Missing support for `input_audio_noise_reduction` in session config

[HVbajoria]

Hi team,

The latest Azure OpenAI real-time audio models support configurable noise reduction, using the input_audio_noise_reduction field with options like near_field and far_field.

Reference:
📄 https://learn.microsoft.com/en-us/azure/ai-services/openai/realtime-audio-reference#realtimeaudioinputaudionoisereductionsettings

Currently, the SDK doesn't expose this field in the session config. This limits transcription quality in noisy or variable microphone environments.

✅ I’ve added support for this in a PR here:
🔗 #135

Example usage:

input_audio_noise_reduction: {
  type: "near_field"
}

This allows users to leverage the platform's built-in noise cancellation aligned with model expectations.

Would appreciate a review and merge — thank you!