diff --git a/web/public/audio/queue-voice/digit-0.wav b/web/public/audio/queue-voice/digit-0.wav
new file mode 100644
index 0000000..ed8b19a
Binary files /dev/null and b/web/public/audio/queue-voice/digit-0.wav differ
diff --git a/web/public/audio/queue-voice/digit-1.wav b/web/public/audio/queue-voice/digit-1.wav
new file mode 100644
index 0000000..0f84aa5
Binary files /dev/null and b/web/public/audio/queue-voice/digit-1.wav differ
diff --git a/web/public/audio/queue-voice/digit-2.wav b/web/public/audio/queue-voice/digit-2.wav
new file mode 100644
index 0000000..9d144bb
Binary files /dev/null and b/web/public/audio/queue-voice/digit-2.wav differ
diff --git a/web/public/audio/queue-voice/digit-3.wav b/web/public/audio/queue-voice/digit-3.wav
new file mode 100644
index 0000000..f199f32
Binary files /dev/null and b/web/public/audio/queue-voice/digit-3.wav differ
diff --git a/web/public/audio/queue-voice/digit-4.wav b/web/public/audio/queue-voice/digit-4.wav
new file mode 100644
index 0000000..7968ef4
Binary files /dev/null and b/web/public/audio/queue-voice/digit-4.wav differ
diff --git a/web/public/audio/queue-voice/digit-5.wav b/web/public/audio/queue-voice/digit-5.wav
new file mode 100644
index 0000000..bfc4797
Binary files /dev/null and b/web/public/audio/queue-voice/digit-5.wav differ
diff --git a/web/public/audio/queue-voice/digit-6.wav b/web/public/audio/queue-voice/digit-6.wav
new file mode 100644
index 0000000..0d4b986
Binary files /dev/null and b/web/public/audio/queue-voice/digit-6.wav differ
diff --git a/web/public/audio/queue-voice/digit-7.wav b/web/public/audio/queue-voice/digit-7.wav
new file mode 100644
index 0000000..26b7bdd
Binary files /dev/null and b/web/public/audio/queue-voice/digit-7.wav differ
diff --git a/web/public/audio/queue-voice/digit-8.wav b/web/public/audio/queue-voice/digit-8.wav
new file mode 100644
index 0000000..acaa325
Binary files /dev/null and b/web/public/audio/queue-voice/digit-8.wav differ
diff --git a/web/public/audio/queue-voice/digit-9.wav b/web/public/audio/queue-voice/digit-9.wav
new file mode 100644
index 0000000..3a1b0e1
Binary files /dev/null and b/web/public/audio/queue-voice/digit-9.wav differ
diff --git a/web/public/audio/queue-voice/entrance.wav b/web/public/audio/queue-voice/entrance.wav
new file mode 100644
index 0000000..6052326
Binary files /dev/null and b/web/public/audio/queue-voice/entrance.wav differ
diff --git a/web/public/audio/queue-voice/gap.wav b/web/public/audio/queue-voice/gap.wav
new file mode 100644
index 0000000..1d8f74a
Binary files /dev/null and b/web/public/audio/queue-voice/gap.wav differ
diff --git a/web/public/audio/queue-voice/number.wav b/web/public/audio/queue-voice/number.wav
new file mode 100644
index 0000000..86ec51f
Binary files /dev/null and b/web/public/audio/queue-voice/number.wav differ
diff --git a/web/public/audio/queue-voice/please.wav b/web/public/audio/queue-voice/please.wav
new file mode 100644
index 0000000..8547cdb
Binary files /dev/null and b/web/public/audio/queue-voice/please.wav differ
diff --git a/web/public/audio/queue-voice/ready.wav b/web/public/audio/queue-voice/ready.wav
new file mode 100644
index 0000000..671c438
Binary files /dev/null and b/web/public/audio/queue-voice/ready.wav differ
diff --git a/web/public/audio/queue-voice/through.wav b/web/public/audio/queue-voice/through.wav
new file mode 100644
index 0000000..92f9775
Binary files /dev/null and b/web/public/audio/queue-voice/through.wav differ
diff --git a/web/src/components/AudioControl.tsx b/web/src/components/AudioControl.tsx
new file mode 100644
index 0000000..73ee621
--- /dev/null
+++ b/web/src/components/AudioControl.tsx
@@ -0,0 +1,37 @@
+import type { RefObject } from "react";
+import type { AudioAnnouncementStatus } from "../hooks/useAudioAnnouncements";
