The Wwise External Source is a source plug-in that allows you to resolve a Voice object with an audio file at runtime. This can be very useful to manage a large amount of dialogue lines that would otherwise require a great deal of overhead, including the creation of Voice objects and Events for each, as well as the time it would take to properly divide and manage these objects and Events into SoundBanks. It can also be very useful if the dialogue lines are already managed through another system such as an AI-driven speech generator.
Depending on how you manage the dialogue in your project, you may also be able to save some runtime memory in the Default Memory pool because the Wwise External Source allows you to play dialogue without having to load many voice assets into memory at any given time.
The Wwise External Source works as follows:
A Sound-Voice 'template' is created in Wwise using the External Source plug-in. This template represents a series of audio files with common properties.
The Wwise External Source can take full advantage of the power and flexibility of the project hierarchy, by placing it within containers, actor-mixers, applying states, RTPCs, effects, and so on.
A play Event is created that calls the external source.
The location and Conversion Settings of the pool of external audio assets that can be used by the Wwise External Source are defined in the External Sources List file. This file is a very simple XML file that contains the location of the external audio assets that need to be converted along with the Conversion Settings you want to use. The location of this file is defined in the Project Settings dialog box in Wwise.
At runtime, the game calls the External Source and then associates the template with one of the external audio files. The actual audio file that is played is left entirely up to the programmer. It is important to note that the management of the source audio files is done external to the Wwise sound engine. While this involves more work, it also gives you more flexibility.
Most of the work related to this feature is performed by the audio programmer within the SDK. For more information, refer to the Wwise SDK documentation.