Add browser-local Piper WebAssembly speech

voice.js

@@ -82,6 +82,9 @@
     let listeningTimer = null;
     let currentUtterance = null;
     let speechPrimed = false;
+    let piperAudioContext = null;
+    let piperSource = null;
+    let piperAudioUnlocked = false;
 
     function setState(next, message = '') {
       state = next;
@@ -95,19 +98,34 @@
       speechTimer = null;
       speechStartTimer = null;
       const wasQueued = currentUtterance !== null;
+      const wasPiper = currentUtterance?.kind === 'piper';
       currentUtterance = null;
       // WebKit versions before Safari 27 can discard a speak() queued in the
       // same turn as cancel(). Do not clear an already-empty native queue.
       if (wasQueued) {
-        try { win.speechSynthesis?.cancel(); } catch (_) { /* Best effort. */ }
-        // cancel() also removes the silent user-gesture primer from the native
-        // queue. A later reply must prime again from the next player gesture.
-        speechPrimed = false;
+        if (wasPiper) {
+          try { piperSource?.stop(); } catch (_) { /* Already stopped. */ }
+          piperSource = null;
+        } else {
+          try { win.speechSynthesis?.cancel(); } catch (_) { /* Best effort. */ }
+          // cancel() also removes the silent user-gesture primer from the native
+          // queue. A later reply must prime again from the next player gesture.
+          speechPrimed = false;
+        }
       }
       if (!recognition) setState('IDLE');
     }
 
     function activateFromUserGesture() {
+      const AudioContext = win.AudioContext || win.webkitAudioContext;
+      if (win.GamePiper && AudioContext) {
+        try {
+          piperAudioContext ||= new AudioContext();
+          if (piperAudioContext.state === 'suspended') piperAudioContext.resume().catch(() => {});
+          piperAudioUnlocked = true;
+          return true;
+        } catch (_) { /* Fall through to the browser speech adapter. */ }
+      }
       if (!win.speechSynthesis || !win.SpeechSynthesisUtterance) return false;
       if (speechPrimed) return true;
       // iOS Safari requires speech output to be initiated by a user gesture.
@@ -211,7 +229,81 @@
       try { return win.speechSynthesis?.getVoices?.() || []; } catch (_) { return []; }
     }
 
+    function speakWithPiper(text, voiceOptions) {
+      const AudioContext = win.AudioContext || win.webkitAudioContext;
+      if (!AudioContext || !piperAudioUnlocked) return false;
+      stopListening();
+      if (currentUtterance) cancelSpeech();
+      const token = ++generation;
+      const playback = {kind: 'piper', token};
+      currentUtterance = playback;
+      setState('QUEUED', 'Förbereder svensk uppläsning. Första gången kan en röstmodell på cirka 63 MB behöva hämtas till enheten.');
+
+      if (speechTimer !== null) win.clearTimeout(speechTimer);
+      speechTimer = win.setTimeout(() => {
+        if (token !== generation) return;
+        generation += 1;
+        currentUtterance = null;
+        try { piperSource?.stop(); } catch (_) { /* Best effort. */ }
+        piperSource = null;
+        speechTimer = null;
+        setState('IDLE', 'Piper tog för lång tid på sig. Dialogen finns kvar som text. Tryck på Läs senaste repliken för att försöka igen.');
+      }, 180000);
+
+      let lastProgress = -1;
+      const progress = event => {
+        if (token !== generation || !event?.total) return;
+        const percent = Math.min(100, Math.floor((event.loaded / event.total) * 100));
+        if (percent === lastProgress) return;
+        lastProgress = percent;
+        setState('PROCESSING', `Hämtar svensk röstmodell till enheten: ${percent} procent.`);
+      };
+
+      Promise.resolve()
+        .then(() => win.GamePiper.synthesize(String(text), progress))
+        .then(async blob => {
+          if (token !== generation) return;
+          piperAudioContext ||= new AudioContext();
+          if (piperAudioContext.state === 'suspended') await piperAudioContext.resume();
+          const buffer = await piperAudioContext.decodeAudioData(await blob.arrayBuffer());
+          if (token !== generation) return;
+          const source = piperAudioContext.createBufferSource();
+          const profile = voiceOptions.characterId ? settings.speakerProfiles?.[voiceOptions.characterId] || {} : {};
+          const identity = characterVoiceProfile(voiceOptions.voiceIdentity, voiceOptions.characterId);
+          const narrator = voiceOptions.role === 'narrator';
+          const requestedRate = Number(voiceOptions.rate ?? profile.rate ?? settings.rate) || 1;
+          source.buffer = buffer;
+          source.playbackRate.value = Math.min(2, Math.max(0.5, requestedRate * (narrator || voiceOptions.rate != null || profile.rate != null ? 1 : identity.rate)));
+          const pitch = narrator ? 1 : Number(voiceOptions.pitch ?? profile.pitch ?? identity.pitch) || 1;
+          if (source.detune) source.detune.value = Math.min(600, Math.max(-600, (pitch - 1) * 1200));
+          source.connect(piperAudioContext.destination);
+          source.onended = () => {
+            if (token !== generation) return;
+            if (speechTimer !== null) win.clearTimeout(speechTimer);
+            speechTimer = null;
+            currentUtterance = null;
+            piperSource = null;
+            setState('IDLE');
+          };
+          piperSource = source;
+          source.start();
+          if (typeof voiceOptions.onStart === 'function') voiceOptions.onStart();
+          setState('SPEAKING', 'Läser upp dialogen med svensk röst.');
+        })
+        .catch(error => {
+          if (token !== generation) return;
+          if (speechTimer !== null) win.clearTimeout(speechTimer);
+          speechTimer = null;
+          currentUtterance = null;
+          piperSource = null;
+          generation += 1;
+          setState('IDLE', `Piper kunde inte skapa eller spela upp ljudet${error?.message ? `: ${error.message}` : '.'} Dialogen finns kvar som text. Tryck på Läs senaste repliken för att försöka igen.`);
+        });
+      return true;
+    }
+
     function speak(text, voiceOptions = {}) {
+      if (text && win.GamePiper?.synthesize && speakWithPiper(text, voiceOptions)) return true;
       if (!text || !win.speechSynthesis || !win.SpeechSynthesisUtterance) {
         setState('IDLE', 'Talsyntes stöds inte i den här webbläsaren. Dialogen finns kvar som text.');
         return false;