[Feature]: Mix extension audio events into offline/native export audio

[crbender]

• I have searched the existing issues

Is your feature request related to a problem?

Yes. Even with captured events, extension sounds still need to be decoded/scheduled/mixed in the offline and native export audio paths.

Describe the solution you'd like

Mix captured extension audio events into final exported audio timeline.

This includes:

1. Decoding extension asset files
2. Scheduling by event timeMs
3. Applying per-event volume
4. Mixing with existing sources (embedded, fallback, audio regions) without regressions

Describe alternatives you've considered

Injecting post-export audio overlay in a separate pass. Rejected for complexity and sync risk.

Additional context

This is the core implementation slice that makes exported output match preview intent.

Acceptance criteria

1. Click events produce audible output in exported MP4 at expected times.
2. Dense click sequences preserve order and do not drop unexpectedly.
3. Existing source audio behavior remains correct.
4. Integration tests verify audible signal near expected click timestamps.

[crbender]

Tracking under parent issue #572 for sequencing and rollout context.