fix(google-meet): clamp audio buffer config

2026-05-06 05:20:43 +00:00 · 2026-05-04 03:38:26 +01:00
parent 571d75aab3
commit c956946b26
6 changed files with 135 additions and 17 deletions
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -43,6 +43,7 @@ Docs: https://docs.openclaw.ai

 - Diagnostics: keep webhook/message OTEL attributes and Prometheus delivery labels low-cardinality and omit raw chat/message IDs from spans, so progress-draft and message-tool modes do not leak high-cardinality messaging identifiers.
 - Google Meet: stop advertising legacy `mode: "realtime"` to agents and config UIs, while keeping it as a hidden compatibility alias for `mode: "agent"`, so new joins use the STT -> OpenClaw agent -> TTS path instead of selecting the direct realtime voice fallback.
+- Google Meet: add `chrome.audioBufferBytes` for generated command-pair SoX audio commands and lower the default buffer from SoX's 8192 bytes to 4096 bytes to reduce Chrome talk-back latency.
 - Telegram: render shared interactive reply buttons in reply delivery so plugin approval messages show inline keyboards. (#76238) Thanks @keshavbotagent.
 - Agents/cli-runner: drop a saved `claude-cli` resume sessionId at preparation time when its on-disk transcript no longer exists in `~/.claude/projects/`, so a stale binding from a half-installed `update.run` cannot trap follow-up runs (auto-reply / Telegram direct) in a `claude --resume` timeout loop; the run starts fresh and the new sessionId is written back through the existing post-run flow. (#77030; refs #77011) Thanks @openperf.
 - Release validation: install the cross-OS TypeScript harness through Windows-safe Node/npm shims so native Windows package checks reach the OpenClaw smoke suites instead of exiting before artifact capture. Thanks @vincentkoc.
--- a/docs/plugins/google-meet.md
+++ b/docs/plugins/google-meet.md
@@ -1016,6 +1016,10 @@ Defaults:
 - `chrome.audioFormat: "pcm16-24khz"`: command-pair audio format. Use
  `"g711-ulaw-8khz"` only for legacy/custom command pairs that still emit
  telephony audio.
+- `chrome.audioBufferBytes: 4096`: SoX processing buffer for generated Chrome
+  command-pair audio commands. This is half of SoX's default 8192-byte buffer,
+  reducing default pipe latency while leaving room to raise it on busy hosts.
+  Values below SoX's minimum are clamped to 17 bytes.
 - `chrome.audioInputCommand`: SoX command reading from CoreAudio `BlackHole 2ch`
  and writing audio in `chrome.audioFormat`
 - `chrome.audioOutputCommand`: SoX command reading audio in `chrome.audioFormat`
@@ -1622,7 +1626,8 @@ Chrome talk-back modes need `BlackHole 2ch` plus either:
  bridge and pipes audio in `chrome.audioFormat` between those commands and the
  selected provider. Agent mode uses realtime transcription plus regular TTS;
  bidi mode uses the realtime voice provider. The default Chrome path is 24 kHz
-  PCM16; 8 kHz G.711 mu-law remains available for legacy command pairs.
+  PCM16 with `chrome.audioBufferBytes: 4096`; 8 kHz G.711 mu-law remains
+  available for legacy command pairs.
 - `chrome.audioBridgeCommand`: an external bridge command owns the whole local
  audio path and must exit after starting or validating its daemon. This is only
  valid for `bidi` because `agent` mode needs direct command-pair access for TTS.
--- a/extensions/google-meet/index.test.ts
+++ b/extensions/google-meet/index.test.ts
@@ -329,9 +329,12 @@ describe("google-meet plugin", () => {
        autoJoin: true,
        waitForInCallMs: 20000,
        audioFormat: "pcm16-24khz",
+        audioBufferBytes: 4096,
        audioInputCommand: [
          "sox",
          "-q",
+          "--buffer",
+          "4096",
          "-t",
          "coreaudio",
          "BlackHole 2ch",
@@ -351,6 +354,8 @@ describe("google-meet plugin", () => {
        audioOutputCommand: [
          "sox",
          "-q",
+          "--buffer",
+          "4096",
          "-t",
          "raw",
          "-r",
@@ -410,18 +415,21 @@ describe("google-meet plugin", () => {
    };

