fix: Hold the 2-way audio backchannel open for the shortest possible time (#2697)

The card claimed a camera's ONVIF audio backchannel in two places that
had
nothing to do with a call: the capability probe at camera init
(`&microphone`),
and a pre-armed `sendonly` audio transceiver on every live WebRTC offer.
Merely
looking at a dashboard occupied the camera's speaker line. Outbound
audio now
travels on its own audio-only WebRTC connection, opened when a call is
answered
and closed when it ends.

- The backchannel is claimed only for the duration of a call. Idle
viewing
  claims nothing.
- Two-way audio now works in `mse`, `mp4` and `mjpeg` modes (note: the
outbound
  audio still traverses WebRTC).
- No renegotiation and no video blink at call start or end.
- A call that cannot carry audio now reports it and ends, instead of
showing a
  live microphone that goes nowhere.
- `live.microphone.always_connected` is now purely about the browser
microphone
  permission prompt.
- Call setup measured at 66ms (LAN) and ~260ms (cellular) for ICE and
DTLS, plus
  ~300ms for `go2rtc` to open an RTSP backchannel.

Verified against a live Frigate + `go2rtc` instance, and by unit tests
at 100%
coverage.

 - Closes #2691
 - Closes #2039
 - Closes #2178

Ref #2299 -- the probe no longer opens a backchannel, but it still runs
per
camera on every load and reconnect, and still dials the camera on the
direct-`go2rtc` path. Caching remains to be done.

Ref AlexxIT/go2rtc#1860 -- once a call has opened a backchannel,
`go2rtc` keeps
that media set up on the camera's RTSP session for the life of the
producer.

Diagnoses #2678

src/camera-manager/camera.ts

@@ -10,6 +10,7 @@ import {
   resolveProxyConfig,
   type EnabledProxyConfig,
 } from '../config/schema/common/proxy';
+import { getGo2RTCMetadataEndpoint, getGo2RTCStreamEndpoint } from '../go2rtc/endpoint';
 import { computeDomain } from '../ha/compute-domain';
 import { matchesEventContext, matchesEventData } from '../ha/event-match';
 import { getTriggerEventType } from '../ha/get-trigger-event-type';
@@ -32,10 +33,6 @@ import type {
   CameraProxyConfig,
 } from './types';
 import { getCameraEntityFromConfig } from './utils/camera-entity-from-config';
-import {
-  getGo2RTCMetadataEndpoint,
-  getGo2RTCStreamEndpoint,
-} from './utils/go2rtc/endpoint';
 import { getConfiguredPTZAction } from './utils/ptz';
 
 interface CapabilityOptions {

src/camera-manager/frigate/camera.ts

@@ -3,6 +3,10 @@ import { format } from 'date-fns';
 import type { ActionsExecutor } from '../../card-controller/actions/types';
 import type { PTZAction, PTZActionPhase } from '../../config/schema/actions/custom/ptz';
 import type { CameraConfig } from '../../config/schema/cameras';
+import {
+  getGo2RTCMetadataEndpoint,
+  getGo2RTCStreamEndpoint,
+} from '../../go2rtc/endpoint';
 import type { Entity, EntityRegistryManager } from '../../ha/registry/entity/types';
 import type { HomeAssistant } from '../../ha/types';
 import {
@@ -16,10 +20,6 @@ import { Camera, type CameraInitializationOptions } from '../camera';
 import { CameraNoEntityError } from '../error';
 import type { CameraEndpoints, CameraEndpointsContext } from '../types';
 import { getCameraEntityFromConfig } from '../utils/camera-entity-from-config';
-import {
-  getGo2RTCMetadataEndpoint,
-  getGo2RTCStreamEndpoint,
-} from '../utils/go2rtc/endpoint';
 import { getPTZCapabilitiesFromCameraConfig, mergePTZCapabilities } from '../utils/ptz';
 import { getPTZInfo } from './requests';
 import {