+
+interface AudioControlProps {
+  audioRef: RefObject<HTMLAudioElement | null>;
+  status: AudioAnnouncementStatus;
+  onEnable: () => void;
+}
+
+export function AudioControl({ audioRef, status, onEnable }: AudioControlProps) {
+  return (
+    <>
+      <audio ref={audioRef} preload="auto" aria-hidden="true" />
+      {status === "ready" ? (
+        <aside className="speech-control speech-control--ready" role="status">
+          <strong>叫号声音已开启</strong>
+          <span>后续叫号将自动播报三次</span>
+        </aside>
+      ) : (
+        <aside className={`speech-control${status === "error" ? " speech-control--error" : ""}`} role="status">
+          <div>
+            <strong>{status === "error" ? "叫号声音开启失败" : "叫号声音尚未开启"}</strong>
+            <span>{status === "error" ? "请检查设备媒体音量后重试" : "请点击一次，之后将自动播报"}</span>
+          </div>
+          <button
+            className="button button--primary"
+            type="button"
+            disabled={status === "enabling"}
+            onClick={onEnable}
+          >
+            {status === "enabling" ? "正在开启叫号声音" : status === "error" ? "重新开启叫号声音" : "开启叫号声音"}
+          </button>
+        </aside>
+      )}
+    </>
+  );
+}
diff --git a/web/src/components/SpeechControl.tsx b/web/src/components/SpeechControl.tsx
deleted file mode 100644
index 874deba..0000000
--- a/web/src/components/SpeechControl.tsx
+++ /dev/null
@@ -1,38 +0,0 @@
-import type { SpeechAnnouncementStatus } from "../hooks/useSpeechAnnouncements";
-
-interface SpeechControlProps {
-  status: SpeechAnnouncementStatus;
-  onEnable: () => void;
-}
-
-export function SpeechControl({ status, onEnable }: SpeechControlProps) {
-  if (status === "ready") {
-    return (
-      <aside className="speech-control speech-control--ready" role="status">
-        <strong>叫号声音已开启</strong>
-        <span>后续叫号将自动播报三次</span>
-      </aside>
-    );
-  }
-
-  const unsupported = status === "unsupported";
-  const failed = status === "error";
-  return (
-    <aside className={`speech-control${failed || unsupported ? " speech-control--error" : ""}`} role="status">
-      <div>
-        <strong>{unsupported ? "当前浏览器不支持叫号声音" : failed ? "叫号声音开启失败" : "叫号声音尚未开启"}</strong>
-        <span>{unsupported || failed ? "请确认系统已安装中文语音服务" : "Chrome 需要先点击一次，之后才允许自动播报"}</span>
-      </div>
-      {!unsupported ? (
-        <button
-          className="button button--primary"
-          type="button"
-          disabled={status === "enabling"}
-          onClick={onEnable}
-        >
-          {status === "enabling" ? "正在开启叫号声音" : failed ? "重新开启叫号声音" : "开启叫号声音"}
-        </button>
-      ) : null}
-    </aside>
-  );
-}
diff --git a/web/src/hooks/useAudioAnnouncements.test.ts b/web/src/hooks/useAudioAnnouncements.test.ts
new file mode 100644
index 0000000..ceabd46
--- /dev/null
+++ b/web/src/hooks/useAudioAnnouncements.test.ts
@@ -0,0 +1,37 @@
+import { describe, expect, it } from "vitest";
+import { announcementAudioQueue } from "./useAudioAnnouncements";
+
+describe("announcementAudioQueue", () => {
+  it("忽略项目名并把单号码播报素材重复三遍", () => {
+    const once = [
+      "/audio/queue-voice/please.wav",
+      "/audio/queue-voice/digit-0.wav",
+      "/audio/queue-voice/digit-0.wav",
+      "/audio/queue-voice/digit-0.wav",
+      "/audio/queue-voice/digit-1.wav",
+      "/audio/queue-voice/digit-6.wav",
+      "/audio/queue-voice/number.wav",
+      "/audio/queue-voice/entrance.wav",
+      "/audio/queue-voice/gap.wav",
+    ];
+
+    expect(announcementAudioQueue(["东门观光车，请零零零一六号，前往入口。"]))
+      .toEqual([...once, ...once, ...once]);
+  });
+
+  it("支持连续号码范围", () => {
+    const queue = announcementAudioQueue(["请一七号至一九号，前往入口。"]);
+
+    expect(queue.slice(0, 9)).toEqual([
