// ── TTS preview ─────────────────────────────────────────────────────────── function backendVoiceId(value) { return typeof value === 'string' ? value : (value?.id || value?.voice || value?.name || JSON.stringify(value)); } function shouldFilterBackendVoices(backend) { return ['voice_clone', 'streaming', 'nvidia_zeroshot', 'nvidia_flow'].includes(backend || ''); } async function activeLibraryVoiceIds() { if (!_voices.length) await loadVoiceLibrary(); return new Set((_voices || []).filter(v => v.enabled !== false).map(v => v.id)); } function cleanReferenceText(text) { return String(text || '').trim(); } function selectedPreviewLibraryVoice() { const id = $('tts-voice-select')?.value || ''; return id ? (_voices || []).find(v => v.id === id) : null; } function previewVoiceWarnings(v) { const warnings = []; const backend = backendById($('tts-backend-select')?.value || ''); if (backend && backend.id && !['voice_clone', 'streaming', 'nvidia_zeroshot', 'nvidia_flow'].includes(backend.id)) { warnings.push(backend.id === 'nvidia_magpie' ? 'NVIDIA Magpie uses fixed speaker voices, not saved WAV clone identity.' : 'This backend may follow style/model voice more than the saved WAV identity.'); } if (backend && backend.id === 'nvidia_zeroshot' && v.duration && (Number(v.duration) < 3 || Number(v.duration) > 10)) { warnings.push('NVIDIA Zeroshot works best with a clear 3-10 second prompt.'); } if (backend && backend.id === 'nvidia_flow' && !v.transcript) { warnings.push('NVIDIA Flow requires the exact saved reference transcript for this voice.'); } if (!v.transcript) warnings.push('No reference transcript is saved; cloned identity is harder to judge.'); if (v.duration && (Number(v.duration) < 3 || Number(v.duration) > 20)) warnings.push('Reference clip length is outside the 3-20 second sweet spot.'); if (v.needs_tts_restart) warnings.push('This voice changed since the last backend refresh; restart or clear restart flags before judging it.'); const healthWarnings = v.health && Array.isArray(v.health.warnings) ? v.health.warnings : []; warnings.push(...healthWarnings.slice(0, 3)); return warnings; } function updatePreviewVoiceMatchPanel() { const panel = $('preview-match-panel'); if (!panel) return; const v = selectedPreviewLibraryVoice(); if (!v) { panel.hidden = true; return; } panel.hidden = false; const lang = v.language || v.lang || (v.id || '').split('_')[0] || '-'; const gender = v.gender || (v.id || '').split('_')[1] || '-'; const db = fmtDbfs(v); const dur = v.duration ? fmtDuration(v.duration) : '-'; $('preview-match-title').textContent = v.id; $('preview-match-detail').textContent = `${lang} · ${gender} · ${dur} · ${db} dBFS`; const warnings = previewVoiceWarnings(v); $('preview-match-warning').textContent = warnings.length ? warnings.join(' ') : 'For a fair voice match check, play the WAV and synthesize the exact saved reference text.'; const transcript = cleanReferenceText(v.transcript || ''); $('preview-match-transcript').textContent = transcript || 'No reference text saved for this voice.'; $('preview-ref-use-text').disabled = !transcript; $('preview-ref-synth').disabled = !transcript; const audio = $('preview-ref-audio'); const expected = voiceFileUrl(v); if (audio.dataset.src !