    expect(entry.configSchema.uiHints).toMatchObject({
+      "chrome.audioBufferBytes": expect.objectContaining({ advanced: true }),
      "chrome.bargeInInputCommand": expect.objectContaining({ advanced: true }),
      "chrome.bargeInRmsThreshold": expect.objectContaining({ advanced: true }),
      "chrome.bargeInPeakThreshold": expect.objectContaining({ advanced: true }),
      "chrome.bargeInCooldownMs": expect.objectContaining({ advanced: true }),
    });
    expect(manifest.uiHints).toMatchObject({
+      "chrome.audioBufferBytes": expect.objectContaining({ advanced: true }),
      "chrome.bargeInInputCommand": expect.objectContaining({ advanced: true }),
      "chrome.bargeInRmsThreshold": expect.objectContaining({ advanced: true }),
      "chrome.bargeInPeakThreshold": expect.objectContaining({ advanced: true }),
      "chrome.bargeInCooldownMs": expect.objectContaining({ advanced: true }),
    });
    expect(manifest.configSchema?.properties?.chrome?.properties).toMatchObject({
+      audioBufferBytes: expect.objectContaining({ type: "number", default: 4096 }),
      bargeInInputCommand: expect.objectContaining({
        type: "array",
        items: { type: "string" },
@@ -467,6 +475,47 @@ describe("google-meet plugin", () => {
    });
  });

+  it("lets generated Chrome audio commands use a configured SoX buffer", () => {
+    const config = resolveGoogleMeetConfig({ chrome: { audioBufferBytes: 2048 } });
+
+    expect(config.chrome.audioBufferBytes).toBe(2048);
+    expect(config.chrome.audioInputCommand).toEqual([
+      "sox",
+      "-q",
+      "--buffer",
+      "2048",
+      "-t",
+      "coreaudio",
+      "BlackHole 2ch",
+      "-t",
+      "raw",
+      "-r",
+      "24000",
+      "-c",
+      "1",
+      "-e",
+      "signed-integer",
+      "-b",
+      "16",
+      "-L",
+      "-",
+    ]);
+    expect(config.chrome.audioOutputCommand?.slice(0, 4)).toEqual([
+      "sox",
+      "-q",
+      "--buffer",
+      "2048",
+    ]);
+  });
+
+  it("clamps configured Chrome audio buffers above SoX's minimum", () => {
+    const config = resolveGoogleMeetConfig({ chrome: { audioBufferBytes: 1 } });
+
+    expect(config.chrome.audioBufferBytes).toBe(17);
+    expect(config.chrome.audioInputCommand?.slice(0, 4)).toEqual(["sox", "-q", "--buffer", "17"]);
+    expect(config.chrome.audioOutputCommand?.slice(0, 4)).toEqual(["sox", "-q", "--buffer", "17"]);
+  });
+
  it("uses env fallbacks for OAuth, preview, and default meeting values", () => {
    expect(
      resolveGoogleMeetConfigWithEnv(
--- a/extensions/google-meet/index.ts
+++ b/extensions/google-meet/index.ts
@@ -82,6 +82,11 @@ const googleMeetConfigSchema = {
      help: "Command-pair audio format. PCM16 24 kHz is the default Chrome/Meet path; G.711 mu-law 8 kHz remains available for legacy command pairs.",
      advanced: true,
    },
+    "chrome.audioBufferBytes": {
+      label: "Audio Buffer Bytes",
+      help: "SoX processing buffer for generated Chrome command-pair audio commands. Lower values reduce latency but may underrun on busy hosts.",
+      advanced: true,
+    },
    "chrome.audioInputCommand": {
      label: "Audio Input Command",
      help: "Command that writes meeting audio to stdout in chrome.audioFormat.",
--- a/extensions/google-meet/openclaw.plugin.json
+++ b/extensions/google-meet/openclaw.plugin.json
@@ -93,6 +93,11 @@
      "help": "Command-pair audio format. PCM16 24 kHz is the default Chrome/Meet path; G.711 mu-law 8 kHz remains available for legacy command pairs.",
      "advanced": true
    },
+    "chrome.audioBufferBytes": {
+      "label": "Audio Buffer Bytes",
+      "help": "SoX processing buffer for generated Chrome command-pair audio commands. Lower values reduce latency but may underrun on busy hosts.",
+      "advanced": true
+    },
    "chrome.audioBridgeCommand": {
      "label": "Audio Bridge Command",
      "advanced": true
@@ -272,11 +277,17 @@
            "enum": ["pcm16-24khz", "g711-ulaw-8khz"],
            "default": "pcm16-24khz"
          },
+          "audioBufferBytes": {
+            "type": "number",
+            "default": 4096
+          },
          "audioInputCommand": {
            "type": "array",
            "default": [
              "sox",
              "-q",
+              "--buffer",
+              "4096",
              "-t",
              "coreaudio",
              "BlackHole 2ch",
@@ -302,6 +313,8 @@
            "default": [
              "sox",
              "-q",
+              "--buffer",
+              "4096",
              "-t",
              "raw",
              "-r",
--- a/extensions/google-meet/src/config.ts
+++ b/extensions/google-meet/src/config.ts
@@ -28,6 +28,7 @@ export type GoogleMeetConfig = {
  chrome: {
    audioBackend: "blackhole-2ch";
    audioFormat: GoogleMeetChromeAudioFormat;
+    audioBufferBytes: number;
    launch: boolean;
    browserProfile?: string;
    guestName: string;
@@ -86,7 +87,15 @@ export type GoogleMeetConfig = {
  };
 };