+      "/audio/queue-voice/please.wav",
+      "/audio/queue-voice/digit-1.wav",
+      "/audio/queue-voice/digit-7.wav",
+      "/audio/queue-voice/number.wav",
+      "/audio/queue-voice/through.wav",
+      "/audio/queue-voice/digit-1.wav",
+      "/audio/queue-voice/digit-9.wav",
+      "/audio/queue-voice/number.wav",
+      "/audio/queue-voice/entrance.wav",
+    ]);
+  });
+});
diff --git a/web/src/hooks/useAudioAnnouncements.ts b/web/src/hooks/useAudioAnnouncements.ts
new file mode 100644
index 0000000..b0e8e9a
--- /dev/null
+++ b/web/src/hooks/useAudioAnnouncements.ts
@@ -0,0 +1,143 @@
+import { useCallback, useEffect, useRef, useState } from "react";
+
+export type AudioAnnouncementStatus = "locked" | "enabling" | "ready" | "error";
+
+const AUDIO_BASE_PATH = "/audio/queue-voice";
+const ANNOUNCEMENT_REPEAT_COUNT = 3;
+const GAP_AUDIO = `${AUDIO_BASE_PATH}/gap.wav`;
+
+const DIGIT_AUDIO: Record<string, string> = {
+  零: `${AUDIO_BASE_PATH}/digit-0.wav`,
+  一: `${AUDIO_BASE_PATH}/digit-1.wav`,
+  二: `${AUDIO_BASE_PATH}/digit-2.wav`,
+  三: `${AUDIO_BASE_PATH}/digit-3.wav`,
+  四: `${AUDIO_BASE_PATH}/digit-4.wav`,
+  五: `${AUDIO_BASE_PATH}/digit-5.wav`,
+  六: `${AUDIO_BASE_PATH}/digit-6.wav`,
+  七: `${AUDIO_BASE_PATH}/digit-7.wav`,
+  八: `${AUDIO_BASE_PATH}/digit-8.wav`,
+  九: `${AUDIO_BASE_PATH}/digit-9.wav`,
+};
+
+const PHRASE_AUDIO = [
+  ["前往入口", `${AUDIO_BASE_PATH}/entrance.wav`],
+  ["请", `${AUDIO_BASE_PATH}/please.wav`],
+  ["号", `${AUDIO_BASE_PATH}/number.wav`],
+  ["至", `${AUDIO_BASE_PATH}/through.wav`],
+] as const;
+
+function clipsForAnnouncement(announcement: string): string[] {
+  const callStart = announcement.indexOf("请");
+  if (callStart < 0) return [];
+
+  const callText = announcement.slice(callStart);
+  const clips: string[] = [];
+  for (let index = 0; index < callText.length;) {
+    const phrase = PHRASE_AUDIO.find(([text]) => callText.startsWith(text, index));
+    if (phrase) {
+      clips.push(phrase[1]);
+      index += phrase[0].length;
+      continue;
+    }
+
+    const digit = DIGIT_AUDIO[callText[index]];
+    if (digit) clips.push(digit);
+    index += 1;
+  }
+
+  return clips.length ? [...clips, GAP_AUDIO] : [];
+}
+
+export function announcementAudioQueue(announcements: string[]): string[] {
+  const queue: string[] = [];
+  for (const announcement of announcements) {
+    const clips = clipsForAnnouncement(announcement);
+    for (let repeat = 0; repeat < ANNOUNCEMENT_REPEAT_COUNT; repeat += 1) {
+      queue.push(...clips);
+    }
+  }
+  return queue;
+}
+
+export function useAudioAnnouncements() {
+  const audioRef = useRef<HTMLAudioElement | null>(null);
+  const playbackIdRef = useRef(0);
+  const [status, setStatus] = useState<AudioAnnouncementStatus>("locked");
+
+  const playQueue = useCallback((sources: string[], onStarted?: () => void) => {
+    const audio = audioRef.current;
+    if (!audio || !sources.length) {
+      setStatus("error");
+      return;
+    }
+
+    const playbackId = ++playbackIdRef.current;
+    audio.pause();
+    audio.onended = null;
+    audio.onerror = null;
+
+    let cursor = 0;
+    let started = false;
+    const markStarted = () => {
+      if (started || playbackIdRef.current !== playbackId) return;
+      started = true;
+      onStarted?.();
+    };
+    const fail = (error?: unknown) => {
+      if (playbackIdRef.current !== playbackId) return;
+      if (error instanceof DOMException && error.name === "AbortError") return;
+      audio.onended = null;
+      audio.onerror = null;
+      setStatus("error");
+    };
+    const playNext = () => {
+      if (playbackIdRef.current !== playbackId) return;
+      if (cursor >= sources.length) {
+        audio.onended = null;
+        audio.onerror = null;