== expected) { audio.pause(); audio.src = expected; audio.dataset.src = expected; } // Persona rewrite button — shown only when voice has a persona const actionsEl = panel.querySelector('.preview-match-actions'); let personaBtn = panel.querySelector('.preview-persona-btn'); if (v.persona) { if (!personaBtn) { personaBtn = document.createElement('button'); personaBtn.className = 'btn-secondary preview-persona-btn'; personaBtn.type = 'button'; personaBtn.textContent = 'Rewrite with persona'; actionsEl?.appendChild(personaBtn); personaBtn.addEventListener('click', async () => { const text = $('preview-text-area').value.trim(); if (!text) { toast('Enter text to rewrite', 'error'); return; } const lv = selectedPreviewLibraryVoice(); if (!lv?.persona) { toast('This voice has no persona', 'error'); return; } personaBtn.disabled = true; personaBtn.textContent = 'Rewriting…'; try { const llmUrl = localStorage.getItem('refine-llm-url') || 'http://localhost:11434/v1'; const r = await fetch('/api/rewrite-with-persona', { method:'POST', headers:{'Content-Type':'application/json'}, body: JSON.stringify({ text, persona: lv.persona, llm_url: llmUrl, mode:'rewrite' }) }); if (!r.ok) { const e = await r.json().catch(()=>({})); throw new Error(e.detail || r.statusText); } const d = await r.json(); $('preview-text-area').value = d.text; toast('Text rewritten in persona style', 'success'); } catch(e) { toast('Persona rewrite failed: ' + e.message, 'error'); } finally { personaBtn.disabled = false; personaBtn.textContent = 'Rewrite with persona'; } }); } } else { personaBtn?.remove(); } } async function synthesizeSelectedReferenceText() { const v = selectedPreviewLibraryVoice(); if (!v) { toast('Select a library voice first', 'error'); return; } let text = cleanReferenceText(v.transcript || ''); if (!text) { toast('This voice has no reference text', 'error'); return; } if (v.needs_tts_restart) { const ok = confirm('This voice is marked as needing a TTS restart. If you already restarted the backend, clear the flag and synthesize anyway?'); if (!ok) return; await clearTtsRestartFlags(); v.needs_tts_restart = false; updatePreviewVoiceMatchPanel(); } const backend = $('tts-backend-select').value; if (!backend) { toast('No available TTS backend', 'error'); return; } const btn = $('preview-ref-synth'); btn.disabled = true; try { $('preview-text-area').value = text; const source = await createTtsAudioSource(v.id, text, backend, $('preview-playback-mode').value, $('preview-style-instruction').value.trim()); previewBlob = source.blob; const audio = $('preview-audio'); audio.src = source.url; audio.style.display = ''; await audio.play(); $('save-preview-mp3-btn').disabled = false; $('save-preview-btn').disabled = source.streaming; toast(source.streaming ? 'Reference text streaming' : 'Reference text synthesized', 'success'); } catch(e) { toast('Reference synthesis failed: ' + e.message, 'error'); } finally { btn.disabled = false; } } $('fetch-tts-voices-btn').addEventListener('click', async () => { $('fetch-tts-voices-btn').disabled = true; try { const backend = $('tts-backend-select')?.value; if (!backend) throw new Error('No available TTS backend'); const rawVoices = await fetch('/api/tts-voices?backend=' + encodeURIComponent(backend)).then(r => r.json()); let voices = Array.isArray(rawVoices) ? rawVoices : []; if (shouldFilterBackendVoices(backend)) { const activeIds = await activeLibraryVoiceIds(); voices = voices.filter(v => activeIds.has(backendVoiceId(v))); } const sel = $('tts-voice-select'), prev = sel.value; sel.innerHTML = ''; voices.forEach(v => { const id = backendVoiceId(v); const opt = document.createElement('option'); opt.value = opt.textContent = id; sel.appendChild(opt); }); if(prev && voices.some(v => backendVoiceId(v) === prev)) sel.value = prev; updatePreviewVoiceMatchPanel(); const suffix = shouldFilterBackendVoices(backend) ? ' active voices' : ' voices'; toast('Fetched '+voices.length+suffix,'success'); } catch(e) { toast('Fetch failed: '+e.message,'error'); } finally { $('fetch-tts-voices-btn').disabled = false; } }); $('tts-backend-select').addEventListener('change', () => { const