Add ability to configure Voice for realtime session.

[cecolson]

Feature Request

Is your feature request related to a problem? Please describe.

No.

Describe the solution you'd like

I would like the ability to change the default voice that is used when establishing a realtime session. The current set voice is "alloy" but the openAI api provides the current voice options as alloy, ash, ballad, coral, echo, sage, shimmer, and verse. Please allow us to set this when configuring the session.

Describe alternatives you've considered

Writing a script that just connects to an OpenAI session through websockets.

Additional context

None.

[StephenHodgson]

Hi @cecolson, you can change the default voice options in the constructor of the session configuration options. Check out my realtime sample behavior.

com.openai.unity/OpenAI/Packages/com.openai.unity/Samples~/Realtime/RealtimeBehaviour.cs

Lines 108 to 114 in a81f086

	session = await openAI.RealtimeEndpoint.CreateSessionAsync(
	new SessionConfiguration(
	model: Model.GPT4oRealtime,
	voice: voice,
	instructions: systemPrompt,
	tools: tools),
	destroyCancellationToken);