+        return;
+      }
+
+      audio.src = sources[cursor];
+      cursor += 1;
+      audio.currentTime = 0;
+      audio.onended = playNext;
+      audio.onerror = () => fail();
+
+      try {
+        const result = audio.play();
+        if (result && typeof result.then === "function") {
+          void result.then(markStarted).catch(fail);
+        } else {
+          markStarted();
+        }
+      } catch (error) {
+        fail(error);
+      }
+    };
+
+    playNext();
+  }, []);
+
+  const enable = useCallback(() => {
+    setStatus("enabling");
+    playQueue([`${AUDIO_BASE_PATH}/ready.wav`], () => setStatus("ready"));
+  }, [playQueue]);
+
+  const announce = useCallback((announcements: string[]) => {
+    if (status !== "ready" || !announcements.length) return;
+    const queue = announcementAudioQueue(announcements);
+    if (queue.length) playQueue(queue);
+  }, [playQueue, status]);
+
+  useEffect(() => () => {
+    playbackIdRef.current += 1;
+    const audio = audioRef.current;
+    if (!audio) return;
+    audio.pause();
+    audio.onended = null;
+    audio.onerror = null;
+  }, []);
+
+  return { announce, audioRef, enable, status };
+}
diff --git a/web/src/hooks/useSpeechAnnouncements.ts b/web/src/hooks/useSpeechAnnouncements.ts
deleted file mode 100644
index 73fbff1..0000000
--- a/web/src/hooks/useSpeechAnnouncements.ts
+++ /dev/null
@@ -1,63 +0,0 @@
-import { useCallback, useState } from "react";
-
-export type SpeechAnnouncementStatus = "locked" | "enabling" | "ready" | "error" | "unsupported";
-
-const ANNOUNCEMENT_REPEAT_COUNT = 3;
-
-function speechSupported(): boolean {
-  return typeof window !== "undefined"
-    && "speechSynthesis" in window
-    && typeof SpeechSynthesisUtterance !== "undefined";
-}
-
-function createUtterance(text: string): SpeechSynthesisUtterance {
-  const utterance = new SpeechSynthesisUtterance(text);
-  utterance.lang = "zh-CN";
-  utterance.rate = 0.9;
-  return utterance;
-}
-
-export function useSpeechAnnouncements() {
-  const [status, setStatus] = useState<SpeechAnnouncementStatus>(() => (
-    speechSupported() ? "locked" : "unsupported"
-  ));
-
-  const enable = useCallback(() => {
-    if (!speechSupported()) {
-      setStatus("unsupported");
-      return;
-    }
-
-    const synthesis = window.speechSynthesis;
-    const utterance = createUtterance("叫号声音已开启。");
-    utterance.onstart = () => setStatus("ready");
-    utterance.onend = () => setStatus("ready");
-    utterance.onerror = (event) => {
-      if (event.error !== "canceled" && event.error !== "interrupted") setStatus("error");
-    };
-
-    setStatus("enabling");
-    synthesis.cancel();
-    if (synthesis.paused) synthesis.resume();
-    synthesis.speak(utterance);
-  }, []);
-
-  const announce = useCallback((announcements: string[]) => {
-    if (status !== "ready" || !announcements.length || !speechSupported()) return;
-
-    const synthesis = window.speechSynthesis;
-    synthesis.cancel();
-    if (synthesis.paused) synthesis.resume();
-    for (const announcement of announcements) {
-      for (let repeat = 0; repeat < ANNOUNCEMENT_REPEAT_COUNT; repeat += 1) {
-        const utterance = createUtterance(announcement);
-        utterance.onerror = (event) => {
-          if (event.error !== "canceled" && event.error !== "interrupted") setStatus("error");
-        };
-        synthesis.speak(utterance);
-      }
-    }
-  }, [status]);
-
-  return { announce, enable, status };
-}
diff --git a/web/src/pages/DisplayPage.test.tsx b/web/src/pages/DisplayPage.test.tsx
index 2a5cca2..38ca706 100644
--- a/web/src/pages/DisplayPage.test.tsx
+++ b/web/src/pages/DisplayPage.test.tsx
@@ -48,6 +48,7 @@ vi.mock("../hooks/usePollingResource", () => ({
 }));
 
