fix(viewer): default x86 audio fallback to ALSA 'default'

[Xzzz]

On x86, get_alsa_audio_device() returns sysdefault:CARD=HID as the fallback, which matches no real ALSA card on Intel/AMD/Nvidia HDA chipsets. This bypasses the container's ~/.asoundrc bind mount and leaves operators with no way to route audio.

The ARM64 branch a few lines above already returns 'default' for the same reason — heterogeneous SoCs, no portable card name — and lets operators override via ~/.asoundrc. This PR mirrors that approach for x86.

Reproduction

• x86 host (Intel HDA PCH card with HDMI output to a TV)
• Debian minimal install, Anthias v2026.05.0+
• mpv launched by the viewer logs --audio-device=alsa/sysdefault:CARD=HID → no audio
• /proc/asound/cards shows only card 0 (HDA Intel PCH); no HID card exists

Verification

After this patch, ~/.asoundrc (e.g. routing default to hw:0,3 for the HDMI sink) is honored, and audio plays correctly on the TV.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[vpetersson]

Validated against an x86 testbed (Intel HDA PCH + a USB Audio+HID dongle) — change is structurally correct, but the PR ships a failing unit test.

Validation

• /proc/asound/cards on the testbed shows card 0 = HDA Intel PCH, card 1 = HID (CSCTEK USB Audio+HID dongle). Minor nit on the PR body: a HID card does exist on this box — it's just a USB peripheral, not the HDMI sink. Softening the wording to "no portable card name across Intel/AMD/Nvidia HDA chipsets" (matching the ARM64 comment this mirrors) would be more accurate; the symptom you describe (audio routed away from HDMI) still applies.
• C++ side handles default cleanly: src/anthias_webview/src/videoview.cpp:380 short-circuits when the extracted card name is literally default and uses QMediaDevices::defaultAudioOutput(). No D-Bus / CARD= parsing regression from dropping the CARD= token.
• ~/.asoundrc bind mount is in place (docker-compose.yml.tmpl:82), so the override hook the PR relies on really exists.

Blocker — CI will fail

tests/test_media_player.py:341 still asserts the old value:

def test_hdmi_on_x86_falls_back_to_hid(alsa_settings: Any) -> None:
    ...
    assert get_alsa_audio_device() == 'sysdefault:CARD=HID'

Running it against this branch:

AssertionError: assert 'default' == 'sysdefault:CARD=HID'

Please rename to test_hdmi_on_x86_falls_back_to_default and assert 'default'. (The status rollup only shows SonarCloud right now — the pytest job will catch this once it runs.)

Nice-to-have (non-blocking)

The ARM64 branch logs the chosen ALSA card once per process via _log_arm64_alsa_default_once() so silent-HDMI reports carry breadcrumbs. The x86 path now has the same "debug via .asoundrc" failure mode — symmetric once-per-process logging here would help future operators. Fine to defer to a follow-up.

Risk

Low. For devices without an .asoundrc, behavior changes from "fall back to a CARD=HID device that probably isn't the intended sink" to "fall back to Qt's platform default" — strictly better, and overridable as documented.