sel = $('tts-voice-select'); sel.innerHTML = ''; updateBackendHelp(); updatePreviewVoiceMatchPanel(); previewBlob = null; $('save-preview-mp3-btn').disabled = true; $('save-preview-btn').disabled = true; }); $('tts-voice-select').addEventListener('change', updatePreviewVoiceMatchPanel); $('preview-ref-play').addEventListener('click', async () => { updatePreviewVoiceMatchPanel(); const audio = $('preview-ref-audio'); try { await audio.play(); } catch(e) { toast('Reference playback failed: ' + e.message, 'error'); } }); $('preview-ref-use-text').addEventListener('click', () => { const v = selectedPreviewLibraryVoice(); const text = cleanReferenceText(v?.transcript || ''); if (!text) { toast('This voice has no reference text', 'error'); return; } $('preview-text-area').value = text; toast('Reference text copied to target text', 'success'); }); $('preview-ref-synth').addEventListener('click', synthesizeSelectedReferenceText); let _ttsStreamHealth = null; function effectiveTtsPlaybackMode(override = 'settings') { if (override && override !== 'settings') return override; return _appSettings.tts_stream_mode || 'auto'; } async function isTtsStreamAvailable(force = false) { if (_ttsStreamHealth && !force) return _ttsStreamHealth.ok; try { _ttsStreamHealth = await fetch('/api/tts-stream-health').then(r => r.json()); return !!_ttsStreamHealth.ok; } catch (_) { _ttsStreamHealth = {ok:false}; return false; } } async function createTtsStreamUrl(voice, text, instruct = '') { if (!await isTtsStreamAvailable()) throw new Error('streaming backend unavailable'); const r = await fetch('/api/tts-stream-session', {method:'POST',headers:{'Content-Type':'application/json'}, body:JSON.stringify({text,voice,instruct})}); if (!r.ok) { const e=await r.json().catch(()=>({})); throw new Error(e.detail || r.statusText); } const data = await r.json(); return data.url; } async function fetchTtsPreviewBlob(voice, text, responseFormat = 'wav', instruct = '', backend = 'voice_clone', applyPersona = false) { const body = {text, voice, response_format: responseFormat, instruct, backend}; if (applyPersona) body.apply_persona = true; const r = await fetch('/api/tts-preview', {method:'POST', headers:{'Content-Type':'application/json'}, body: JSON.stringify(body)}); if (!r.ok) { const e=await r.json().catch(()=>({})); throw new Error(e.detail || r.statusText); } return await r.blob(); } async function createTtsAudioSource(voice, text, backend = 'voice_clone', modeOverride = 'settings', instruct = '', applyPersona = false) { const mode = effectiveTtsPlaybackMode(modeOverride); if (backend !== 'streaming' || mode === 'buffered') { const blob = await fetchTtsPreviewBlob(voice, text, 'wav', instruct, backend, applyPersona); return {url: URL.createObjectURL(blob), blob, streaming:false, label:'buffered'}; } try { return {url: await createTtsStreamUrl(voice, text, instruct), blob:null, streaming:true, label:'streaming'}; } catch (e) { if (mode === 'streaming') throw e; const blob = await fetchTtsPreviewBlob(voice, text, 'wav', instruct, backend, applyPersona); return {url: URL.createObjectURL(blob), blob, streaming:false, label:'buffered'}; } } let previewBlob = null; const PREVIEW_SAMPLE_TEXT = 'Hello! This is a voice preview from TTS Voice Creator - Clone and Design.'; $('preview-text-area').addEventListener('focus', () => { if ($('preview-text-area').value === PREVIEW_SAMPLE_TEXT) $('preview-text-area').value = ''; }, { once:true }); function _onPreviewGenerated(source, voice, text, backend) { if (typeof effectsSourceBlob !