 import { DisplayPage, forecastRows } from "./DisplayPage";
+import { announcementAudioQueue } from "../hooks/useAudioAnnouncements";
 
 describe("DisplayPage", () => {
   beforeEach(() => {
@@ -94,33 +95,22 @@ describe("DisplayPage", () => {
     ]);
   });
 
-  it("新批次出现时连续播报三次叫号内容", () => {
-    const speak = vi.fn();
-    const cancel = vi.fn();
-    class SpeechSynthesisUtteranceMock {
-      text: string;
-      lang = "";
-      rate = 1;
-      onstart: (() => void) | null = null;
-      onend: (() => void) | null = null;
-      onerror: ((event: { error: string }) => void) | null = null;
+  it("新批次出现时用内置音频连续播报三次叫号内容", async () => {
+    const playedSources: string[] = [];
+    vi.spyOn(HTMLMediaElement.prototype, "pause").mockImplementation(() => undefined);
+    vi.spyOn(HTMLMediaElement.prototype, "play").mockImplementation(function (this: HTMLMediaElement) {
+      playedSources.push(this.getAttribute("src") ?? "");
+      return Promise.resolve();
+    });
+    const { container, rerender } = render(<DisplayPage />);
 
-      constructor(text: string) {
-        this.text = text;
-      }
-    }
-    vi.stubGlobal("SpeechSynthesisUtterance", SpeechSynthesisUtteranceMock);
-    vi.stubGlobal("speechSynthesis", { speak, cancel, paused: false, resume: vi.fn() });
-    const { rerender } = render(<DisplayPage />);
-
-    fireEvent.click(screen.getByRole("button", { name: "开启叫号声音" }));
-    act(() => {
-      speak.mock.calls[0][0].onstart?.();
-      speak.mock.calls[0][0].onerror?.({ error: "interrupted" });
+    await act(async () => {
+      fireEvent.click(screen.getByRole("button", { name: "开启叫号声音" }));
+      await Promise.resolve();
     });
     expect(screen.getByText("叫号声音已开启")).toBeVisible();
-    speak.mockClear();
-    cancel.mockClear();
+    expect(playedSources).toEqual(["/audio/queue-voice/ready.wav"]);
+    playedSources.length = 0;
 
     pollingResource.data.current_batch = {
       batch_number: 3,
@@ -134,14 +124,12 @@ describe("DisplayPage", () => {
     rerender(<DisplayPage />);
     rerender(<DisplayPage />);
 