src/camera-manager/utils/go2rtc/audio.ts

@@ -1,76 +0,0 @@
-import type { EnabledProxyConfig } from '../../../config/schema/common/proxy';
-import { homeAssistantSignAndFetch } from '../../../ha/fetch';
-import type { HomeAssistant } from '../../../ha/types';
-import { createProxiedEndpointIfNecessary } from '../../../ha/web-proxy';
-import type { Endpoint } from '../../../types';
-import { errorToConsole } from '../../../utils/basic';
-import { go2RTCStreamInfoSchema, type Go2RTCStreamInfo } from './types';
-
-const getGo2RTCStreamMetadata = async (
-  hass: HomeAssistant,
-  endpoint: Endpoint,
-  timeoutSeconds: number,
-): Promise<Go2RTCStreamInfo | null> => {
-  try {
-    return await homeAssistantSignAndFetch(hass, endpoint, go2RTCStreamInfoSchema, {
-      timeoutSeconds,
-    });
-  } catch (e) {
-    errorToConsole(e);
-    return null;
-  }
-};
-
-const streamSupports2WayAudio = (streamInfo: Go2RTCStreamInfo | null): boolean => {
-  if (!streamInfo?.producers) {
-    return false;
-  }
-  return streamInfo.producers.some(
-    (producer) =>
-      producer.medias?.some(
-        (media) =>
-          media.includes('audio') &&
-          (media.includes('sendonly') || media.includes('sendrecv')),
-      ) ?? false,
-  );
-};
-
-/**
- * Fetch go2rtc metadata and determine if the stream supports 2-way audio.
- * Handles proxy transformation if proxy config requires it.
- * Returns false if the endpoint is not available or fetch fails.
- *
- * Note: Caller is responsible for checking if live_provider is 'go2rtc' before calling.
- *
- * @param hass Home Assistant instance.
- * @param go2rtcMetadataEndpoint The go2rtc metadata endpoint.
- * @param proxyConfig The resolved proxy configuration for live streams.
- * @returns True if supports 2-way audio, false otherwise.
- */
-export const supports2WayAudio = async (
-  hass: HomeAssistant,
-  metadataFetchTimeoutSeconds: number,
-  go2rtcMetadataEndpoint?: Endpoint | null,
-  proxyConfig?: EnabledProxyConfig,
-): Promise<boolean> => {
-  if (!go2rtcMetadataEndpoint) {
-    return false;
-  }
-
-  const endpoint = await createProxiedEndpointIfNecessary(
-    hass,
-    go2rtcMetadataEndpoint,
-    proxyConfig,
-    { openLimit: 1 },
-  );
-  if (!endpoint) {
-    return false;
-  }
-
-  const streamInfo = await getGo2RTCStreamMetadata(
-    hass,
-    endpoint,
-    metadataFetchTimeoutSeconds,
-  );
-  return streamSupports2WayAudio(streamInfo);
-};

src/camera-manager/utils/go2rtc/endpoint.ts

@@ -1,51 +0,0 @@
-import type { CameraConfig } from '../../../config/schema/cameras';
-import type { Endpoint } from '../../../types';
-
-interface EndpointOptions {
-  url?: string;
-  stream?: string;
-}
-
-const buildGo2RTCEndpoint = (
-  cameraConfig: CameraConfig,
-  pathBuilder: (url: string, stream: string) => string,
-  options?: EndpointOptions,
-): Endpoint | null => {
-  const url = options?.url ?? cameraConfig.go2rtc?.url;
-  const stream = options?.stream ?? cameraConfig.go2rtc?.stream;
-
-  if (!url || !stream) {
-    return null;
-  }
-
-  const endpoint = pathBuilder(url, stream);
-  return {
-    endpoint,
-    // Only sign the endpoint if it's local to HA.
-    sign: endpoint.startsWith('/'),
-  };
-};
-
-export const getGo2RTCStreamEndpoint = (
-  cameraConfig: CameraConfig,
-  options?: EndpointOptions,
-): Endpoint | null => {
-  return buildGo2RTCEndpoint(
-    cameraConfig,
-    (url, stream) => `${url}/api/ws?src=${stream}`,
-    options,
-  );
-};
-
-export const getGo2RTCMetadataEndpoint = (
-  cameraConfig: CameraConfig,
-  options?: EndpointOptions,
-): Endpoint | null => {
-  return buildGo2RTCEndpoint(
-    cameraConfig,
-    // Use probe parameters to trigger active stream detection.
-    // Without these, go2rtc only returns static config without producer medias.
-    (url, stream) => `${url}/api/streams?src=${stream}&video=all&audio=all&microphone`,
-    options,
-  );
-};

src/camera-manager/utils/go2rtc/types.ts

@@ -1,16 +0,0 @@
-import { z } from 'zod';
-
-const go2RTCProducerSchema = z.object({
-  medias: z.array(z.string()).optional(),
-});
-
-/**
- * Zod schema for Go2RTC stream information. Schema only covers the minimum
- * required by the card.
- * Response from `/api/streams?src=${stream}&video=all&audio=all&microphone`
- */
-export const go2RTCStreamInfoSchema = z.object({
-  producers: z.array(go2RTCProducerSchema).optional(),
-});
-
-export type Go2RTCStreamInfo = z.infer<typeof go2RTCStreamInfoSchema>;