== 'undefined') window._effectsSourceBlob = null; const ea = $('effects-apply-btn'); if (ea && source.blob) ea.disabled = false; const ap = $('add-to-playlist-btn'); if (ap && source.blob) ap.disabled = false; if (typeof historyPush === 'function' && source.blob) historyPush(voice, text, backend, source.blob, source.url); } $('preview-btn').addEventListener('click', async () => { const voice=$('tts-voice-select').value, backend=$('tts-backend-select').value, text=$('preview-text-area').value.trim(), instruct=$('preview-style-instruction').value.trim(); const applyPersona = $('preview-persona-toggle')?.checked || false; if(!backend) { toast('No available TTS backend','error'); return; } if(!voice) { toast('Select a TTS voice','error'); return; } if(!text) { toast('Enter preview text','error'); return; } $('preview-btn').disabled=true; $('save-preview-mp3-btn').disabled=true; $('save-preview-btn').disabled=true; if($('add-to-playlist-btn')) $('add-to-playlist-btn').disabled=true; if($('effects-apply-btn')) $('effects-apply-btn').disabled=true; try { const audio = $('preview-audio'); const useChunked = $('preview-chunked-toggle')?.checked && text.length > 200 && typeof generateChunkedTts === 'function'; const source = useChunked ? await generateChunkedTts(voice, text, backend, instruct) : await createTtsAudioSource(voice, text, backend, $('preview-playback-mode').value, instruct, applyPersona); previewBlob = source.blob; window._previewVoice = voice; window._previewBackend = backend; window._previewText = text; audio.src = source.url; audio.style.display=''; await audio.play(); $('save-preview-mp3-btn').disabled = false; $('save-preview-btn').disabled = source.streaming; _onPreviewGenerated(source, voice, text, backend); toast(source.streaming ? 'Streaming preview playing' : source.label === 'chunked' ? `Chunked (${text.length} chars) playing` : 'Preview playing', 'success'); } catch(e) { toast('TTS failed: '+e.message,'error'); } finally { $('preview-btn').disabled=false; } }); $('save-preview-mp3-btn').addEventListener('click', async () => { const voice=$('tts-voice-select').value, backend=$('tts-backend-select').value, text=$('preview-text-area').value.trim(), instruct=$('preview-style-instruction').value.trim(); if(!backend) { toast('No available TTS backend','error'); return; } if(!voice || !text) return; const btn = $('save-preview-mp3-btn'); btn.disabled = true; try { const blob = await fetchTtsPreviewBlob(voice, text, 'mp3', instruct, backend); const a = document.createElement('a'); a.href = URL.createObjectURL(blob); a.download = (voice||'preview')+'_preview.mp3'; a.click(); toast('MP3 saved', 'success'); } catch(e) { toast('MP3 save failed: '+e.message,'error'); } finally { btn.disabled = false; } }); $('save-preview-btn').addEventListener('click', () => { if(!previewBlob) return; const a = document.createElement('a'); a.href = URL.createObjectURL(previewBlob); a.download = ($('tts-voice-select').value||'preview')+'_preview.wav'; a.click(); }); // ── Performance benchmark ───────────────────────────────────────────────── const PERF_HISTORY_KEY = 'vcf-perf-history'; const PERF_HISTORY_MAX = 50; function perfHistoryLoad() { try { return JSON.parse(localStorage.getItem(PERF_HISTORY_KEY) || '[]'); } catch(_) { return []; } } function perfHistorySave(entries) { try { localStorage.setItem(PERF_HISTORY_KEY, JSON.stringify(entries.slice(-PERF_HISTORY_MAX))); } catch(_) {} } function perfHistoryAdd(entry) { const h = perfHistoryLoad(); h.push(entry); perfHistorySave(h); } function perfSparklineSvg(rtfValues) { if (!rtfValues.length) return ''; const W = 120, H = 32, PAD = 2, barW = Math.max(4, Math.floor((W - PAD * 2) / rtfValues.length) - 1); const maxV = Math.max(...rtfValues, 1); const bars = rtfValues.map((v, i) => { const bh = Math.max(3, Math.round((v / maxV) * (H - PAD * 2))); const x = PAD + i * (barW + 1); const y = H - PAD - bh; const col = v < 1 ? 