-    expect(cancel).toHaveBeenCalledOnce();
-    expect(speak).toHaveBeenCalledTimes(3);
-    for (const [utterance] of speak.mock.calls) {
-      expect(utterance).toMatchObject({
-        text: "请零零零零五号至零零零零六号，前往入口。",
-        lang: "zh-CN",
-        rate: 0.9,
-      });
+    const expectedQueue = announcementAudioQueue(["请零零零零五号至零零零零六号，前往入口。"]).slice();
+    const audio = container.querySelector("audio");
+    expect(audio).not.toBeNull();
+    for (let index = 1; index < expectedQueue.length; index += 1) {
+      fireEvent.ended(audio!);
     }
+    expect(playedSources).toEqual(expectedQueue);
   });
 });
diff --git a/web/src/pages/DisplayPage.tsx b/web/src/pages/DisplayPage.tsx
index 94727d2..ce89338 100644
--- a/web/src/pages/DisplayPage.tsx
+++ b/web/src/pages/DisplayPage.tsx
@@ -1,10 +1,10 @@
 import { useEffect, useRef } from "react";
 import { useParams } from "react-router-dom";
 import { api } from "../api";
+import { AudioControl } from "../components/AudioControl";
 import { ProjectScreenTile } from "../components/ProjectScreenTile";
-import { SpeechControl } from "../components/SpeechControl";
+import { useAudioAnnouncements } from "../hooks/useAudioAnnouncements";
 import { usePollingResource } from "../hooks/usePollingResource";
-import { useSpeechAnnouncements } from "../hooks/useSpeechAnnouncements";
 import { formatCallAnnouncement, formatDateTime, isTimestampStale } from "../lib/format";
 import type { AdminProjectDto } from "../types";
 
@@ -12,7 +12,7 @@ export { forecastRows } from "../lib/display";
 
 export function DisplayPage() {
   const { token = "" } = useParams();
-  const { announce, enable: enableSpeech, status: speechStatus } = useSpeechAnnouncements();
+  const { announce, audioRef, enable: enableAudio, status: audioStatus } = useAudioAnnouncements();
   const resource = usePollingResource((signal) => api.display(token, signal), {
     enabled: Boolean(token),
     intervalMs: 3_000,
@@ -70,7 +70,7 @@ export function DisplayPage() {
 
   return (
     <main className="single-display-screen">
-      <SpeechControl status={speechStatus} onEnable={enableSpeech} />
+      <AudioControl audioRef={audioRef} status={audioStatus} onEnable={enableAudio} />
       {(resource.offline || stale || resource.error) ? (
         <div className="display-alert" role="status">
           <strong>{resource.offline ? "连接已中断" : "数据更新延迟"}</strong>
diff --git a/web/src/pages/PublicDisplayCenterPage.test.tsx b/web/src/pages/PublicDisplayCenterPage.test.tsx
index 5ab968a..c1a844b 100644
--- a/web/src/pages/PublicDisplayCenterPage.test.tsx
+++ b/web/src/pages/PublicDisplayCenterPage.test.tsx
@@ -38,6 +38,7 @@ vi.mock("../hooks/usePollingResource", () => ({
 }));
 
 import { PublicDisplayCenterPage } from "./PublicDisplayCenterPage";
+import { announcementAudioQueue } from "../hooks/useAudioAnnouncements";
 
 describe("PublicDisplayCenterPage", () => {
   beforeEach(() => {
@@ -62,28 +63,22 @@ describe("PublicDisplayCenterPage", () => {
     expect(screen.queryByRole("navigation", { name: "管理任务" })).not.toBeInTheDocument();
   });
 
-  it("新叫号批次出现时连续三次播报项目和号码", () => {
-    const speak = vi.fn();
-    class SpeechSynthesisUtteranceMock {
-      text: string;
-      lang = "";
-      rate = 1;
-      onstart: (() => void) | null = null;
-      onend: (() => void) | null = null;
-      onerror: ((event: { error: string }) => void) | null = null;
+  it("新叫号批次出现时只用内置音频连续三次播报号码", async () => {
+    const playedSources: string[] = [];
+    vi.spyOn(HTMLMediaElement.prototype, "pause").mockImplementation(() => undefined);
+    vi.spyOn(HTMLMediaElement.prototype, "play").mockImplementation(function (this: HTMLMediaElement) {
+      playedSources.push(this.getAttribute("src") ?? "");
+      return Promise.resolve();
+    });
+    const { container, rerender } = render(<PublicDisplayCenterPage />);
 
