feat(audio): add offline local neural voice via Piper (sv + en)

Announcements can use Piper, a free local neural TTS, instead of the browser's system
voice. No account, no paid service, and it keeps working with no internet at the track.

- npm run tts:install (or Settings -> Audio -> Install) downloads the standalone Piper
  release for Linux/Windows/macOS plus Swedish and English voices into piper/ once
- server/tts.js generates each line once, one at a time, and caches it as WAV under
  data/tts/; a new line takes ~0.3 s with the standalone build
- src/piper_speech.js plays it and falls back to the browser voice if Piper is missing,
  errors, or takes over 2.5 s, so no announcement is lost and timing is never held up
- "Pre-generate phrases and driver names" renders every fixed phrase and name in both
  languages before race day

Browser voice stays the default. Generation routes fall under the admin PIN.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

src/piper_speech.js

@@ -0,0 +1,60 @@
+// Plays announcements generated by the server's local Piper voice. Any failure - Piper not
+// installed, generation too slow, network hiccup - falls back to the browser's own voice,
+// so an announcement is never lost and never holds up timing.
+const PIPER_TIMEOUT_MS = 2500;
+let current = null;
+
+export function stopPiperSpeech() {
+  if (!current) return;
+  current.cancelled = true;
+  current.controller.abort();
+  if (current.audio) current.audio.pause();
+  if (current.url) URL.revokeObjectURL(current.url);
+  current = null;
+}
+
+export function isPiperSpeaking() {
+  return Boolean(current);
+}
+
+// Returns true when the line was accepted (it will play via Piper or the fallback).
+export function speakWithPiper(text, { language, rate, volume = 1, interrupt, onEnd, onError, fallback }) {
+  if (current && !interrupt) return false;
+  stopPiperSpeech();
+  const job = { cancelled: false, controller: new AbortController(), audio: null, url: "" };
+  current = job;
+  const finish = (callback) => {
+    if (job.url) URL.revokeObjectURL(job.url);
+    if (current === job) current = null;
+    callback?.();
+  };
+  const timer = setTimeout(() => job.controller.abort(), PIPER_TIMEOUT_MS);
+  fetch("/api/tts/speech", {
+    method: "POST",
+    headers: { "Content-Type": "application/json" },
+    body: JSON.stringify({ text, language, rate }),
+    signal: job.controller.signal,
+  })
+    .then((response) => {
+      if (!response.ok) throw new Error(`tts ${response.status}`);
+      return response.blob();
+    })
+    .then((blob) => {
+      clearTimeout(timer);
+      if (job.cancelled) return;
+      job.url = URL.createObjectURL(blob);
+      job.audio = new Audio(job.url);
+      job.audio.volume = Math.max(0, Math.min(1, Number(volume)));
+      job.audio.onended = () => finish(onEnd);
+      job.audio.onerror = () => finish(onError);
+      return job.audio.play();
+    })
+    .catch(() => {
+      clearTimeout(timer);
+      if (job.cancelled) return;
+      finish(() => {
+        if (!fallback?.()) onError?.();
+      });
+    });
+  return true;
+}