-export const DEFAULT_GOOGLE_MEET_AUDIO_INPUT_COMMAND = [
+const SOX_DEFAULT_BUFFER_BYTES = 8192;
+const SOX_MIN_BUFFER_BYTES = 17;
+export const DEFAULT_GOOGLE_MEET_AUDIO_BUFFER_BYTES = SOX_DEFAULT_BUFFER_BYTES / 2;
+
+function withSoxBuffer(command: readonly string[], bufferBytes: number): string[] {
+  return [command[0] ?? "sox", "-q", "--buffer", String(bufferBytes), ...command.slice(2)];
+}
+
+const DEFAULT_GOOGLE_MEET_AUDIO_INPUT_COMMAND_BASE = [
  "sox",
  "-q",
  "-t",
@@ -106,7 +115,7 @@ export const DEFAULT_GOOGLE_MEET_AUDIO_INPUT_COMMAND = [
  "-",
 ] as const;

-export const DEFAULT_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND = [
+const DEFAULT_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND_BASE = [
  "sox",
  "-q",
  "-t",
@@ -126,7 +135,7 @@ export const DEFAULT_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND = [
  "BlackHole 2ch",
 ] as const;

-const LEGACY_GOOGLE_MEET_AUDIO_INPUT_COMMAND = [
+const LEGACY_GOOGLE_MEET_AUDIO_INPUT_COMMAND_BASE = [
  "rec",
  "-q",
  "-t",
@@ -142,7 +151,7 @@ const LEGACY_GOOGLE_MEET_AUDIO_INPUT_COMMAND = [
  "-",
 ] as const;

-const LEGACY_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND = [
+const LEGACY_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND_BASE = [
  "play",
  "-q",
  "-t",
@@ -158,6 +167,16 @@ const LEGACY_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND = [
  "-",
 ] as const;

+export const DEFAULT_GOOGLE_MEET_AUDIO_INPUT_COMMAND = withSoxBuffer(
+  DEFAULT_GOOGLE_MEET_AUDIO_INPUT_COMMAND_BASE,
+  DEFAULT_GOOGLE_MEET_AUDIO_BUFFER_BYTES,
+);
+
+export const DEFAULT_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND = withSoxBuffer(
+  DEFAULT_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND_BASE,
+  DEFAULT_GOOGLE_MEET_AUDIO_BUFFER_BYTES,
+);
+
 const DEFAULT_GOOGLE_MEET_CHROME_AUDIO_FORMAT: GoogleMeetChromeAudioFormat = "pcm16-24khz";
 const DEFAULT_GOOGLE_MEET_BARGE_IN_RMS_THRESHOLD = 650;
 const DEFAULT_GOOGLE_MEET_BARGE_IN_PEAK_THRESHOLD = 2500;
@@ -177,6 +196,7 @@ const DEFAULT_GOOGLE_MEET_CONFIG: GoogleMeetConfig = {
  chrome: {
    audioBackend: "blackhole-2ch",
    audioFormat: DEFAULT_GOOGLE_MEET_CHROME_AUDIO_FORMAT,
+    audioBufferBytes: DEFAULT_GOOGLE_MEET_AUDIO_BUFFER_BYTES,
    launch: true,
    guestName: "OpenClaw Agent",
    reuseExistingTab: true,
@@ -361,16 +381,36 @@ function resolveChromeAudioFormat(value: unknown): GoogleMeetChromeAudioFormat |
  }
 }