-      constructor(text: string) {
-        this.text = text;
-      }
-    }
-    vi.stubGlobal("SpeechSynthesisUtterance", SpeechSynthesisUtteranceMock);
-    vi.stubGlobal("speechSynthesis", { speak, cancel: vi.fn(), paused: false, resume: vi.fn() });
-    const { rerender } = render(<PublicDisplayCenterPage />);
-
-    fireEvent.click(screen.getByRole("button", { name: "开启叫号声音" }));
-    act(() => speak.mock.calls[0][0].onstart?.());
+    await act(async () => {
+      fireEvent.click(screen.getByRole("button", { name: "开启叫号声音" }));
+      await Promise.resolve();
+    });
     expect(screen.getByText("叫号声音已开启")).toBeVisible();
-    speak.mockClear();
+    expect(playedSources).toEqual(["/audio/queue-voice/ready.wav"]);
+    playedSources.length = 0;
 
     pollingResource.data = {
       ...pollingResource.data,
@@ -104,12 +99,12 @@ describe("PublicDisplayCenterPage", () => {
     };
     rerender(<PublicDisplayCenterPage />);
 
-    expect(speak).toHaveBeenCalledTimes(3);
-    for (const [utterance] of speak.mock.calls) {
-      expect(utterance).toMatchObject({
-        text: "东门观光车，请零零零一六号，前往入口。",
-        lang: "zh-CN",
-      });
+    const expectedQueue = announcementAudioQueue(["请零零零一六号，前往入口。"]).slice();
+    const audio = container.querySelector("audio");
+    expect(audio).not.toBeNull();
+    for (let index = 1; index < expectedQueue.length; index += 1) {
+      fireEvent.ended(audio!);
     }
+    expect(playedSources).toEqual(expectedQueue);
   });
 });
diff --git a/web/src/pages/PublicDisplayCenterPage.tsx b/web/src/pages/PublicDisplayCenterPage.tsx
index 9672213..1892663 100644
--- a/web/src/pages/PublicDisplayCenterPage.tsx
+++ b/web/src/pages/PublicDisplayCenterPage.tsx
@@ -1,9 +1,9 @@
 import { useEffect, useRef } from "react";
 import { api } from "../api";
+import { AudioControl } from "../components/AudioControl";
 import { FeedbackBanner, FreshnessBanner, LoadingState } from "../components/Feedback";
-import { SpeechControl } from "../components/SpeechControl";
+import { useAudioAnnouncements } from "../hooks/useAudioAnnouncements";
 import { usePollingResource } from "../hooks/usePollingResource";
-import { useSpeechAnnouncements } from "../hooks/useSpeechAnnouncements";
 import { formatCallAnnouncement, isTimestampStale } from "../lib/format";
 import type { CallBatchDto } from "../types";
 import { DisplayCenter } from "./AdminPage";
@@ -13,7 +13,7 @@ function projectCurrentBatch(value: unknown): CallBatchDto | null {
 }
 
 export function PublicDisplayCenterPage() {
-  const { announce, enable: enableSpeech, status: speechStatus } = useSpeechAnnouncements();
+  const { announce, audioRef, enable: enableAudio, status: audioStatus } = useAudioAnnouncements();
   const resource = usePollingResource((signal) => api.displayOverview(signal), { intervalMs: 5_000 });
   const data = resource.data;
   const stale = Boolean(data) && isTimestampStale(resource.lastClientSuccessAt, 30_000);
@@ -29,7 +29,7 @@ export function PublicDisplayCenterPage() {
       nextCalls.set(project.id, key);
       if (!previousCalls?.has(project.id) || !key || previousCalls.get(project.id) === key) continue;
       const announcement = formatCallAnnouncement(batch);
-      if (announcement) announcements.push(`${project.name}，${announcement}`);
+      if (announcement) announcements.push(announcement);
     }
     lastCallsRef.current = nextCalls;
     announce(announcements);
@@ -38,7 +38,7 @@ export function PublicDisplayCenterPage() {
   return (
     <div className="app-shell app-shell--admin app-shell--no-primary-action">
       <a className="skip-link" href="#main-content">跳到主要内容</a>
-      <SpeechControl status={speechStatus} onEnable={enableSpeech} />
+      <AudioControl audioRef={audioRef} status={audioStatus} onEnable={enableAudio} />
       <header className="app-header">
         <div className="brand-lockup" aria-label="景区排队叫号系统">
           <span className="brand-logo-frame" aria-hidden="true"><img className="brand-logo" src="/xiaoqikong-logo.jpg" alt="" /></span>