Fixing Dual Output vertical audio

[aleksandr-voitenko]

Description

This PR changes advanced streaming/recording/replay-buffer outputs to create their own per-output audio encoders from user-visible track configuration.

This PR fixes audio issues for the Vertical canvas when using dual output via Factory API.

Motivation and Context

The main behavioral goal is to allow multiple advanced outputs, such as horizontal and vertical streams or future simultaneous horizontal/vertical recordings, to use the same user audio track without relying on a "hidden" track 2 in desktop or sharing one global OBS audio encoder object.

Changes

1. AudioTrack no longer owns an OBS encoder.

It now represents only user-visible advanced track configuration:

• bitrate
• name

2. AdvancedStreaming has its own audio encoder
   When starting, it validates the configured track, then creates or reuses a private encoder for that output. This means two advanced streams can use the same user-level audio track without sharing the same native encoder object.

The destructor now releases the streaming-owned audio encoder reference after deleting the output. If the encoder is unexpectedly still active, it logs a warning before releasing the owner reference.

3. The same is applied to AdvancedRecording and AdvancedReplayBuffer.

How Has This Been Tested?

Manually, Windows only.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• Tweak (non-breaking change to improve existing functionality)