-function defaultAudioInputCommand(format: GoogleMeetChromeAudioFormat): readonly string[] {
-  return format === "g711-ulaw-8khz"
-    ? LEGACY_GOOGLE_MEET_AUDIO_INPUT_COMMAND
-    : DEFAULT_GOOGLE_MEET_AUDIO_INPUT_COMMAND;
+function resolveAudioBufferBytes(value: unknown, fallback: number): number {
+  const number = resolveNumber(value, fallback);
+  if (!Number.isFinite(number) || number <= 0) {
+    return fallback;
+  }
+  return Math.max(SOX_MIN_BUFFER_BYTES, Math.trunc(number));
 }

-function defaultAudioOutputCommand(format: GoogleMeetChromeAudioFormat): readonly string[] {
-  return format === "g711-ulaw-8khz"
-    ? LEGACY_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND
-    : DEFAULT_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND;
+function defaultAudioInputCommand(
+  format: GoogleMeetChromeAudioFormat,
+  bufferBytes: number,
+): string[] {
+  return withSoxBuffer(
+    format === "g711-ulaw-8khz"
+      ? LEGACY_GOOGLE_MEET_AUDIO_INPUT_COMMAND_BASE
+      : DEFAULT_GOOGLE_MEET_AUDIO_INPUT_COMMAND_BASE,
+    bufferBytes,
+  );
+}
+
+function defaultAudioOutputCommand(
+  format: GoogleMeetChromeAudioFormat,
+  bufferBytes: number,
+): string[] {
+  return withSoxBuffer(
+    format === "g711-ulaw-8khz"
+      ? LEGACY_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND_BASE
+      : DEFAULT_GOOGLE_MEET_AUDIO_OUTPUT_COMMAND_BASE,
+    bufferBytes,
+  );
 }

 export function resolveGoogleMeetConfig(input: unknown): GoogleMeetConfig {
@@ -392,6 +432,10 @@ export function resolveGoogleMeetConfigWithEnv(
  const audioFormat =
    resolveChromeAudioFormat(chrome.audioFormat) ??
    (hasCustomAudioCommand ? "g711-ulaw-8khz" : DEFAULT_GOOGLE_MEET_CONFIG.chrome.audioFormat);
+  const audioBufferBytes = resolveAudioBufferBytes(
+    chrome.audioBufferBytes,
+    DEFAULT_GOOGLE_MEET_CONFIG.chrome.audioBufferBytes,
+  );
  const chromeNode = asRecord(raw.chromeNode);
  const twilio = asRecord(raw.twilio);
  const voiceCall = asRecord(raw.voiceCall);
@@ -421,6 +465,7 @@ export function resolveGoogleMeetConfigWithEnv(
    chrome: {
      audioBackend: "blackhole-2ch",
      audioFormat,
+      audioBufferBytes,
      launch: resolveBoolean(chrome.launch, DEFAULT_GOOGLE_MEET_CONFIG.chrome.launch),
      browserProfile: normalizeOptionalString(chrome.browserProfile),
      guestName:
@@ -438,10 +483,10 @@ export function resolveGoogleMeetConfigWithEnv(
        chrome.waitForInCallMs,
        DEFAULT_GOOGLE_MEET_CONFIG.chrome.waitForInCallMs,
      ),
-      audioInputCommand: configuredAudioInputCommand ?? [...defaultAudioInputCommand(audioFormat)],
-      audioOutputCommand: configuredAudioOutputCommand ?? [
-        ...defaultAudioOutputCommand(audioFormat),
-      ],
+      audioInputCommand:
+        configuredAudioInputCommand ?? defaultAudioInputCommand(audioFormat, audioBufferBytes),
+      audioOutputCommand:
+        configuredAudioOutputCommand ?? defaultAudioOutputCommand(audioFormat, audioBufferBytes),
      bargeInInputCommand: resolveStringArray(chrome.bargeInInputCommand),
      bargeInRmsThreshold: resolveNumber(
        chrome.bargeInRmsThreshold,