MULTICHANNEL MAC INTEGRATION

A short gameplay run-through captured as discrete 5.1 PCM, recorded from the multichannel build. This requires a surround hardware setup with a 5.1 capable player such as QuickTime or VLC.

The backend now queries the device's preferred channel layout, the same speaker configuration set in Audio MIDI Setup, and translates each channel label into the engine's channel map. 7.1 layouts resolve sides and rears correctly, Quad falls out of the same parse, and any layout with two or fewer labeled channels takes the original stereo path. Although not advised, Speaker configurations are read on each stream open (Play in PIE), so changes can be applied mid-session without a restart.

The engine renders a compact interleaved stream of just the labeled channels and reports the real channel list to the mixer, which activates Unreal's existing surround panner with zero mixer changes. A CoreAudio channel map then scatters each rendered channel to its physical device output, verified using an 88-output aggregate (Universal Audio Apollo x8P, x16, and x8PG2). Spatialized sources pan across every speaker, including true center and LFE routing.

The stock sample-rate query used a wrong property selector which failed on every launch, so the engine always rendered 48 kHz with the AudioUnit resampling with a log warning. This patch allows the Mac platform settings to set a fixed sample rate, or a value of zero to follow the output device's native rate. Buffer size, buffers to enqueue, max channels, and async source workers are now honored, moving per-voice DSP off the render callback and bringing macOS to desktop parity.

Verified audibly per speaker on a calibrated 5.1 monitoring rig across stereo, Quad, 5.1, and 7.1, with the full sample-rate matrix at 48 and 96 kHz, pinned and device-follow, plus mid-session speaker reconfiguration. Render cost measures about 1.2 ms against a 21.3 ms callback budget at six channels.