fix: precise TTS word-boundary timing and remove wasted sync calculations

- ffmpeg-service/main.py: replace ffsubsync path with exact word-boundary timing from edge_tts for TTS subtitles. Karaoke styling preserved.
- server/services/ttsSync.ts: fix word count to match TTS-cleaned text, remove artificial punctuationPause subtraction, strip punctuation tokens from count.
- server/routes/reels.ts: remove redundant ttsSyncService calls in preview/background; word_duration is ignored by Python, these only wasted TTS generations.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

server/routes/reels.ts

@@ -350,17 +350,7 @@ reelsRouter.post('/reels/preview', async (req: Request, res: Response) => {
             console.error('Error fetching watermark configuration', e);
         }
 
-        // Calculer le word_duration synchronisé si TTS activé
-        let finalWordDuration = wordDuration;
-        if (ttsEnabled && overlayText) {
-            try {
-                const sync = await ttsSyncService.calculateSyncTiming(overlayText, ttsVoice);
-                finalWordDuration = sync.wordDuration;
-                console.log(`🎯 TTS Sync (preview): ${sync.wordCount} mots, ${sync.audioDuration.toFixed(2)}s audio, word_duration=${finalWordDuration.toFixed(3)}s`);
-            } catch (syncErr) {
-                console.warn('⚠️ TTS sync failed (preview), using default word_duration:', syncErr);
-            }
-        }
+        const finalWordDuration = wordDuration;
 
         // Traiter la vidéo via FFmpeg
         const result = await ffmpegService.processReelFromUrl(resolveInternalUrl(media.originalUrl), {
@@ -593,17 +583,7 @@ async function processReelBackground(
             console.error('Error fetching watermark configuration', e);
         }
 
-        // Calculer le word_duration synchronisé si TTS activé
-        let finalWordDuration = wordDuration ?? 0.6;
-        if (ttsEnabled && overlayText && ttsVoice) {
-            try {
-                const sync = await ttsSyncService.calculateSyncTiming(overlayText, ttsVoice);
-                finalWordDuration = sync.wordDuration;
-                console.log(`🎯 TTS Sync (background): ${sync.wordCount} mots, ${sync.audioDuration.toFixed(2)}s audio, word_duration=${finalWordDuration.toFixed(3)}s`);
-            } catch (syncErr) {
-                console.warn('⚠️ TTS sync failed (background), using default word_duration:', syncErr);
-            }
-        }
+        const finalWordDuration = wordDuration ?? 0.6;
 
         const ffmpegResult = await ffmpegService.processReelFromUrl(resolveInternalUrl(media.originalUrl), {
             text: overlayText,

server/services/ttsSync.ts

@@ -16,11 +16,10 @@ export class TtsSyncService {
    * avec la durée réelle de la voix TTS générée.
    */
   async calculateSyncTiming(text: string, voice: string): Promise<SyncTiming> {
-    const cleanTtsText = this.cleanForTts(text);
-    const cleanDisplayText = this.cleanForDisplay(text);
+    const cleanText = this.cleanText(text);
 
     // 1. Générer le TTS preview et mesurer sa durée exacte
-    const ttsResult = await ffmpegService.previewTTS(cleanTtsText, voice);
+    const ttsResult = await ffmpegService.previewTTS(cleanText, voice);
     if (!ttsResult.success || !ttsResult.audioBase64) {
       throw new Error('TTS preview failed: ' + (ttsResult.error || 'unknown'));
     }
@@ -29,14 +28,11 @@ export class TtsSyncService {
     const metadata = await musicMetadata.parseBuffer(audioBuffer, 'audio/mpeg');
     const audioDuration = metadata.format.duration || 0;
 
-    // 2. Analyser le texte
-    const words = cleanDisplayText.split(/\s+/).filter(w => w.length > 0);
-    const wordCount = words.length;
-    const punctuationPause = this.calculatePunctuationPauses(cleanDisplayText);
+    // 2. Analyser le texte (compte les mots réellement lus par la voix)
+    const wordCount = this.calculateWordCount(cleanText);
 
-    // 3. Calculer le word_duration ajusté
-    const effectiveDuration = Math.max(audioDuration - punctuationPause, 0.5);
-    const wordDuration = wordCount > 0 ? effectiveDuration / wordCount : 0.6;
+    // 3. Calculer le word_duration
+    const wordDuration = wordCount > 0 ? audioDuration / wordCount : 0.6;
 
     // 4. Validation
     const warnings: string[] = [];
@@ -52,13 +48,13 @@ export class TtsSyncService {
       wordDuration,
       audioDuration,
       wordCount,
-      punctuationPause,
+      punctuationPause: 0,
       isHealthy,
       warnings,
     };
   }
 
-  private cleanForTts(text: string): string {
+  private cleanText(text: string): string {
     return text
       .replace(/#\w+/g, '')
       .replace(/https?:\/\/\S+/g, '')
@@ -66,23 +62,9 @@ export class TtsSyncService {
       .trim();
   }
 
-  private cleanForDisplay(text: string): string {
-    return text.trim();
-  }
-
-  private calculatePunctuationPauses(text: string): number {
-    let pause = 0;
-    const chars = text.split('');
-    for (const c of chars) {
-      if (',;'.includes(c)) pause += 0.3;
-      else if ('.!?'.includes(c)) pause += 0.6;
-      else if (':'.includes(c)) pause += 0.4;
-      else if (c === '\n') pause += 0.3;
-    }
-    // "..." compte comme un seul point mais pause plus longue
-    const ellipsisCount = (text.match(/\.\.\./g) || []).length;
-    pause += ellipsisCount * 0.3; // bonus pour les points de suspension
-    return pause;
+  private calculateWordCount(text: string): number {
+    const tokens = text.split(/\s+/).filter(w => w.length > 0);
+    return tokens.filter(w => /[a-zA-Z0-9À-ſ]/.test(w)).length;
   }
 }