice data, or until a specified timeout is reached. This ensures that you are able to capture the voice data you need.

The library manages all requirements for capture including operating the microphone. Once a long enough voice sample has been captured, the library will automatically stop microphone capture and create a package for use with the Voice Liveness Server.

Design Features

Voice Capture SDK

• Light-weight on-device size.
• Easy to use API.
• No direct UI activity to allow for a fully customizable user experience.
• Handles operation of device hardware.
• Simplistic choice of Workflows to optimize data collection for back-end analysis.
• No direct server interaction. Applications have full control over when and how they interact with the back-end service.