'var(--green)' : 'var(--yellow)'; return ``; }).join(''); return ``; } function renderPerfHistory() { const histList = $('perf-history-list'); if (!histList) return; const filterEl = $('perf-history-filter-current'); const filterOn = filterEl?.checked; const curBack = $('perf-backend-select')?.value; const curVoice = $('perf-voice-select')?.value; let entries = perfHistoryLoad().slice().reverse(); if (filterOn && curBack) entries = entries.filter(e => e.backend === curBack && e.voice === curVoice); if (!entries.length) { histList.innerHTML = '
' + (filterOn ? 'No history for this backend/voice yet.' : 'No benchmark history yet. Run a benchmark above to start tracking.') + '
'; return; } const head = `
Date / TimeBackendVoice Avg latencyMinAvg RTF
`; const rows = entries.map(e => { const dt = new Date(e.ts).toLocaleString([], {month:'2-digit',day:'2-digit',hour:'2-digit',minute:'2-digit'}); const rtfCls = e.avgRtf < 1 ? 'perf-good' : 'perf-slow'; return `
${escHtml(dt)} ${escHtml(e.backend)} ${escHtml(e.voice)} ${Math.round(e.avgLatencyMs)} ms ${Math.round(e.minLatencyMs)} ms ${e.avgRtf.toFixed(2)}
`; }).join(''); histList.innerHTML = head + rows; histList.querySelectorAll('.perf-history-del').forEach(btn => { btn.addEventListener('click', () => { const ts = Number(btn.dataset.ts); perfHistorySave(perfHistoryLoad().filter(e => e.ts !== ts)); renderPerfHistory(); }); }); } (function initPerfBenchmark() { const perfBackendSel = $('perf-backend-select'); const perfVoiceSel = $('perf-voice-select'); const perfFetchBtn = $('perf-fetch-voices-btn'); const perfRunBtn = $('perf-run-btn'); const perfClearBtn = $('perf-clear-btn'); const perfProgress = $('perf-progress'); const perfResultsCard = $('perf-results-card'); const perfSummary = $('perf-summary'); const perfTbody = $('perf-tbody'); const perfText = $('perf-text'); const perfRunsSel = $('perf-runs'); if (!perfRunBtn) return; let perfRows = []; function populatePerfBackends() { if (!perfBackendSel) return; const cur = perfBackendSel.value; perfBackendSel.innerHTML = availableTtsBackends().map(b => `` ).join('') || ''; } populatePerfBackends(); perfFetchBtn.addEventListener('click', async () => { const backend = perfBackendSel.value; if (!backend) { toast('Select a backend first', 'error'); return; } perfFetchBtn.disabled = true; try { const rawVoices = await fetch('/api/tts-voices?backend=' + encodeURIComponent(backend)).then(r => r.json()); const cur = perfVoiceSel.value; perfVoiceSel.innerHTML = rawVoices.map(v => { const id = backendVoiceId(v); return ``; }).join('') || ''; } catch(e) { toast('Fetch voices failed: '+e.message, 'error'); } finally { perfFetchBtn.disabled = false; } }); function updateTrendDisplay(backend, voice, currentAvgRtf) { const trendRow = $('perf-trend-row'); const trendBadge = $('perf-trend-badge'); const sparkWrap = $('perf-sparkline-wrap') || trendRow?.querySelector('.perf-sparkline-wrap'); if (!trendRow) return; const history = perfHistoryLoad().filter(e => e.backend === backend && e.voice === voice && typeof e.avgRtf === 'number'); if (history.length === 0) { trendRow.style.display = 'none'; return; } const prevRtf = history[history.length - 1].avgRtf; const delta = currentAvgRtf - prevRtf; const pct = Math.abs(delta / Math.max(prevRtf, 0.01)) * 100; let cls, label; if (pct < 5) { cls = 'perf-trend-stable'; label = ' Stable'; } else if (delta < 0) { cls = 'perf-trend-better'; label = ` ${pct.toFixed(0)}% faster`; } else { cls = 'perf-trend-worse'; label = ` ${pct.toFixed(0)}% slower`; } trendBadge.className = 'perf-trend-badge ' + cls; trendBadge.innerHTML = label; const rtfValues = [...history.slice(-9).map(e => e.avgRtf), currentAvgRtf]; if (sparkWrap) sparkWrap.innerHTML = perfSparklineSvg(rtfValues); trendRow.style.display = ''; } function renderPerfTable(sessionDone = false) { if (!perfRows.length) { perfResultsCard.style.display='none'; return; } perfResultsCard.style.display = ''; perfTbody.innerHTML = perfRows.map((r, i) => { const rtf = r.audioDuration > 0 ? (r.latencyMs / 1000 / r.audioDuration).toFixed(2) : '—'; const ok = r.ok; return ` ${i+1} ${escHtml(r.backend)} ${escHtml(r.voice)} ${ok ? r.latencyMs : '—'} ${ok && r.audioDuration > 0 ? r.audioDuration.toFixed(2) : '—'} ${ok ? rtf : '—'} ${ok ? 'OK' : `${escHtml(r.error||'Error')}`} `; }).join(''); const ok = perfRows.filter(r => r.ok); if (ok.length) { const avg = ok.reduce((s,r) => s + r.latencyMs, 0) / ok.length; const minL = Math.min(...ok.map(r => r.latencyMs)); const maxL = Math.max(...ok.map(r => r.latencyMs)); const rtfArr = ok.filter(r=>r.audioDuration>0).map(r=>r.latencyMs/1000/r.audioDuration); const avgRtf = rtfArr.length ? rtfArr.reduce((s,v)=>s+v,0)/rtfArr.length : 0; const labelEl = $('perf-results-label'); if (labelEl) labelEl.textContent = `${ok.length} run${ok.length>1?'s':''} — ${perfBackendSel.value} / ${perfVoiceSel.value}`; perfSummary.innerHTML = ` ${Math.round(avg)} ms avg latency ${minL} ms best ${maxL} ms worst ${avgRtf.toFixed(2)} avg RTF ${avgRtf < 1 ? ' Real-time capable' : ' Slower than real-time'} `; if (sessionDone && rtfArr.length) { updateTrendDisplay(perfBackendSel.value, perfVoiceSel.value, avgRtf); perfHistoryAdd({ ts: Date.now(), backend: perfBackendSel.value, voice: perfVoiceSel.value, textLen: perfText.value.trim().length, avgLatencyMs: avg, minLatencyMs: minL, maxLatencyMs: maxL, avgRtf, runCount: ok.length, allOk: ok.length === perfRows.length, }); renderPerfHistory(); } } else { perfSummary.innerHTML = 'All runs failed'; } } perfClearBtn.addEventListener('click', () => { perfRows = []; renderPerfTable(); if ($('perf-trend-row')) $('perf-trend-row').style.display = 'none'; perfProgress.style.display = 'none'; }); perfRunBtn.addEventListener('click', async () => { const backend = perfBackendSel.value; const voice = perfVoiceSel.value; const text = perfText.value.trim(); const runs = parseInt(perfRunsSel.value) || 3; if (!backend) { toast('Select a backend first', 'error'); return; } if (!voice) { toast('Fetch and select a voice first', 'error'); return; } if (!text) { toast('Enter sample text', 'error'); return; } perfRows = []; perfRunBtn.disabled = true; perfProgress.style.display = ''; for (let i = 0; i < runs; i++) { perfProgress.textContent = `Run ${i+1} / ${runs}…`; const row = { backend, voice, ok: false, latencyMs: 0, audioDuration: 0, error: '' }; try { const t0 = performance.now(); const blob = await fetchTtsPreviewBlob(voice, text, 'wav', '', backend); row.latencyMs = Math.round(performance.now() - t0); row.ok = true; try { const audioCtx = new (window.AudioContext || window.webkitAudioContext)(); const buf = await audioCtx.decodeAudioData(await blob.arrayBuffer()); row.audioDuration = buf.duration; audioCtx.close(); } catch(_) {} } catch(e) { row.error = e.message; } perfRows.push(row); renderPerfTable(false); } perfProgress.textContent = `Done — ${runs} run${runs>1?'s':''} completed.`; renderPerfTable(true); perfRunBtn.disabled = false; }); // History filter toggle $('perf-history-filter-current')?.addEventListener('change', renderPerfHistory); $('perf-history-clear-btn')?.addEventListener('click', () => { perfHistorySave([]); renderPerfHistory(); toast('Benchmark history cleared', 'success'); }); renderPerfHistory(); })();