Line-of-Sight audio Occlusion and Obstruction

[phevosccp]

Description

This PR implements basic occlusion/obstruction using LPF and implementing SetObjectObstructionAndOcclusion() from the Wwise API.

In addition to that it manages emitterIDs that are occluded/passthrough and handle fade in/out internals.
This implementation differs from the Acoustics API in the sense that it completely relies on Carbon's (destiny) line-of-sight ray tracing checks to know what is occluded and what not.

The whole path end to end in short is the following: Destiny → Python → trinity → CarbonAudio → Wwise.

Technical details:

• New class AudObstructionOcclusion : Owns all per emitter obstruction/occlusion and talks to the wwise API. We store the emitter IDs and states in a unordered map and handle their states, in addition we also handle fade/in out intervals
• The class has no concept of the actual emitter states so it reads that from AudManager
• It's also important to note that currently we disable basic occlusion/obstruction when acoustics are ON to avoid attenuation stacking (e.g -12dB occlusion + -12 dB tranmission = -24dB reduction which is unexpected ). This will potentially change with the introduction of trigger volumes and we will keep both systems alive.

Lifecycle

• Update loop : AudManager::Process(), once per tick, before RenderAudio()): advances all fades using a clock delta and sends only changed values to Wwise.
• Emitter validation : values are only accepted/tracked for emitters that actually exist, see WithCallbackGameObject
• Culling : emitters that are currently culled are skipped but keep fading
• Cleanup : entries that fade fully back to unblocked are dropped to keep the map small , see UnregisterGameObject()

Testing

Tested and implemented for EVE Frontier.

[ccpgames-carbon]

All committers have signed the CLA.

[ccp-zoetrope]

This pull request looks almost ready!

The only blocker is that there are no tests covering this new feature. The logic that needs to have some testing around it is the interpolation of occlusion values to create a fade as it seems the most brittle to future contributors.

I also made a few small comments about method name changes which, if you agree, you can go ahead with but it is not a blocker.

[ccp-zoetrope]